Annotation of gforth/doc/gforth.ds, revision 1.145

1.1       anton       1: \input texinfo   @c -*-texinfo-*-
                      2: @comment The source is gforth.ds, from which gforth.texi is generated
1.28      crook       3: 
1.21      crook       4: @comment TODO: nac29jan99 - a list of things to add in the next edit:
1.28      crook       5: @comment 1. x-ref all ambiguous or implementation-defined features?
                      6: @comment 2. Describe the use of Auser Avariable AConstant A, etc.
                      7: @comment 3. words in miscellaneous section need a home.
                      8: @comment 4. search for TODO for other minor and major works required.
                      9: @comment 5. [rats] change all @var to @i in Forth source so that info
                     10: @comment    file looks decent.
1.36      anton      11: @c          Not an improvement IMO - anton
                     12: @c          and anyway, this should be taken up
                     13: @c          with Karl Berry (the texinfo guy) - anton
1.113     anton      14: @c
                     15: @c Karl Berry writes:
                     16: @c  If they don't like the all-caps for @var Info output, all I can say is
                     17: @c  that it's always been that way, and the usage of all-caps for
                     18: @c  metavariables has a long tradition.  I think it's best to just let it be
                     19: @c  what it is, for the sake of consistency among manuals.
                     20: @c
1.29      crook      21: @comment .. would be useful to have a word that identified all deferred words
                     22: @comment should semantics stuff in intro be moved to another section
                     23: 
1.66      anton      24: @c POSTPONE, COMPILE, [COMPILE], LITERAL should have their own section
1.28      crook      25: 
1.1       anton      26: @comment %**start of header (This is for running Texinfo on a region.)
                     27: @setfilename gforth.info
1.113     anton      28: @include version.texi
1.1       anton      29: @settitle Gforth Manual
1.113     anton      30: @c @syncodeindex pg cp
1.49      anton      31: 
1.12      anton      32: @macro progstyle {}
                     33: Programming style note:
1.3       anton      34: @end macro
1.48      anton      35: 
                     36: @macro assignment {}
                     37: @table @i
                     38: @item Assignment:
                     39: @end macro
                     40: @macro endassignment {}
                     41: @end table
                     42: @end macro
                     43: 
1.29      crook      44: @comment macros for beautifying glossary entries
                     45: @macro GLOSS-START {}
                     46: @iftex
                     47: @ninerm
                     48: @end iftex
                     49: @end macro
                     50: 
                     51: @macro GLOSS-END {}
                     52: @iftex
                     53: @rm
                     54: @end iftex
                     55: @end macro
                     56: 
1.113     anton      57: @comment %**end of header (This is for running Texinfo on a region.)
                     58: @copying
1.125     anton      59: This manual is for Gforth (version @value{VERSION}, @value{UPDATED}),
                     60: a fast and portable implementation of the ANS Forth language.  It
                     61: serves as reference manual, but it also contains an introduction to
                     62: Forth and a Forth tutorial.
1.29      crook      63: 
1.142     anton      64: Copyright @copyright{} 1995, 1996, 1997, 1998, 2000, 2003, 2004,2005 Free Software Foundation, Inc.
1.29      crook      65: 
1.113     anton      66: @quotation
                     67: Permission is granted to copy, distribute and/or modify this document
                     68: under the terms of the GNU Free Documentation License, Version 1.1 or
                     69: any later version published by the Free Software Foundation; with no
                     70: Invariant Sections, with the Front-Cover texts being ``A GNU Manual,''
                     71: and with the Back-Cover Texts as in (a) below.  A copy of the
                     72: license is included in the section entitled ``GNU Free Documentation
                     73: License.''
                     74: 
                     75: (a) The FSF's Back-Cover Text is: ``You have freedom to copy and modify
                     76: this GNU Manual, like GNU software.  Copies published by the Free
                     77: Software Foundation raise funds for GNU development.''
                     78: @end quotation
                     79: @end copying
1.10      anton      80: 
1.113     anton      81: @dircategory Software development
                     82: @direntry
                     83: * Gforth: (gforth).             A fast interpreter for the Forth language.
                     84: @end direntry
                     85: @c The Texinfo manual also recommends doing this, but for Gforth it may
                     86: @c  not make much sense
                     87: @c @dircategory Individual utilities
                     88: @c @direntry
                     89: @c * Gforth: (gforth)Invoking Gforth.      gforth, gforth-fast, gforthmi
                     90: @c @end direntry
1.1       anton      91: 
                     92: @titlepage
1.113     anton      93: @title Gforth
                     94: @subtitle for version @value{VERSION}, @value{UPDATED}
                     95: @author Neal Crook
                     96: @author Anton Ertl
1.114     anton      97: @author David Kuehling
1.113     anton      98: @author Bernd Paysan
                     99: @author Jens Wilke
1.1       anton     100: @page
                    101: @vskip 0pt plus 1filll
1.113     anton     102: @insertcopying
                    103: @end titlepage
1.1       anton     104: 
1.113     anton     105: @contents
1.1       anton     106: 
1.113     anton     107: @ifnottex
                    108: @node Top, Goals, (dir), (dir)
                    109: @top Gforth
1.1       anton     110: 
1.113     anton     111: @insertcopying
1.49      anton     112: @end ifnottex
1.1       anton     113: 
                    114: @menu
1.26      crook     115: * Goals::                       About the Gforth Project
1.29      crook     116: * Gforth Environment::          Starting (and exiting) Gforth
1.48      anton     117: * Tutorial::                    Hands-on Forth Tutorial
1.21      crook     118: * Introduction::                An introduction to ANS Forth
1.1       anton     119: * Words::                       Forth words available in Gforth
1.24      anton     120: * Error messages::              How to interpret them
1.1       anton     121: * Tools::                       Programming tools
                    122: * ANS conformance::             Implementation-defined options etc.
1.65      anton     123: * Standard vs Extensions::      Should I use extensions?
1.1       anton     124: * Model::                       The abstract machine of Gforth
                    125: * Integrating Gforth::          Forth as scripting language for applications
                    126: * Emacs and Gforth::            The Gforth Mode
                    127: * Image Files::                 @code{.fi} files contain compiled code
                    128: * Engine::                      The inner interpreter and the primitives
1.13      pazsan    129: * Cross Compiler::              The Cross Compiler
1.1       anton     130: * Bugs::                        How to report them
                    131: * Origin::                      Authors and ancestors of Gforth
1.21      crook     132: * Forth-related information::   Books and places to look on the WWW
1.113     anton     133: * Licenses::                    
1.1       anton     134: * Word Index::                  An item for each Forth word
                    135: * Concept Index::               A menu covering many topics
1.12      anton     136: 
1.91      anton     137: @detailmenu
                    138:  --- The Detailed Node Listing ---
1.12      anton     139: 
1.29      crook     140: Gforth Environment
                    141: 
1.32      anton     142: * Invoking Gforth::             Getting in
                    143: * Leaving Gforth::              Getting out
                    144: * Command-line editing::        
1.48      anton     145: * Environment variables::       that affect how Gforth starts up
1.32      anton     146: * Gforth Files::                What gets installed and where
1.112     anton     147: * Gforth in pipes::             
1.48      anton     148: * Startup speed::               When 35ms is not fast enough ...
                    149: 
                    150: Forth Tutorial
                    151: 
                    152: * Starting Gforth Tutorial::    
                    153: * Syntax Tutorial::             
                    154: * Crash Course Tutorial::       
                    155: * Stack Tutorial::              
                    156: * Arithmetics Tutorial::        
                    157: * Stack Manipulation Tutorial::  
                    158: * Using files for Forth code Tutorial::  
                    159: * Comments Tutorial::           
                    160: * Colon Definitions Tutorial::  
                    161: * Decompilation Tutorial::      
                    162: * Stack-Effect Comments Tutorial::  
                    163: * Types Tutorial::              
                    164: * Factoring Tutorial::          
                    165: * Designing the stack effect Tutorial::  
                    166: * Local Variables Tutorial::    
                    167: * Conditional execution Tutorial::  
                    168: * Flags and Comparisons Tutorial::  
                    169: * General Loops Tutorial::      
                    170: * Counted loops Tutorial::      
                    171: * Recursion Tutorial::          
                    172: * Leaving definitions or loops Tutorial::  
                    173: * Return Stack Tutorial::       
                    174: * Memory Tutorial::             
                    175: * Characters and Strings Tutorial::  
                    176: * Alignment Tutorial::          
1.87      anton     177: * Files Tutorial::              
1.48      anton     178: * Interpretation and Compilation Semantics and Immediacy Tutorial::  
                    179: * Execution Tokens Tutorial::   
                    180: * Exceptions Tutorial::         
                    181: * Defining Words Tutorial::     
                    182: * Arrays and Records Tutorial::  
                    183: * POSTPONE Tutorial::           
                    184: * Literal Tutorial::            
                    185: * Advanced macros Tutorial::    
                    186: * Compilation Tokens Tutorial::  
                    187: * Wordlists and Search Order Tutorial::  
1.29      crook     188: 
1.24      anton     189: An Introduction to ANS Forth
                    190: 
1.67      anton     191: * Introducing the Text Interpreter::  
                    192: * Stacks and Postfix notation::  
                    193: * Your first definition::       
                    194: * How does that work?::         
                    195: * Forth is written in Forth::   
                    196: * Review - elements of a Forth system::  
                    197: * Where to go next::            
                    198: * Exercises::                   
1.24      anton     199: 
1.12      anton     200: Forth Words
                    201: 
                    202: * Notation::                    
1.65      anton     203: * Case insensitivity::          
                    204: * Comments::                    
                    205: * Boolean Flags::               
1.12      anton     206: * Arithmetic::                  
                    207: * Stack Manipulation::          
                    208: * Memory::                      
                    209: * Control Structures::          
                    210: * Defining Words::              
1.65      anton     211: * Interpretation and Compilation Semantics::  
1.47      crook     212: * Tokens for Words::            
1.81      anton     213: * Compiling words::             
1.65      anton     214: * The Text Interpreter::        
1.111     anton     215: * The Input Stream::            
1.65      anton     216: * Word Lists::                  
                    217: * Environmental Queries::       
1.12      anton     218: * Files::                       
                    219: * Blocks::                      
                    220: * Other I/O::                   
1.121     anton     221: * OS command line arguments::   
1.78      anton     222: * Locals::                      
                    223: * Structures::                  
                    224: * Object-oriented Forth::       
1.12      anton     225: * Programming Tools::           
                    226: * Assembler and Code Words::    
                    227: * Threading Words::             
1.65      anton     228: * Passing Commands to the OS::  
                    229: * Keeping track of Time::       
                    230: * Miscellaneous Words::         
1.12      anton     231: 
                    232: Arithmetic
                    233: 
                    234: * Single precision::            
1.67      anton     235: * Double precision::            Double-cell integer arithmetic
1.12      anton     236: * Bitwise operations::          
1.67      anton     237: * Numeric comparison::          
1.32      anton     238: * Mixed precision::             Operations with single and double-cell integers
1.12      anton     239: * Floating Point::              
                    240: 
                    241: Stack Manipulation
                    242: 
                    243: * Data stack::                  
                    244: * Floating point stack::        
                    245: * Return stack::                
                    246: * Locals stack::                
                    247: * Stack pointer manipulation::  
                    248: 
                    249: Memory
                    250: 
1.32      anton     251: * Memory model::                
                    252: * Dictionary allocation::       
                    253: * Heap Allocation::             
                    254: * Memory Access::               
                    255: * Address arithmetic::          
                    256: * Memory Blocks::               
1.12      anton     257: 
                    258: Control Structures
                    259: 
1.41      anton     260: * Selection::                   IF ... ELSE ... ENDIF
                    261: * Simple Loops::                BEGIN ...
1.32      anton     262: * Counted Loops::               DO
1.67      anton     263: * Arbitrary control structures::  
                    264: * Calls and returns::           
1.12      anton     265: * Exception Handling::          
                    266: 
                    267: Defining Words
                    268: 
1.67      anton     269: * CREATE::                      
1.44      crook     270: * Variables::                   Variables and user variables
1.67      anton     271: * Constants::                   
1.44      crook     272: * Values::                      Initialised variables
1.67      anton     273: * Colon Definitions::           
1.44      crook     274: * Anonymous Definitions::       Definitions without names
1.71      anton     275: * Supplying names::             Passing definition names as strings
1.67      anton     276: * User-defined Defining Words::  
1.44      crook     277: * Deferred words::              Allow forward references
1.67      anton     278: * Aliases::                     
1.47      crook     279: 
1.63      anton     280: User-defined Defining Words
                    281: 
                    282: * CREATE..DOES> applications::  
                    283: * CREATE..DOES> details::       
                    284: * Advanced does> usage example::  
1.91      anton     285: * @code{Const-does>}::          
1.63      anton     286: 
1.47      crook     287: Interpretation and Compilation Semantics
                    288: 
1.67      anton     289: * Combined words::              
1.12      anton     290: 
1.71      anton     291: Tokens for Words
                    292: 
                    293: * Execution token::             represents execution/interpretation semantics
                    294: * Compilation token::           represents compilation semantics
                    295: * Name token::                  represents named words
                    296: 
1.82      anton     297: Compiling words
                    298: 
                    299: * Literals::                    Compiling data values
                    300: * Macros::                      Compiling words
                    301: 
1.21      crook     302: The Text Interpreter
                    303: 
1.67      anton     304: * Input Sources::               
                    305: * Number Conversion::           
                    306: * Interpret/Compile states::    
                    307: * Interpreter Directives::      
1.21      crook     308: 
1.26      crook     309: Word Lists
                    310: 
1.75      anton     311: * Vocabularies::                
1.67      anton     312: * Why use word lists?::         
1.75      anton     313: * Word list example::           
1.26      crook     314: 
                    315: Files
                    316: 
1.48      anton     317: * Forth source files::          
                    318: * General files::               
                    319: * Search Paths::                
                    320: 
                    321: Search Paths
                    322: 
1.75      anton     323: * Source Search Paths::         
1.26      crook     324: * General Search Paths::        
                    325: 
                    326: Other I/O
                    327: 
1.32      anton     328: * Simple numeric output::       Predefined formats
                    329: * Formatted numeric output::    Formatted (pictured) output
                    330: * String Formats::              How Forth stores strings in memory
1.67      anton     331: * Displaying characters and strings::  Other stuff
1.32      anton     332: * Input::                       Input
1.112     anton     333: * Pipes::                       How to create your own pipes
1.26      crook     334: 
                    335: Locals
                    336: 
                    337: * Gforth locals::               
                    338: * ANS Forth locals::            
                    339: 
                    340: Gforth locals
                    341: 
                    342: * Where are locals visible by name?::  
                    343: * How long do locals live?::    
1.78      anton     344: * Locals programming style::    
                    345: * Locals implementation::       
1.26      crook     346: 
1.12      anton     347: Structures
                    348: 
                    349: * Why explicit structure support?::  
                    350: * Structure Usage::             
                    351: * Structure Naming Convention::  
                    352: * Structure Implementation::    
                    353: * Structure Glossary::          
                    354: 
                    355: Object-oriented Forth
                    356: 
1.48      anton     357: * Why object-oriented programming?::  
                    358: * Object-Oriented Terminology::  
                    359: * Objects::                     
                    360: * OOF::                         
                    361: * Mini-OOF::                    
1.23      crook     362: * Comparison with other object models::  
1.12      anton     363: 
1.24      anton     364: The @file{objects.fs} model
1.12      anton     365: 
                    366: * Properties of the Objects model::  
                    367: * Basic Objects Usage::         
1.41      anton     368: * The Objects base class::      
1.12      anton     369: * Creating objects::            
                    370: * Object-Oriented Programming Style::  
                    371: * Class Binding::               
                    372: * Method conveniences::         
                    373: * Classes and Scoping::         
1.41      anton     374: * Dividing classes::            
1.12      anton     375: * Object Interfaces::           
                    376: * Objects Implementation::      
                    377: * Objects Glossary::            
                    378: 
1.24      anton     379: The @file{oof.fs} model
1.12      anton     380: 
1.67      anton     381: * Properties of the OOF model::  
                    382: * Basic OOF Usage::             
                    383: * The OOF base class::          
                    384: * Class Declaration::           
                    385: * Class Implementation::        
1.12      anton     386: 
1.24      anton     387: The @file{mini-oof.fs} model
1.23      crook     388: 
1.48      anton     389: * Basic Mini-OOF Usage::        
                    390: * Mini-OOF Example::            
                    391: * Mini-OOF Implementation::     
1.23      crook     392: 
1.78      anton     393: Programming Tools
                    394: 
                    395: * Examining::                   
                    396: * Forgetting words::            
                    397: * Debugging::                   Simple and quick.
                    398: * Assertions::                  Making your programs self-checking.
                    399: * Singlestep Debugger::         Executing your program word by word.
                    400: 
                    401: Assembler and Code Words
                    402: 
                    403: * Code and ;code::              
                    404: * Common Assembler::            Assembler Syntax
                    405: * Common Disassembler::         
                    406: * 386 Assembler::               Deviations and special cases
                    407: * Alpha Assembler::             Deviations and special cases
                    408: * MIPS assembler::              Deviations and special cases
                    409: * Other assemblers::            How to write them
                    410: 
1.12      anton     411: Tools
                    412: 
                    413: * ANS Report::                  Report the words used, sorted by wordset.
1.127     anton     414: * Stack depth changes::         Where does this stack item come from?
1.12      anton     415: 
                    416: ANS conformance
                    417: 
                    418: * The Core Words::              
                    419: * The optional Block word set::  
                    420: * The optional Double Number word set::  
                    421: * The optional Exception word set::  
                    422: * The optional Facility word set::  
                    423: * The optional File-Access word set::  
                    424: * The optional Floating-Point word set::  
                    425: * The optional Locals word set::  
                    426: * The optional Memory-Allocation word set::  
                    427: * The optional Programming-Tools word set::  
                    428: * The optional Search-Order word set::  
                    429: 
                    430: The Core Words
                    431: 
                    432: * core-idef::                   Implementation Defined Options                   
                    433: * core-ambcond::                Ambiguous Conditions                
                    434: * core-other::                  Other System Documentation                  
                    435: 
                    436: The optional Block word set
                    437: 
                    438: * block-idef::                  Implementation Defined Options
                    439: * block-ambcond::               Ambiguous Conditions               
                    440: * block-other::                 Other System Documentation                 
                    441: 
                    442: The optional Double Number word set
                    443: 
                    444: * double-ambcond::              Ambiguous Conditions              
                    445: 
                    446: The optional Exception word set
                    447: 
                    448: * exception-idef::              Implementation Defined Options              
                    449: 
                    450: The optional Facility word set
                    451: 
                    452: * facility-idef::               Implementation Defined Options               
                    453: * facility-ambcond::            Ambiguous Conditions            
                    454: 
                    455: The optional File-Access word set
                    456: 
                    457: * file-idef::                   Implementation Defined Options
                    458: * file-ambcond::                Ambiguous Conditions                
                    459: 
                    460: The optional Floating-Point word set
                    461: 
                    462: * floating-idef::               Implementation Defined Options
                    463: * floating-ambcond::            Ambiguous Conditions            
                    464: 
                    465: The optional Locals word set
                    466: 
                    467: * locals-idef::                 Implementation Defined Options                 
                    468: * locals-ambcond::              Ambiguous Conditions              
                    469: 
                    470: The optional Memory-Allocation word set
                    471: 
                    472: * memory-idef::                 Implementation Defined Options                 
                    473: 
                    474: The optional Programming-Tools word set
                    475: 
                    476: * programming-idef::            Implementation Defined Options            
                    477: * programming-ambcond::         Ambiguous Conditions         
                    478: 
                    479: The optional Search-Order word set
                    480: 
                    481: * search-idef::                 Implementation Defined Options                 
                    482: * search-ambcond::              Ambiguous Conditions              
                    483: 
1.109     anton     484: Emacs and Gforth
                    485: 
                    486: * Installing gforth.el::        Making Emacs aware of Forth.
                    487: * Emacs Tags::                  Viewing the source of a word in Emacs.
                    488: * Hilighting::                  Making Forth code look prettier.
                    489: * Auto-Indentation::            Customizing auto-indentation.
                    490: * Blocks Files::                Reading and writing blocks files.
                    491: 
1.12      anton     492: Image Files
                    493: 
1.24      anton     494: * Image Licensing Issues::      Distribution terms for images.
                    495: * Image File Background::       Why have image files?
1.67      anton     496: * Non-Relocatable Image Files::  don't always work.
1.24      anton     497: * Data-Relocatable Image Files::  are better.
1.67      anton     498: * Fully Relocatable Image Files::  better yet.
1.24      anton     499: * Stack and Dictionary Sizes::  Setting the default sizes for an image.
1.32      anton     500: * Running Image Files::         @code{gforth -i @i{file}} or @i{file}.
1.24      anton     501: * Modifying the Startup Sequence::  and turnkey applications.
1.12      anton     502: 
                    503: Fully Relocatable Image Files
                    504: 
1.27      crook     505: * gforthmi::                    The normal way
1.12      anton     506: * cross.fs::                    The hard way
                    507: 
                    508: Engine
                    509: 
                    510: * Portability::                 
                    511: * Threading::                   
                    512: * Primitives::                  
                    513: * Performance::                 
                    514: 
                    515: Threading
                    516: 
                    517: * Scheduling::                  
                    518: * Direct or Indirect Threaded?::  
1.109     anton     519: * Dynamic Superinstructions::   
1.12      anton     520: * DOES>::                       
                    521: 
                    522: Primitives
                    523: 
                    524: * Automatic Generation::        
                    525: * TOS Optimization::            
                    526: * Produced code::               
1.13      pazsan    527: 
                    528: Cross Compiler
                    529: 
1.67      anton     530: * Using the Cross Compiler::    
                    531: * How the Cross Compiler Works::  
1.13      pazsan    532: 
1.113     anton     533: Licenses
                    534: 
                    535: * GNU Free Documentation License::  License for copying this manual.
                    536: * Copying::                         GPL (for copying this software).
                    537: 
1.24      anton     538: @end detailmenu
1.1       anton     539: @end menu
                    540: 
1.113     anton     541: @c ----------------------------------------------------------
1.1       anton     542: @iftex
                    543: @unnumbered Preface
                    544: @cindex Preface
1.21      crook     545: This manual documents Gforth. Some introductory material is provided for
                    546: readers who are unfamiliar with Forth or who are migrating to Gforth
                    547: from other Forth compilers. However, this manual is primarily a
                    548: reference manual.
1.1       anton     549: @end iftex
                    550: 
1.28      crook     551: @comment TODO much more blurb here.
1.26      crook     552: 
                    553: @c ******************************************************************
1.113     anton     554: @node Goals, Gforth Environment, Top, Top
1.26      crook     555: @comment node-name,     next,           previous, up
                    556: @chapter Goals of Gforth
                    557: @cindex goals of the Gforth project
                    558: The goal of the Gforth Project is to develop a standard model for
                    559: ANS Forth. This can be split into several subgoals:
                    560: 
                    561: @itemize @bullet
                    562: @item
                    563: Gforth should conform to the ANS Forth Standard.
                    564: @item
                    565: It should be a model, i.e. it should define all the
                    566: implementation-dependent things.
                    567: @item
                    568: It should become standard, i.e. widely accepted and used. This goal
                    569: is the most difficult one.
                    570: @end itemize
                    571: 
                    572: To achieve these goals Gforth should be
                    573: @itemize @bullet
                    574: @item
                    575: Similar to previous models (fig-Forth, F83)
                    576: @item
                    577: Powerful. It should provide for all the things that are considered
                    578: necessary today and even some that are not yet considered necessary.
                    579: @item
                    580: Efficient. It should not get the reputation of being exceptionally
                    581: slow.
                    582: @item
                    583: Free.
                    584: @item
                    585: Available on many machines/easy to port.
                    586: @end itemize
                    587: 
                    588: Have we achieved these goals? Gforth conforms to the ANS Forth
                    589: standard. It may be considered a model, but we have not yet documented
                    590: which parts of the model are stable and which parts we are likely to
                    591: change. It certainly has not yet become a de facto standard, but it
                    592: appears to be quite popular. It has some similarities to and some
                    593: differences from previous models. It has some powerful features, but not
                    594: yet everything that we envisioned. We certainly have achieved our
1.65      anton     595: execution speed goals (@pxref{Performance})@footnote{However, in 1998
                    596: the bar was raised when the major commercial Forth vendors switched to
                    597: native code compilers.}.  It is free and available on many machines.
1.29      crook     598: 
1.26      crook     599: @c ******************************************************************
1.48      anton     600: @node Gforth Environment, Tutorial, Goals, Top
1.29      crook     601: @chapter Gforth Environment
                    602: @cindex Gforth environment
1.21      crook     603: 
1.45      crook     604: Note: ultimately, the Gforth man page will be auto-generated from the
1.29      crook     605: material in this chapter.
1.21      crook     606: 
                    607: @menu
1.29      crook     608: * Invoking Gforth::             Getting in
                    609: * Leaving Gforth::              Getting out
                    610: * Command-line editing::        
1.48      anton     611: * Environment variables::       that affect how Gforth starts up
1.29      crook     612: * Gforth Files::                What gets installed and where
1.112     anton     613: * Gforth in pipes::             
1.48      anton     614: * Startup speed::               When 35ms is not fast enough ...
1.21      crook     615: @end menu
                    616: 
1.49      anton     617: For related information about the creation of images see @ref{Image Files}.
1.29      crook     618: 
1.21      crook     619: @comment ----------------------------------------------
1.48      anton     620: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1.29      crook     621: @section Invoking Gforth
                    622: @cindex invoking Gforth
                    623: @cindex running Gforth
                    624: @cindex command-line options
                    625: @cindex options on the command line
                    626: @cindex flags on the command line
1.21      crook     627: 
1.30      anton     628: Gforth is made up of two parts; an executable ``engine'' (named
1.109     anton     629: @command{gforth} or @command{gforth-fast}) and an image file. To start it, you
1.30      anton     630: will usually just say @code{gforth} -- this automatically loads the
                    631: default image file @file{gforth.fi}. In many other cases the default
                    632: Gforth image will be invoked like this:
1.21      crook     633: @example
1.30      anton     634: gforth [file | -e forth-code] ...
1.21      crook     635: @end example
1.29      crook     636: @noindent
                    637: This interprets the contents of the files and the Forth code in the order they
                    638: are given.
1.21      crook     639: 
1.109     anton     640: In addition to the @command{gforth} engine, there is also an engine
                    641: called @command{gforth-fast}, which is faster, but gives less
                    642: informative error messages (@pxref{Error messages}) and may catch some
                    643: stack underflows later or not at all.  You should use it for debugged,
                    644: performance-critical programs.
                    645: 
                    646: Moreover, there is an engine called @command{gforth-itc}, which is
                    647: useful in some backwards-compatibility situations (@pxref{Direct or
                    648: Indirect Threaded?}).
1.30      anton     649: 
1.29      crook     650: In general, the command line looks like this:
1.21      crook     651: 
                    652: @example
1.30      anton     653: gforth[-fast] [engine options] [image options]
1.21      crook     654: @end example
                    655: 
1.30      anton     656: The engine options must come before the rest of the command
1.29      crook     657: line. They are:
1.26      crook     658: 
1.29      crook     659: @table @code
                    660: @cindex -i, command-line option
                    661: @cindex --image-file, command-line option
                    662: @item --image-file @i{file}
                    663: @itemx -i @i{file}
                    664: Loads the Forth image @i{file} instead of the default
                    665: @file{gforth.fi} (@pxref{Image Files}).
1.21      crook     666: 
1.39      anton     667: @cindex --appl-image, command-line option
                    668: @item --appl-image @i{file}
                    669: Loads the image @i{file} and leaves all further command-line arguments
1.65      anton     670: to the image (instead of processing them as engine options).  This is
                    671: useful for building executable application images on Unix, built with
1.39      anton     672: @code{gforthmi --application ...}.
                    673: 
1.29      crook     674: @cindex --path, command-line option
                    675: @cindex -p, command-line option
                    676: @item --path @i{path}
                    677: @itemx -p @i{path}
                    678: Uses @i{path} for searching the image file and Forth source code files
                    679: instead of the default in the environment variable @code{GFORTHPATH} or
                    680: the path specified at installation time (e.g.,
                    681: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
                    682: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1.21      crook     683: 
1.29      crook     684: @cindex --dictionary-size, command-line option
                    685: @cindex -m, command-line option
                    686: @cindex @i{size} parameters for command-line options
                    687: @cindex size of the dictionary and the stacks
                    688: @item --dictionary-size @i{size}
                    689: @itemx -m @i{size}
                    690: Allocate @i{size} space for the Forth dictionary space instead of
                    691: using the default specified in the image (typically 256K). The
                    692: @i{size} specification for this and subsequent options consists of
                    693: an integer and a unit (e.g.,
                    694: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
                    695: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
                    696: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
                    697: @code{e} is used.
1.21      crook     698: 
1.29      crook     699: @cindex --data-stack-size, command-line option
                    700: @cindex -d, command-line option
                    701: @item --data-stack-size @i{size}
                    702: @itemx -d @i{size}
                    703: Allocate @i{size} space for the data stack instead of using the
                    704: default specified in the image (typically 16K).
1.21      crook     705: 
1.29      crook     706: @cindex --return-stack-size, command-line option
                    707: @cindex -r, command-line option
                    708: @item --return-stack-size @i{size}
                    709: @itemx -r @i{size}
                    710: Allocate @i{size} space for the return stack instead of using the
                    711: default specified in the image (typically 15K).
1.21      crook     712: 
1.29      crook     713: @cindex --fp-stack-size, command-line option
                    714: @cindex -f, command-line option
                    715: @item --fp-stack-size @i{size}
                    716: @itemx -f @i{size}
                    717: Allocate @i{size} space for the floating point stack instead of
                    718: using the default specified in the image (typically 15.5K). In this case
                    719: the unit specifier @code{e} refers to floating point numbers.
1.21      crook     720: 
1.48      anton     721: @cindex --locals-stack-size, command-line option
                    722: @cindex -l, command-line option
                    723: @item --locals-stack-size @i{size}
                    724: @itemx -l @i{size}
                    725: Allocate @i{size} space for the locals stack instead of using the
                    726: default specified in the image (typically 14.5K).
                    727: 
                    728: @cindex -h, command-line option
                    729: @cindex --help, command-line option
                    730: @item --help
                    731: @itemx -h
                    732: Print a message about the command-line options
                    733: 
                    734: @cindex -v, command-line option
                    735: @cindex --version, command-line option
                    736: @item --version
                    737: @itemx -v
                    738: Print version and exit
                    739: 
                    740: @cindex --debug, command-line option
                    741: @item --debug
                    742: Print some information useful for debugging on startup.
                    743: 
                    744: @cindex --offset-image, command-line option
                    745: @item --offset-image
                    746: Start the dictionary at a slightly different position than would be used
                    747: otherwise (useful for creating data-relocatable images,
                    748: @pxref{Data-Relocatable Image Files}).
                    749: 
                    750: @cindex --no-offset-im, command-line option
                    751: @item --no-offset-im
                    752: Start the dictionary at the normal position.
                    753: 
                    754: @cindex --clear-dictionary, command-line option
                    755: @item --clear-dictionary
                    756: Initialize all bytes in the dictionary to 0 before loading the image
                    757: (@pxref{Data-Relocatable Image Files}).
                    758: 
                    759: @cindex --die-on-signal, command-line-option
                    760: @item --die-on-signal
                    761: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
                    762: or the segmentation violation SIGSEGV) by translating it into a Forth
                    763: @code{THROW}. With this option, Gforth exits if it receives such a
                    764: signal. This option is useful when the engine and/or the image might be
                    765: severely broken (such that it causes another signal before recovering
                    766: from the first); this option avoids endless loops in such cases.
1.109     anton     767: 
1.119     anton     768: @cindex --no-dynamic, command-line option
                    769: @cindex --dynamic, command-line option
1.109     anton     770: @item --no-dynamic
                    771: @item --dynamic
                    772: Disable or enable dynamic superinstructions with replication
                    773: (@pxref{Dynamic Superinstructions}).
                    774: 
1.119     anton     775: @cindex --no-super, command-line option
1.109     anton     776: @item --no-super
1.110     anton     777: Disable dynamic superinstructions, use just dynamic replication; this is
                    778: useful if you want to patch threaded code (@pxref{Dynamic
                    779: Superinstructions}).
1.119     anton     780: 
                    781: @cindex --ss-number, command-line option
                    782: @item --ss-number=@var{N}
                    783: Use only the first @var{N} static superinstructions compiled into the
                    784: engine (default: use them all; note that only @code{gforth-fast} has
                    785: any).  This option is useful for measuring the performance impact of
                    786: static superinstructions.
                    787: 
                    788: @cindex --ss-min-..., command-line options
                    789: @item --ss-min-codesize
                    790: @item --ss-min-ls
                    791: @item --ss-min-lsu
                    792: @item --ss-min-nexts
                    793: Use specified metric for determining the cost of a primitive or static
                    794: superinstruction for static superinstruction selection.  @code{Codesize}
                    795: is the native code size of the primive or static superinstruction,
                    796: @code{ls} is the number of loads and stores, @code{lsu} is the number of
                    797: loads, stores, and updates, and @code{nexts} is the number of dispatches
                    798: (not taking dynamic superinstructions into account), i.e. every
                    799: primitive or static superinstruction has cost 1. Default:
                    800: @code{codesize} if you use dynamic code generation, otherwise
                    801: @code{nexts}.
                    802: 
                    803: @cindex --ss-greedy, command-line option
                    804: @item --ss-greedy
                    805: This option is useful for measuring the performance impact of static
                    806: superinstructions.  By default, an optimal shortest-path algorithm is
                    807: used for selecting static superinstructions.  With @option{--ss-greedy}
                    808: this algorithm is modified to assume that anything after the static
                    809: superinstruction currently under consideration is not combined into
                    810: static superinstructions.  With @option{--ss-min-nexts} this produces
                    811: the same result as a greedy algorithm that always selects the longest
                    812: superinstruction available at the moment.  E.g., if there are
                    813: superinstructions AB and BCD, then for the sequence A B C D the optimal
                    814: algorithm will select A BCD and the greedy algorithm will select AB C D.
                    815: 
                    816: @cindex --print-metrics, command-line option
                    817: @item --print-metrics
                    818: Prints some metrics used during static superinstruction selection:
                    819: @code{code size} is the actual size of the dynamically generated code.
                    820: @code{Metric codesize} is the sum of the codesize metrics as seen by
                    821: static superinstruction selection; there is a difference from @code{code
                    822: size}, because not all primitives and static superinstructions are
                    823: compiled into dynamically generated code, and because of markers.  The
                    824: other metrics correspond to the @option{ss-min-...} options.  This
                    825: option is useful for evaluating the effects of the @option{--ss-...}
                    826: options.
1.109     anton     827: 
1.48      anton     828: @end table
                    829: 
                    830: @cindex loading files at startup
                    831: @cindex executing code on startup
                    832: @cindex batch processing with Gforth
                    833: As explained above, the image-specific command-line arguments for the
                    834: default image @file{gforth.fi} consist of a sequence of filenames and
                    835: @code{-e @var{forth-code}} options that are interpreted in the sequence
                    836: in which they are given. The @code{-e @var{forth-code}} or
1.121     anton     837: @code{--evaluate @var{forth-code}} option evaluates the Forth code. This
                    838: option takes only one argument; if you want to evaluate more Forth
                    839: words, you have to quote them or use @code{-e} several times. To exit
1.48      anton     840: after processing the command line (instead of entering interactive mode)
1.121     anton     841: append @code{-e bye} to the command line.  You can also process the
                    842: command-line arguments with a Forth program (@pxref{OS command line
                    843: arguments}).
1.48      anton     844: 
                    845: @cindex versions, invoking other versions of Gforth
                    846: If you have several versions of Gforth installed, @code{gforth} will
                    847: invoke the version that was installed last. @code{gforth-@i{version}}
                    848: invokes a specific version. If your environment contains the variable
                    849: @code{GFORTHPATH}, you may want to override it by using the
                    850: @code{--path} option.
                    851: 
                    852: Not yet implemented:
                    853: On startup the system first executes the system initialization file
                    854: (unless the option @code{--no-init-file} is given; note that the system
                    855: resulting from using this option may not be ANS Forth conformant). Then
                    856: the user initialization file @file{.gforth.fs} is executed, unless the
1.62      crook     857: option @code{--no-rc} is given; this file is searched for in @file{.},
1.48      anton     858: then in @file{~}, then in the normal path (see above).
                    859: 
                    860: 
                    861: 
                    862: @comment ----------------------------------------------
                    863: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
                    864: @section Leaving Gforth
                    865: @cindex Gforth - leaving
                    866: @cindex leaving Gforth
                    867: 
                    868: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
                    869: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
                    870: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1.49      anton     871: data are discarded.  For ways of saving the state of the system before
                    872: leaving Gforth see @ref{Image Files}.
1.48      anton     873: 
                    874: doc-bye
                    875: 
                    876: 
                    877: @comment ----------------------------------------------
1.65      anton     878: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1.48      anton     879: @section Command-line editing
                    880: @cindex command-line editing
                    881: 
                    882: Gforth maintains a history file that records every line that you type to
                    883: the text interpreter. This file is preserved between sessions, and is
                    884: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
                    885: repeatedly you can recall successively older commands from this (or
                    886: previous) session(s). The full list of command-line editing facilities is:
                    887: 
                    888: @itemize @bullet
                    889: @item
                    890: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
                    891: commands from the history buffer.
                    892: @item
                    893: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
                    894: from the history buffer.
                    895: @item
                    896: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
                    897: @item
                    898: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
                    899: @item
                    900: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
                    901: closing up the line.
                    902: @item
                    903: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
                    904: @item
                    905: @kbd{Ctrl-a} to move the cursor to the start of the line.
                    906: @item
                    907: @kbd{Ctrl-e} to move the cursor to the end of the line.
                    908: @item
                    909: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
                    910: line.
                    911: @item
                    912: @key{TAB} to step through all possible full-word completions of the word
                    913: currently being typed.
                    914: @item
1.65      anton     915: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
                    916: using @code{bye}). 
                    917: @item
                    918: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
                    919: character under the cursor.
1.48      anton     920: @end itemize
                    921: 
                    922: When editing, displayable characters are inserted to the left of the
                    923: cursor position; the line is always in ``insert'' (as opposed to
                    924: ``overstrike'') mode.
                    925: 
                    926: @cindex history file
                    927: @cindex @file{.gforth-history}
                    928: On Unix systems, the history file is @file{~/.gforth-history} by
                    929: default@footnote{i.e. it is stored in the user's home directory.}. You
                    930: can find out the name and location of your history file using:
                    931: 
                    932: @example 
                    933: history-file type \ Unix-class systems
                    934: 
                    935: history-file type \ Other systems
                    936: history-dir  type
                    937: @end example
                    938: 
                    939: If you enter long definitions by hand, you can use a text editor to
                    940: paste them out of the history file into a Forth source file for reuse at
                    941: a later time.
                    942: 
                    943: Gforth never trims the size of the history file, so you should do this
                    944: periodically, if necessary.
                    945: 
                    946: @comment this is all defined in history.fs
                    947: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
                    948: @comment chosen?
                    949: 
                    950: 
                    951: @comment ----------------------------------------------
1.65      anton     952: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1.48      anton     953: @section Environment variables
                    954: @cindex environment variables
                    955: 
                    956: Gforth uses these environment variables:
                    957: 
                    958: @itemize @bullet
                    959: @item
                    960: @cindex @code{GFORTHHIST} -- environment variable
                    961: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
                    962: open/create the history file, @file{.gforth-history}. Default:
                    963: @code{$HOME}.
                    964: 
                    965: @item
                    966: @cindex @code{GFORTHPATH} -- environment variable
                    967: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
                    968: for Forth source-code files.
                    969: 
                    970: @item
1.129     anton     971: @cindex @code{GFORTHSYSTEMPREFIX} -- environment variable
                    972: 
                    973: @code{GFORTHSYSTEMPREFIX} -- specifies what to prepend to the argument
                    974: of @code{system} before passing it to C's @code{system()}.  Default:
1.130     anton     975: @code{"./$COMSPEC /c "} on Windows, @code{""} on other OSs.  The prefix
1.129     anton     976: and the command are directly concatenated, so if a space between them is
                    977: necessary, append it to the prefix.
                    978: 
                    979: @item
1.48      anton     980: @cindex @code{GFORTH} -- environment variable
1.49      anton     981: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48      anton     982: 
                    983: @item
                    984: @cindex @code{GFORTHD} -- environment variable
1.62      crook     985: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48      anton     986: 
                    987: @item
                    988: @cindex @code{TMP}, @code{TEMP} - environment variable
                    989: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
                    990: location for the history file.
                    991: @end itemize
                    992: 
                    993: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
                    994: @comment mentioning these.
                    995: 
                    996: All the Gforth environment variables default to sensible values if they
                    997: are not set.
                    998: 
                    999: 
                   1000: @comment ----------------------------------------------
1.112     anton    1001: @node Gforth Files, Gforth in pipes, Environment variables, Gforth Environment
1.48      anton    1002: @section Gforth files
                   1003: @cindex Gforth files
                   1004: 
                   1005: When you install Gforth on a Unix system, it installs files in these
                   1006: locations by default:
                   1007: 
                   1008: @itemize @bullet
                   1009: @item
                   1010: @file{/usr/local/bin/gforth}
                   1011: @item
                   1012: @file{/usr/local/bin/gforthmi}
                   1013: @item
                   1014: @file{/usr/local/man/man1/gforth.1} - man page.
                   1015: @item
                   1016: @file{/usr/local/info} - the Info version of this manual.
                   1017: @item
                   1018: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
                   1019: @item
                   1020: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
                   1021: @item
                   1022: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
                   1023: @item
                   1024: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
                   1025: @end itemize
                   1026: 
                   1027: You can select different places for installation by using
                   1028: @code{configure} options (listed with @code{configure --help}).
                   1029: 
                   1030: @comment ----------------------------------------------
1.112     anton    1031: @node Gforth in pipes, Startup speed, Gforth Files, Gforth Environment
                   1032: @section Gforth in pipes
                   1033: @cindex pipes, Gforth as part of
                   1034: 
                   1035: Gforth can be used in pipes created elsewhere (described here).  It can
                   1036: also create pipes on its own (@pxref{Pipes}).
                   1037: 
                   1038: @cindex input from pipes
                   1039: If you pipe into Gforth, your program should read with @code{read-file}
                   1040: or @code{read-line} from @code{stdin} (@pxref{General files}).
                   1041: @code{Key} does not recognize the end of input.  Words like
                   1042: @code{accept} echo the input and are therefore usually not useful for
                   1043: reading from a pipe.  You have to invoke the Forth program with an OS
                   1044: command-line option, as you have no chance to use the Forth command line
                   1045: (the text interpreter would try to interpret the pipe input).
                   1046: 
                   1047: @cindex output in pipes
                   1048: You can output to a pipe with @code{type}, @code{emit}, @code{cr} etc.
                   1049: 
                   1050: @cindex silent exiting from Gforth
                   1051: When you write to a pipe that has been closed at the other end, Gforth
                   1052: receives a SIGPIPE signal (``pipe broken'').  Gforth translates this
                   1053: into the exception @code{broken-pipe-error}.  If your application does
                   1054: not catch that exception, the system catches it and exits, usually
                   1055: silently (unless you were working on the Forth command line; then it
                   1056: prints an error message and exits).  This is usually the desired
                   1057: behaviour.
                   1058: 
                   1059: If you do not like this behaviour, you have to catch the exception
                   1060: yourself, and react to it.
                   1061: 
                   1062: Here's an example of an invocation of Gforth that is usable in a pipe:
                   1063: 
                   1064: @example
                   1065: gforth -e ": foo begin pad dup 10 stdin read-file throw dup while \
                   1066:  type repeat ; foo bye"
                   1067: @end example
                   1068: 
                   1069: This example just copies the input verbatim to the output.  A very
                   1070: simple pipe containing this example looks like this:
                   1071: 
                   1072: @example
                   1073: cat startup.fs |
                   1074: gforth -e ": foo begin pad dup 80 stdin read-file throw dup while \
                   1075:  type repeat ; foo bye"|
                   1076: head
                   1077: @end example
                   1078: 
                   1079: @cindex stderr and pipes
                   1080: Pipes involving Gforth's @code{stderr} output do not work.
                   1081: 
                   1082: @comment ----------------------------------------------
                   1083: @node Startup speed,  , Gforth in pipes, Gforth Environment
1.48      anton    1084: @section Startup speed
                   1085: @cindex Startup speed
                   1086: @cindex speed, startup
                   1087: 
                   1088: If Gforth is used for CGI scripts or in shell scripts, its startup
                   1089: speed may become a problem.  On a 300MHz 21064a under Linux-2.2.13 with
                   1090: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
                   1091: system time.
                   1092: 
                   1093: If startup speed is a problem, you may consider the following ways to
                   1094: improve it; or you may consider ways to reduce the number of startups
1.62      crook    1095: (for example, by using Fast-CGI).
1.48      anton    1096: 
1.112     anton    1097: An easy step that influences Gforth startup speed is the use of the
                   1098: @option{--no-dynamic} option; this decreases image loading speed, but
                   1099: increases compile-time and run-time.
                   1100: 
                   1101: Another step to improve startup speed is to statically link Gforth, by
1.48      anton    1102: building it with @code{XLDFLAGS=-static}.  This requires more memory for
                   1103: the code and will therefore slow down the first invocation, but
                   1104: subsequent invocations avoid the dynamic linking overhead.  Another
                   1105: disadvantage is that Gforth won't profit from library upgrades.  As a
                   1106: result, @code{gforth-static -e bye} takes about 17.1ms user and
                   1107: 8.2ms system time.
                   1108: 
                   1109: The next step to improve startup speed is to use a non-relocatable image
1.65      anton    1110: (@pxref{Non-Relocatable Image Files}).  You can create this image with
1.48      anton    1111: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
                   1112: @code{gforth -i gforthnr.fi ...}.  This avoids the relocation overhead
                   1113: and a part of the copy-on-write overhead.  The disadvantage is that the
1.62      crook    1114: non-relocatable image does not work if the OS gives Gforth a different
1.48      anton    1115: address for the dictionary, for whatever reason; so you better provide a
                   1116: fallback on a relocatable image.  @code{gforth-static -i gforthnr.fi -e
                   1117: bye} takes about 15.3ms user and 7.5ms system time.
                   1118: 
                   1119: The final step is to disable dictionary hashing in Gforth.  Gforth
                   1120: builds the hash table on startup, which takes much of the startup
                   1121: overhead. You can do this by commenting out the @code{include hash.fs}
                   1122: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
                   1123: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
                   1124: The disadvantages are that functionality like @code{table} and
                   1125: @code{ekey} is missing and that text interpretation (e.g., compiling)
                   1126: now takes much longer. So, you should only use this method if there is
                   1127: no significant text interpretation to perform (the script should be
1.62      crook    1128: compiled into the image, amongst other things).  @code{gforth-static -i
1.48      anton    1129: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
                   1130: 
                   1131: @c ******************************************************************
                   1132: @node Tutorial, Introduction, Gforth Environment, Top
                   1133: @chapter Forth Tutorial
                   1134: @cindex Tutorial
                   1135: @cindex Forth Tutorial
                   1136: 
1.67      anton    1137: @c Topics from nac's Introduction that could be mentioned:
                   1138: @c press <ret> after each line
                   1139: @c Prompt
                   1140: @c numbers vs. words in dictionary on text interpretation
                   1141: @c what happens on redefinition
                   1142: @c parsing words (in particular, defining words)
                   1143: 
1.83      anton    1144: The difference of this chapter from the Introduction
                   1145: (@pxref{Introduction}) is that this tutorial is more fast-paced, should
                   1146: be used while sitting in front of a computer, and covers much more
                   1147: material, but does not explain how the Forth system works.
                   1148: 
1.62      crook    1149: This tutorial can be used with any ANS-compliant Forth; any
                   1150: Gforth-specific features are marked as such and you can skip them if you
                   1151: work with another Forth.  This tutorial does not explain all features of
                   1152: Forth, just enough to get you started and give you some ideas about the
                   1153: facilities available in Forth.  Read the rest of the manual and the
                   1154: standard when you are through this.
1.48      anton    1155: 
                   1156: The intended way to use this tutorial is that you work through it while
                   1157: sitting in front of the console, take a look at the examples and predict
                   1158: what they will do, then try them out; if the outcome is not as expected,
                   1159: find out why (e.g., by trying out variations of the example), so you
                   1160: understand what's going on.  There are also some assignments that you
                   1161: should solve.
                   1162: 
                   1163: This tutorial assumes that you have programmed before and know what,
                   1164: e.g., a loop is.
                   1165: 
                   1166: @c !! explain compat library
                   1167: 
                   1168: @menu
                   1169: * Starting Gforth Tutorial::    
                   1170: * Syntax Tutorial::             
                   1171: * Crash Course Tutorial::       
                   1172: * Stack Tutorial::              
                   1173: * Arithmetics Tutorial::        
                   1174: * Stack Manipulation Tutorial::  
                   1175: * Using files for Forth code Tutorial::  
                   1176: * Comments Tutorial::           
                   1177: * Colon Definitions Tutorial::  
                   1178: * Decompilation Tutorial::      
                   1179: * Stack-Effect Comments Tutorial::  
                   1180: * Types Tutorial::              
                   1181: * Factoring Tutorial::          
                   1182: * Designing the stack effect Tutorial::  
                   1183: * Local Variables Tutorial::    
                   1184: * Conditional execution Tutorial::  
                   1185: * Flags and Comparisons Tutorial::  
                   1186: * General Loops Tutorial::      
                   1187: * Counted loops Tutorial::      
                   1188: * Recursion Tutorial::          
                   1189: * Leaving definitions or loops Tutorial::  
                   1190: * Return Stack Tutorial::       
                   1191: * Memory Tutorial::             
                   1192: * Characters and Strings Tutorial::  
                   1193: * Alignment Tutorial::          
1.87      anton    1194: * Files Tutorial::              
1.48      anton    1195: * Interpretation and Compilation Semantics and Immediacy Tutorial::  
                   1196: * Execution Tokens Tutorial::   
                   1197: * Exceptions Tutorial::         
                   1198: * Defining Words Tutorial::     
                   1199: * Arrays and Records Tutorial::  
                   1200: * POSTPONE Tutorial::           
                   1201: * Literal Tutorial::            
                   1202: * Advanced macros Tutorial::    
                   1203: * Compilation Tokens Tutorial::  
                   1204: * Wordlists and Search Order Tutorial::  
                   1205: @end menu
                   1206: 
                   1207: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
                   1208: @section Starting Gforth
1.66      anton    1209: @cindex starting Gforth tutorial
1.48      anton    1210: You can start Gforth by typing its name:
                   1211: 
                   1212: @example
                   1213: gforth
                   1214: @end example
                   1215: 
                   1216: That puts you into interactive mode; you can leave Gforth by typing
                   1217: @code{bye}.  While in Gforth, you can edit the command line and access
                   1218: the command line history with cursor keys, similar to bash.
                   1219: 
                   1220: 
                   1221: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
                   1222: @section Syntax
1.66      anton    1223: @cindex syntax tutorial
1.48      anton    1224: 
                   1225: A @dfn{word} is a sequence of arbitrary characters (expcept white
                   1226: space).  Words are separated by white space.  E.g., each of the
                   1227: following lines contains exactly one word:
                   1228: 
                   1229: @example
                   1230: word
                   1231: !@@#$%^&*()
                   1232: 1234567890
                   1233: 5!a
                   1234: @end example
                   1235: 
                   1236: A frequent beginner's error is to leave away necessary white space,
                   1237: resulting in an error like @samp{Undefined word}; so if you see such an
                   1238: error, check if you have put spaces wherever necessary.
                   1239: 
                   1240: @example
                   1241: ." hello, world" \ correct
                   1242: ."hello, world"  \ gives an "Undefined word" error
                   1243: @end example
                   1244: 
1.65      anton    1245: Gforth and most other Forth systems ignore differences in case (they are
1.48      anton    1246: case-insensitive), i.e., @samp{word} is the same as @samp{Word}.  If
                   1247: your system is case-sensitive, you may have to type all the examples
                   1248: given here in upper case.
                   1249: 
                   1250: 
                   1251: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
                   1252: @section Crash Course
                   1253: 
                   1254: Type
                   1255: 
                   1256: @example
                   1257: 0 0 !
                   1258: here execute
                   1259: ' catch >body 20 erase abort
                   1260: ' (quit) >body 20 erase
                   1261: @end example
                   1262: 
                   1263: The last two examples are guaranteed to destroy parts of Gforth (and
                   1264: most other systems), so you better leave Gforth afterwards (if it has
                   1265: not finished by itself).  On some systems you may have to kill gforth
                   1266: from outside (e.g., in Unix with @code{kill}).
                   1267: 
                   1268: Now that you know how to produce crashes (and that there's not much to
                   1269: them), let's learn how to produce meaningful programs.
                   1270: 
                   1271: 
                   1272: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
                   1273: @section Stack
1.66      anton    1274: @cindex stack tutorial
1.48      anton    1275: 
                   1276: The most obvious feature of Forth is the stack.  When you type in a
                   1277: number, it is pushed on the stack.  You can display the content of the
                   1278: stack with @code{.s}.
                   1279: 
                   1280: @example
                   1281: 1 2 .s
                   1282: 3 .s
                   1283: @end example
                   1284: 
                   1285: @code{.s} displays the top-of-stack to the right, i.e., the numbers
                   1286: appear in @code{.s} output as they appeared in the input.
                   1287: 
                   1288: You can print the top of stack element with @code{.}.
                   1289: 
                   1290: @example
                   1291: 1 2 3 . . .
                   1292: @end example
                   1293: 
                   1294: In general, words consume their stack arguments (@code{.s} is an
                   1295: exception).
                   1296: 
1.141     anton    1297: @quotation Assignment
1.48      anton    1298: What does the stack contain after @code{5 6 7 .}?
1.141     anton    1299: @end quotation
1.48      anton    1300: 
                   1301: 
                   1302: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
                   1303: @section Arithmetics
1.66      anton    1304: @cindex arithmetics tutorial
1.48      anton    1305: 
                   1306: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
                   1307: operate on the top two stack items:
                   1308: 
                   1309: @example
1.67      anton    1310: 2 2 .s
                   1311: + .s
                   1312: .
1.48      anton    1313: 2 1 - .
                   1314: 7 3 mod .
                   1315: @end example
                   1316: 
                   1317: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
                   1318: as in the corresponding infix expression (this is generally the case in
                   1319: Forth).
                   1320: 
                   1321: Parentheses are superfluous (and not available), because the order of
                   1322: the words unambiguously determines the order of evaluation and the
                   1323: operands:
                   1324: 
                   1325: @example
                   1326: 3 4 + 5 * .
                   1327: 3 4 5 * + .
                   1328: @end example
                   1329: 
1.141     anton    1330: @quotation Assignment
1.48      anton    1331: What are the infix expressions corresponding to the Forth code above?
                   1332: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
                   1333: known as Postfix or RPN (Reverse Polish Notation).}.
1.141     anton    1334: @end quotation
1.48      anton    1335: 
                   1336: To change the sign, use @code{negate}:
                   1337: 
                   1338: @example
                   1339: 2 negate .
                   1340: @end example
                   1341: 
1.141     anton    1342: @quotation Assignment
1.48      anton    1343: Convert -(-3)*4-5 to Forth.
1.141     anton    1344: @end quotation
1.48      anton    1345: 
                   1346: @code{/mod} performs both @code{/} and @code{mod}.
                   1347: 
                   1348: @example
                   1349: 7 3 /mod . .
                   1350: @end example
                   1351: 
1.66      anton    1352: Reference: @ref{Arithmetic}.
                   1353: 
                   1354: 
1.48      anton    1355: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
                   1356: @section Stack Manipulation
1.66      anton    1357: @cindex stack manipulation tutorial
1.48      anton    1358: 
                   1359: Stack manipulation words rearrange the data on the stack.
                   1360: 
                   1361: @example
                   1362: 1 .s drop .s
                   1363: 1 .s dup .s drop drop .s
                   1364: 1 2 .s over .s drop drop drop
                   1365: 1 2 .s swap .s drop drop
                   1366: 1 2 3 .s rot .s drop drop drop
                   1367: @end example
                   1368: 
                   1369: These are the most important stack manipulation words.  There are also
                   1370: variants that manipulate twice as many stack items:
                   1371: 
                   1372: @example
                   1373: 1 2 3 4 .s 2swap .s 2drop 2drop
                   1374: @end example
                   1375: 
                   1376: Two more stack manipulation words are:
                   1377: 
                   1378: @example
                   1379: 1 2 .s nip .s drop
                   1380: 1 2 .s tuck .s 2drop drop
                   1381: @end example
                   1382: 
1.141     anton    1383: @quotation Assignment
1.48      anton    1384: Replace @code{nip} and @code{tuck} with combinations of other stack
                   1385: manipulation words.
                   1386: 
                   1387: @example
                   1388: Given:          How do you get:
                   1389: 1 2 3           3 2 1           
                   1390: 1 2 3           1 2 3 2                 
                   1391: 1 2 3           1 2 3 3                 
                   1392: 1 2 3           1 3 3           
                   1393: 1 2 3           2 1 3           
                   1394: 1 2 3 4         4 3 2 1         
                   1395: 1 2 3           1 2 3 1 2 3             
                   1396: 1 2 3 4         1 2 3 4 1 2             
                   1397: 1 2 3
                   1398: 1 2 3           1 2 3 4                 
                   1399: 1 2 3           1 3             
                   1400: @end example
1.141     anton    1401: @end quotation
1.48      anton    1402: 
                   1403: @example
                   1404: 5 dup * .
                   1405: @end example
                   1406: 
1.141     anton    1407: @quotation Assignment
1.48      anton    1408: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
                   1409: Write a piece of Forth code that expects two numbers on the stack
                   1410: (@var{a} and @var{b}, with @var{b} on top) and computes
                   1411: @code{(a-b)(a+1)}.
1.141     anton    1412: @end quotation
1.48      anton    1413: 
1.66      anton    1414: Reference: @ref{Stack Manipulation}.
                   1415: 
                   1416: 
1.48      anton    1417: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
                   1418: @section Using files for Forth code
1.66      anton    1419: @cindex loading Forth code, tutorial
                   1420: @cindex files containing Forth code, tutorial
1.48      anton    1421: 
                   1422: While working at the Forth command line is convenient for one-line
                   1423: examples and short one-off code, you probably want to store your source
                   1424: code in files for convenient editing and persistence.  You can use your
                   1425: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1.102     anton    1426: Gforth}) to create @var{file.fs} and use
1.48      anton    1427: 
                   1428: @example
1.102     anton    1429: s" @var{file.fs}" included
1.48      anton    1430: @end example
                   1431: 
                   1432: to load it into your Forth system.  The file name extension I use for
                   1433: Forth files is @samp{.fs}.
                   1434: 
                   1435: You can easily start Gforth with some files loaded like this:
                   1436: 
                   1437: @example
1.102     anton    1438: gforth @var{file1.fs} @var{file2.fs}
1.48      anton    1439: @end example
                   1440: 
                   1441: If an error occurs during loading these files, Gforth terminates,
                   1442: whereas an error during @code{INCLUDED} within Gforth usually gives you
                   1443: a Gforth command line.  Starting the Forth system every time gives you a
                   1444: clean start every time, without interference from the results of earlier
                   1445: tries.
                   1446: 
                   1447: I often put all the tests in a file, then load the code and run the
                   1448: tests with
                   1449: 
                   1450: @example
1.102     anton    1451: gforth @var{code.fs} @var{tests.fs} -e bye
1.48      anton    1452: @end example
                   1453: 
                   1454: (often by performing this command with @kbd{C-x C-e} in Emacs).  The
                   1455: @code{-e bye} ensures that Gforth terminates afterwards so that I can
                   1456: restart this command without ado.
                   1457: 
                   1458: The advantage of this approach is that the tests can be repeated easily
                   1459: every time the program ist changed, making it easy to catch bugs
                   1460: introduced by the change.
                   1461: 
1.66      anton    1462: Reference: @ref{Forth source files}.
                   1463: 
1.48      anton    1464: 
                   1465: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
                   1466: @section Comments
1.66      anton    1467: @cindex comments tutorial
1.48      anton    1468: 
                   1469: @example
                   1470: \ That's a comment; it ends at the end of the line
                   1471: ( Another comment; it ends here: )  .s
                   1472: @end example
                   1473: 
                   1474: @code{\} and @code{(} are ordinary Forth words and therefore have to be
                   1475: separated with white space from the following text.
                   1476: 
                   1477: @example
                   1478: \This gives an "Undefined word" error
                   1479: @end example
                   1480: 
                   1481: The first @code{)} ends a comment started with @code{(}, so you cannot
                   1482: nest @code{(}-comments; and you cannot comment out text containing a
                   1483: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
                   1484: avoid @code{)} in word names.}.
                   1485: 
                   1486: I use @code{\}-comments for descriptive text and for commenting out code
                   1487: of one or more line; I use @code{(}-comments for describing the stack
                   1488: effect, the stack contents, or for commenting out sub-line pieces of
                   1489: code.
                   1490: 
                   1491: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
                   1492: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
                   1493: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
                   1494: with @kbd{M-q}.
                   1495: 
1.66      anton    1496: Reference: @ref{Comments}.
                   1497: 
1.48      anton    1498: 
                   1499: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
                   1500: @section Colon Definitions
1.66      anton    1501: @cindex colon definitions, tutorial
                   1502: @cindex definitions, tutorial
                   1503: @cindex procedures, tutorial
                   1504: @cindex functions, tutorial
1.48      anton    1505: 
                   1506: are similar to procedures and functions in other programming languages.
                   1507: 
                   1508: @example
                   1509: : squared ( n -- n^2 )
                   1510:    dup * ;
                   1511: 5 squared .
                   1512: 7 squared .
                   1513: @end example
                   1514: 
                   1515: @code{:} starts the colon definition; its name is @code{squared}.  The
                   1516: following comment describes its stack effect.  The words @code{dup *}
                   1517: are not executed, but compiled into the definition.  @code{;} ends the
                   1518: colon definition.
                   1519: 
                   1520: The newly-defined word can be used like any other word, including using
                   1521: it in other definitions:
                   1522: 
                   1523: @example
                   1524: : cubed ( n -- n^3 )
                   1525:    dup squared * ;
                   1526: -5 cubed .
                   1527: : fourth-power ( n -- n^4 )
                   1528:    squared squared ;
                   1529: 3 fourth-power .
                   1530: @end example
                   1531: 
1.141     anton    1532: @quotation Assignment
1.48      anton    1533: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
                   1534: @code{/mod} in terms of other Forth words, and check if they work (hint:
                   1535: test your tests on the originals first).  Don't let the
                   1536: @samp{redefined}-Messages spook you, they are just warnings.
1.141     anton    1537: @end quotation
1.48      anton    1538: 
1.66      anton    1539: Reference: @ref{Colon Definitions}.
                   1540: 
1.48      anton    1541: 
                   1542: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
                   1543: @section Decompilation
1.66      anton    1544: @cindex decompilation tutorial
                   1545: @cindex see tutorial
1.48      anton    1546: 
                   1547: You can decompile colon definitions with @code{see}:
                   1548: 
                   1549: @example
                   1550: see squared
                   1551: see cubed
                   1552: @end example
                   1553: 
                   1554: In Gforth @code{see} shows you a reconstruction of the source code from
                   1555: the executable code.  Informations that were present in the source, but
                   1556: not in the executable code, are lost (e.g., comments).
                   1557: 
1.65      anton    1558: You can also decompile the predefined words:
                   1559: 
                   1560: @example
                   1561: see .
                   1562: see +
                   1563: @end example
                   1564: 
                   1565: 
1.48      anton    1566: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
                   1567: @section Stack-Effect Comments
1.66      anton    1568: @cindex stack-effect comments, tutorial
                   1569: @cindex --, tutorial
1.48      anton    1570: By convention the comment after the name of a definition describes the
                   1571: stack effect: The part in from of the @samp{--} describes the state of
                   1572: the stack before the execution of the definition, i.e., the parameters
                   1573: that are passed into the colon definition; the part behind the @samp{--}
                   1574: is the state of the stack after the execution of the definition, i.e.,
                   1575: the results of the definition.  The stack comment only shows the top
                   1576: stack items that the definition accesses and/or changes.
                   1577: 
                   1578: You should put a correct stack effect on every definition, even if it is
                   1579: just @code{( -- )}.  You should also add some descriptive comment to
                   1580: more complicated words (I usually do this in the lines following
                   1581: @code{:}).  If you don't do this, your code becomes unreadable (because
1.117     anton    1582: you have to work through every definition before you can understand
1.48      anton    1583: any).
                   1584: 
1.141     anton    1585: @quotation Assignment
1.48      anton    1586: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
                   1587: x2 x1}.  Describe the stack effect of @code{-}, @code{drop}, @code{dup},
                   1588: @code{over}, @code{rot}, @code{nip}, and @code{tuck}.  Hint: When you
1.65      anton    1589: are done, you can compare your stack effects to those in this manual
1.48      anton    1590: (@pxref{Word Index}).
1.141     anton    1591: @end quotation
1.48      anton    1592: 
                   1593: Sometimes programmers put comments at various places in colon
                   1594: definitions that describe the contents of the stack at that place (stack
                   1595: comments); i.e., they are like the first part of a stack-effect
                   1596: comment. E.g.,
                   1597: 
                   1598: @example
                   1599: : cubed ( n -- n^3 )
                   1600:    dup squared  ( n n^2 ) * ;
                   1601: @end example
                   1602: 
                   1603: In this case the stack comment is pretty superfluous, because the word
                   1604: is simple enough.  If you think it would be a good idea to add such a
                   1605: comment to increase readability, you should also consider factoring the
                   1606: word into several simpler words (@pxref{Factoring Tutorial,,
1.60      anton    1607: Factoring}), which typically eliminates the need for the stack comment;
1.48      anton    1608: however, if you decide not to refactor it, then having such a comment is
                   1609: better than not having it.
                   1610: 
                   1611: The names of the stack items in stack-effect and stack comments in the
                   1612: standard, in this manual, and in many programs specify the type through
                   1613: a type prefix, similar to Fortran and Hungarian notation.  The most
                   1614: frequent prefixes are:
                   1615: 
                   1616: @table @code
                   1617: @item n
                   1618: signed integer
                   1619: @item u
                   1620: unsigned integer
                   1621: @item c
                   1622: character
                   1623: @item f
                   1624: Boolean flags, i.e. @code{false} or @code{true}.
                   1625: @item a-addr,a-
                   1626: Cell-aligned address
                   1627: @item c-addr,c-
                   1628: Char-aligned address (note that a Char may have two bytes in Windows NT)
                   1629: @item xt
                   1630: Execution token, same size as Cell
                   1631: @item w,x
                   1632: Cell, can contain an integer or an address.  It usually takes 32, 64 or
                   1633: 16 bits (depending on your platform and Forth system). A cell is more
                   1634: commonly known as machine word, but the term @emph{word} already means
                   1635: something different in Forth.
                   1636: @item d
                   1637: signed double-cell integer
                   1638: @item ud
                   1639: unsigned double-cell integer
                   1640: @item r
                   1641: Float (on the FP stack)
                   1642: @end table
                   1643: 
                   1644: You can find a more complete list in @ref{Notation}.
                   1645: 
1.141     anton    1646: @quotation Assignment
1.48      anton    1647: Write stack-effect comments for all definitions you have written up to
                   1648: now.
1.141     anton    1649: @end quotation
1.48      anton    1650: 
                   1651: 
                   1652: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
                   1653: @section Types
1.66      anton    1654: @cindex types tutorial
1.48      anton    1655: 
                   1656: In Forth the names of the operations are not overloaded; so similar
                   1657: operations on different types need different names; e.g., @code{+} adds
                   1658: integers, and you have to use @code{f+} to add floating-point numbers.
                   1659: The following prefixes are often used for related operations on
                   1660: different types:
                   1661: 
                   1662: @table @code
                   1663: @item (none)
                   1664: signed integer
                   1665: @item u
                   1666: unsigned integer
                   1667: @item c
                   1668: character
                   1669: @item d
                   1670: signed double-cell integer
                   1671: @item ud, du
                   1672: unsigned double-cell integer
                   1673: @item 2
                   1674: two cells (not-necessarily double-cell numbers)
                   1675: @item m, um
                   1676: mixed single-cell and double-cell operations
                   1677: @item f
                   1678: floating-point (note that in stack comments @samp{f} represents flags,
1.66      anton    1679: and @samp{r} represents FP numbers).
1.48      anton    1680: @end table
                   1681: 
                   1682: If there are no differences between the signed and the unsigned variant
                   1683: (e.g., for @code{+}), there is only the prefix-less variant.
                   1684: 
                   1685: Forth does not perform type checking, neither at compile time, nor at
                   1686: run time.  If you use the wrong oeration, the data are interpreted
                   1687: incorrectly:
                   1688: 
                   1689: @example
                   1690: -1 u.
                   1691: @end example
                   1692: 
                   1693: If you have only experience with type-checked languages until now, and
                   1694: have heard how important type-checking is, don't panic!  In my
                   1695: experience (and that of other Forthers), type errors in Forth code are
                   1696: usually easy to find (once you get used to it), the increased vigilance
                   1697: of the programmer tends to catch some harder errors in addition to most
                   1698: type errors, and you never have to work around the type system, so in
                   1699: most situations the lack of type-checking seems to be a win (projects to
                   1700: add type checking to Forth have not caught on).
                   1701: 
                   1702: 
                   1703: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
                   1704: @section Factoring
1.66      anton    1705: @cindex factoring tutorial
1.48      anton    1706: 
                   1707: If you try to write longer definitions, you will soon find it hard to
                   1708: keep track of the stack contents.  Therefore, good Forth programmers
                   1709: tend to write only short definitions (e.g., three lines).  The art of
                   1710: finding meaningful short definitions is known as factoring (as in
                   1711: factoring polynomials).
                   1712: 
                   1713: Well-factored programs offer additional advantages: smaller, more
                   1714: general words, are easier to test and debug and can be reused more and
                   1715: better than larger, specialized words.
                   1716: 
                   1717: So, if you run into difficulties with stack management, when writing
                   1718: code, try to define meaningful factors for the word, and define the word
                   1719: in terms of those.  Even if a factor contains only two words, it is
                   1720: often helpful.
                   1721: 
1.65      anton    1722: Good factoring is not easy, and it takes some practice to get the knack
                   1723: for it; but even experienced Forth programmers often don't find the
                   1724: right solution right away, but only when rewriting the program.  So, if
                   1725: you don't come up with a good solution immediately, keep trying, don't
                   1726: despair.
1.48      anton    1727: 
                   1728: @c example !!
                   1729: 
                   1730: 
                   1731: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
                   1732: @section Designing the stack effect
1.66      anton    1733: @cindex Stack effect design, tutorial
                   1734: @cindex design of stack effects, tutorial
1.48      anton    1735: 
                   1736: In other languages you can use an arbitrary order of parameters for a
1.65      anton    1737: function; and since there is only one result, you don't have to deal with
1.48      anton    1738: the order of results, either.
                   1739: 
1.117     anton    1740: In Forth (and other stack-based languages, e.g., PostScript) the
1.48      anton    1741: parameter and result order of a definition is important and should be
                   1742: designed well.  The general guideline is to design the stack effect such
                   1743: that the word is simple to use in most cases, even if that complicates
                   1744: the implementation of the word.  Some concrete rules are:
                   1745: 
                   1746: @itemize @bullet
                   1747: 
                   1748: @item
                   1749: Words consume all of their parameters (e.g., @code{.}).
                   1750: 
                   1751: @item
                   1752: If there is a convention on the order of parameters (e.g., from
                   1753: mathematics or another programming language), stick with it (e.g.,
                   1754: @code{-}).
                   1755: 
                   1756: @item
                   1757: If one parameter usually requires only a short computation (e.g., it is
                   1758: a constant), pass it on the top of the stack.  Conversely, parameters
                   1759: that usually require a long sequence of code to compute should be passed
                   1760: as the bottom (i.e., first) parameter.  This makes the code easier to
                   1761: read, because reader does not need to keep track of the bottom item
                   1762: through a long sequence of code (or, alternatively, through stack
1.49      anton    1763: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1.48      anton    1764: address on top of the stack because it is usually simpler to compute
                   1765: than the stored value (often the address is just a variable).
                   1766: 
                   1767: @item
                   1768: Similarly, results that are usually consumed quickly should be returned
                   1769: on the top of stack, whereas a result that is often used in long
                   1770: computations should be passed as bottom result.  E.g., the file words
                   1771: like @code{open-file} return the error code on the top of stack, because
                   1772: it is usually consumed quickly by @code{throw}; moreover, the error code
                   1773: has to be checked before doing anything with the other results.
                   1774: 
                   1775: @end itemize
                   1776: 
                   1777: These rules are just general guidelines, don't lose sight of the overall
                   1778: goal to make the words easy to use.  E.g., if the convention rule
                   1779: conflicts with the computation-length rule, you might decide in favour
                   1780: of the convention if the word will be used rarely, and in favour of the
                   1781: computation-length rule if the word will be used frequently (because
                   1782: with frequent use the cost of breaking the computation-length rule would
                   1783: be quite high, and frequent use makes it easier to remember an
                   1784: unconventional order).
                   1785: 
                   1786: @c example !! structure package
                   1787: 
1.65      anton    1788: 
1.48      anton    1789: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
                   1790: @section Local Variables
1.66      anton    1791: @cindex local variables, tutorial
1.48      anton    1792: 
                   1793: You can define local variables (@emph{locals}) in a colon definition:
                   1794: 
                   1795: @example
                   1796: : swap @{ a b -- b a @}
                   1797:   b a ;
                   1798: 1 2 swap .s 2drop
                   1799: @end example
                   1800: 
                   1801: (If your Forth system does not support this syntax, include
                   1802: @file{compat/anslocals.fs} first).
                   1803: 
                   1804: In this example @code{@{ a b -- b a @}} is the locals definition; it
                   1805: takes two cells from the stack, puts the top of stack in @code{b} and
                   1806: the next stack element in @code{a}.  @code{--} starts a comment ending
                   1807: with @code{@}}.  After the locals definition, using the name of the
                   1808: local will push its value on the stack.  You can leave the comment
                   1809: part (@code{-- b a}) away:
                   1810: 
                   1811: @example
                   1812: : swap ( x1 x2 -- x2 x1 )
                   1813:   @{ a b @} b a ;
                   1814: @end example
                   1815: 
                   1816: In Gforth you can have several locals definitions, anywhere in a colon
                   1817: definition; in contrast, in a standard program you can have only one
                   1818: locals definition per colon definition, and that locals definition must
                   1819: be outside any controll structure.
                   1820: 
                   1821: With locals you can write slightly longer definitions without running
                   1822: into stack trouble.  However, I recommend trying to write colon
                   1823: definitions without locals for exercise purposes to help you gain the
                   1824: essential factoring skills.
                   1825: 
1.141     anton    1826: @quotation Assignment
1.48      anton    1827: Rewrite your definitions until now with locals
1.141     anton    1828: @end quotation
1.48      anton    1829: 
1.66      anton    1830: Reference: @ref{Locals}.
                   1831: 
1.48      anton    1832: 
                   1833: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
                   1834: @section Conditional execution
1.66      anton    1835: @cindex conditionals, tutorial
                   1836: @cindex if, tutorial
1.48      anton    1837: 
                   1838: In Forth you can use control structures only inside colon definitions.
                   1839: An @code{if}-structure looks like this:
                   1840: 
                   1841: @example
                   1842: : abs ( n1 -- +n2 )
                   1843:     dup 0 < if
                   1844:         negate
                   1845:     endif ;
                   1846: 5 abs .
                   1847: -5 abs .
                   1848: @end example
                   1849: 
                   1850: @code{if} takes a flag from the stack.  If the flag is non-zero (true),
                   1851: the following code is performed, otherwise execution continues after the
1.51      pazsan   1852: @code{endif} (or @code{else}).  @code{<} compares the top two stack
1.48      anton    1853: elements and prioduces a flag:
                   1854: 
                   1855: @example
                   1856: 1 2 < .
                   1857: 2 1 < .
                   1858: 1 1 < .
                   1859: @end example
                   1860: 
                   1861: Actually the standard name for @code{endif} is @code{then}.  This
                   1862: tutorial presents the examples using @code{endif}, because this is often
                   1863: less confusing for people familiar with other programming languages
                   1864: where @code{then} has a different meaning.  If your system does not have
                   1865: @code{endif}, define it with
                   1866: 
                   1867: @example
                   1868: : endif postpone then ; immediate
                   1869: @end example
                   1870: 
                   1871: You can optionally use an @code{else}-part:
                   1872: 
                   1873: @example
                   1874: : min ( n1 n2 -- n )
                   1875:   2dup < if
                   1876:     drop
                   1877:   else
                   1878:     nip
                   1879:   endif ;
                   1880: 2 3 min .
                   1881: 3 2 min .
                   1882: @end example
                   1883: 
1.141     anton    1884: @quotation Assignment
1.48      anton    1885: Write @code{min} without @code{else}-part (hint: what's the definition
                   1886: of @code{nip}?).
1.141     anton    1887: @end quotation
1.48      anton    1888: 
1.66      anton    1889: Reference: @ref{Selection}.
                   1890: 
1.48      anton    1891: 
                   1892: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
                   1893: @section Flags and Comparisons
1.66      anton    1894: @cindex flags tutorial
                   1895: @cindex comparison tutorial
1.48      anton    1896: 
                   1897: In a false-flag all bits are clear (0 when interpreted as integer).  In
                   1898: a canonical true-flag all bits are set (-1 as a twos-complement signed
                   1899: integer); in many contexts (e.g., @code{if}) any non-zero value is
                   1900: treated as true flag.
                   1901: 
                   1902: @example
                   1903: false .
                   1904: true .
                   1905: true hex u. decimal
                   1906: @end example
                   1907: 
                   1908: Comparison words produce canonical flags:
                   1909: 
                   1910: @example
                   1911: 1 1 = .
                   1912: 1 0= .
                   1913: 0 1 < .
                   1914: 0 0 < .
                   1915: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
                   1916: -1 1 < .
                   1917: @end example
                   1918: 
1.66      anton    1919: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
                   1920: (or none) and the comparisons @code{= <> < > <= >=}.  Only a part of
                   1921: these combinations are standard (for details see the standard,
                   1922: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
1.48      anton    1923: 
                   1924: You can use @code{and or xor invert} can be used as operations on
                   1925: canonical flags.  Actually they are bitwise operations:
                   1926: 
                   1927: @example
                   1928: 1 2 and .
                   1929: 1 2 or .
                   1930: 1 3 xor .
                   1931: 1 invert .
                   1932: @end example
                   1933: 
                   1934: You can convert a zero/non-zero flag into a canonical flag with
                   1935: @code{0<>} (and complement it on the way with @code{0=}).
                   1936: 
                   1937: @example
                   1938: 1 0= .
                   1939: 1 0<> .
                   1940: @end example
                   1941: 
1.65      anton    1942: You can use the all-bits-set feature of canonical flags and the bitwise
1.48      anton    1943: operation of the Boolean operations to avoid @code{if}s:
                   1944: 
                   1945: @example
                   1946: : foo ( n1 -- n2 )
                   1947:   0= if
                   1948:     14
                   1949:   else
                   1950:     0
                   1951:   endif ;
                   1952: 0 foo .
                   1953: 1 foo .
                   1954: 
                   1955: : foo ( n1 -- n2 )
                   1956:   0= 14 and ;
                   1957: 0 foo .
                   1958: 1 foo .
                   1959: @end example
                   1960: 
1.141     anton    1961: @quotation Assignment
1.48      anton    1962: Write @code{min} without @code{if}.
1.141     anton    1963: @end quotation
1.48      anton    1964: 
1.66      anton    1965: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
                   1966: @ref{Bitwise operations}.
                   1967: 
1.48      anton    1968: 
                   1969: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
                   1970: @section General Loops
1.66      anton    1971: @cindex loops, indefinite, tutorial
1.48      anton    1972: 
                   1973: The endless loop is the most simple one:
                   1974: 
                   1975: @example
                   1976: : endless ( -- )
                   1977:   0 begin
                   1978:     dup . 1+
                   1979:   again ;
                   1980: endless
                   1981: @end example
                   1982: 
                   1983: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth).  @code{begin}
                   1984: does nothing at run-time, @code{again} jumps back to @code{begin}.
                   1985: 
                   1986: A loop with one exit at any place looks like this:
                   1987: 
                   1988: @example
                   1989: : log2 ( +n1 -- n2 )
                   1990: \ logarithmus dualis of n1>0, rounded down to the next integer
                   1991:   assert( dup 0> )
                   1992:   2/ 0 begin
                   1993:     over 0> while
                   1994:       1+ swap 2/ swap
                   1995:   repeat
                   1996:   nip ;
                   1997: 7 log2 .
                   1998: 8 log2 .
                   1999: @end example
                   2000: 
                   2001: At run-time @code{while} consumes a flag; if it is 0, execution
1.51      pazsan   2002: continues behind the @code{repeat}; if the flag is non-zero, execution
1.48      anton    2003: continues behind the @code{while}.  @code{Repeat} jumps back to
                   2004: @code{begin}, just like @code{again}.
                   2005: 
                   2006: In Forth there are many combinations/abbreviations, like @code{1+}.
1.90      anton    2007: However, @code{2/} is not one of them; it shifts its argument right by
1.48      anton    2008: one bit (arithmetic shift right):
                   2009: 
                   2010: @example
                   2011: -5 2 / .
                   2012: -5 2/ .
                   2013: @end example
                   2014: 
                   2015: @code{assert(} is no standard word, but you can get it on systems other
                   2016: then Gforth by including @file{compat/assert.fs}.  You can see what it
                   2017: does by trying
                   2018: 
                   2019: @example
                   2020: 0 log2 .
                   2021: @end example
                   2022: 
                   2023: Here's a loop with an exit at the end:
                   2024: 
                   2025: @example
                   2026: : log2 ( +n1 -- n2 )
                   2027: \ logarithmus dualis of n1>0, rounded down to the next integer
                   2028:   assert( dup 0 > )
                   2029:   -1 begin
                   2030:     1+ swap 2/ swap
                   2031:     over 0 <=
                   2032:   until
                   2033:   nip ;
                   2034: @end example
                   2035: 
                   2036: @code{Until} consumes a flag; if it is non-zero, execution continues at
                   2037: the @code{begin}, otherwise after the @code{until}.
                   2038: 
1.141     anton    2039: @quotation Assignment
1.48      anton    2040: Write a definition for computing the greatest common divisor.
1.141     anton    2041: @end quotation
1.48      anton    2042: 
1.66      anton    2043: Reference: @ref{Simple Loops}.
                   2044: 
1.48      anton    2045: 
                   2046: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
                   2047: @section Counted loops
1.66      anton    2048: @cindex loops, counted, tutorial
1.48      anton    2049: 
                   2050: @example
                   2051: : ^ ( n1 u -- n )
                   2052: \ n = the uth power of u1
                   2053:   1 swap 0 u+do
                   2054:     over *
                   2055:   loop
                   2056:   nip ;
                   2057: 3 2 ^ .
                   2058: 4 3 ^ .
                   2059: @end example
                   2060: 
                   2061: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
                   2062: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
                   2063: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
                   2064: times (or not at all, if @code{u3-u4<0}).
                   2065: 
                   2066: You can see the stack effect design rules at work in the stack effect of
                   2067: the loop start words: Since the start value of the loop is more
                   2068: frequently constant than the end value, the start value is passed on
                   2069: the top-of-stack.
                   2070: 
                   2071: You can access the counter of a counted loop with @code{i}:
                   2072: 
                   2073: @example
                   2074: : fac ( u -- u! )
                   2075:   1 swap 1+ 1 u+do
                   2076:     i *
                   2077:   loop ;
                   2078: 5 fac .
                   2079: 7 fac .
                   2080: @end example
                   2081: 
                   2082: There is also @code{+do}, which expects signed numbers (important for
                   2083: deciding whether to enter the loop).
                   2084: 
1.141     anton    2085: @quotation Assignment
1.48      anton    2086: Write a definition for computing the nth Fibonacci number.
1.141     anton    2087: @end quotation
1.48      anton    2088: 
1.65      anton    2089: You can also use increments other than 1:
                   2090: 
                   2091: @example
                   2092: : up2 ( n1 n2 -- )
                   2093:   +do
                   2094:     i .
                   2095:   2 +loop ;
                   2096: 10 0 up2
                   2097: 
                   2098: : down2 ( n1 n2 -- )
                   2099:   -do
                   2100:     i .
                   2101:   2 -loop ;
                   2102: 0 10 down2
                   2103: @end example
1.48      anton    2104: 
1.66      anton    2105: Reference: @ref{Counted Loops}.
                   2106: 
1.48      anton    2107: 
                   2108: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
                   2109: @section Recursion
1.66      anton    2110: @cindex recursion tutorial
1.48      anton    2111: 
                   2112: Usually the name of a definition is not visible in the definition; but
                   2113: earlier definitions are usually visible:
                   2114: 
                   2115: @example
                   2116: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
                   2117: : / ( n1 n2 -- n )
                   2118:   dup 0= if
                   2119:     -10 throw \ report division by zero
                   2120:   endif
                   2121:   /           \ old version
                   2122: ;
                   2123: 1 0 /
                   2124: @end example
                   2125: 
                   2126: For recursive definitions you can use @code{recursive} (non-standard) or
                   2127: @code{recurse}:
                   2128: 
                   2129: @example
                   2130: : fac1 ( n -- n! ) recursive
                   2131:  dup 0> if
                   2132:    dup 1- fac1 *
                   2133:  else
                   2134:    drop 1
                   2135:  endif ;
                   2136: 7 fac1 .
                   2137: 
                   2138: : fac2 ( n -- n! )
                   2139:  dup 0> if
                   2140:    dup 1- recurse *
                   2141:  else
                   2142:    drop 1
                   2143:  endif ;
                   2144: 8 fac2 .
                   2145: @end example
                   2146: 
1.141     anton    2147: @quotation Assignment
1.48      anton    2148: Write a recursive definition for computing the nth Fibonacci number.
1.141     anton    2149: @end quotation
1.48      anton    2150: 
1.66      anton    2151: Reference (including indirect recursion): @xref{Calls and returns}.
                   2152: 
1.48      anton    2153: 
                   2154: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
                   2155: @section Leaving definitions or loops
1.66      anton    2156: @cindex leaving definitions, tutorial
                   2157: @cindex leaving loops, tutorial
1.48      anton    2158: 
                   2159: @code{EXIT} exits the current definition right away.  For every counted
                   2160: loop that is left in this way, an @code{UNLOOP} has to be performed
                   2161: before the @code{EXIT}:
                   2162: 
                   2163: @c !! real examples
                   2164: @example
                   2165: : ...
                   2166:  ... u+do
                   2167:    ... if
                   2168:      ... unloop exit
                   2169:    endif
                   2170:    ...
                   2171:  loop
                   2172:  ... ;
                   2173: @end example
                   2174: 
                   2175: @code{LEAVE} leaves the innermost counted loop right away:
                   2176: 
                   2177: @example
                   2178: : ...
                   2179:  ... u+do
                   2180:    ... if
                   2181:      ... leave
                   2182:    endif
                   2183:    ...
                   2184:  loop
                   2185:  ... ;
                   2186: @end example
                   2187: 
1.65      anton    2188: @c !! example
1.48      anton    2189: 
1.66      anton    2190: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
                   2191: 
                   2192: 
1.48      anton    2193: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
                   2194: @section Return Stack
1.66      anton    2195: @cindex return stack tutorial
1.48      anton    2196: 
                   2197: In addition to the data stack Forth also has a second stack, the return
                   2198: stack; most Forth systems store the return addresses of procedure calls
                   2199: there (thus its name).  Programmers can also use this stack:
                   2200: 
                   2201: @example
                   2202: : foo ( n1 n2 -- )
                   2203:  .s
                   2204:  >r .s
1.50      anton    2205:  r@@ .
1.48      anton    2206:  >r .s
1.50      anton    2207:  r@@ .
1.48      anton    2208:  r> .
1.50      anton    2209:  r@@ .
1.48      anton    2210:  r> . ;
                   2211: 1 2 foo
                   2212: @end example
                   2213: 
                   2214: @code{>r} takes an element from the data stack and pushes it onto the
                   2215: return stack; conversely, @code{r>} moves an elementm from the return to
                   2216: the data stack; @code{r@@} pushes a copy of the top of the return stack
                   2217: on the return stack.
                   2218: 
                   2219: Forth programmers usually use the return stack for storing data
                   2220: temporarily, if using the data stack alone would be too complex, and
                   2221: factoring and locals are not an option:
                   2222: 
                   2223: @example
                   2224: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
                   2225:  rot >r rot r> ;
                   2226: @end example
                   2227: 
                   2228: The return address of the definition and the loop control parameters of
                   2229: counted loops usually reside on the return stack, so you have to take
                   2230: all items, that you have pushed on the return stack in a colon
                   2231: definition or counted loop, from the return stack before the definition
                   2232: or loop ends.  You cannot access items that you pushed on the return
                   2233: stack outside some definition or loop within the definition of loop.
                   2234: 
                   2235: If you miscount the return stack items, this usually ends in a crash:
                   2236: 
                   2237: @example
                   2238: : crash ( n -- )
                   2239:   >r ;
                   2240: 5 crash
                   2241: @end example
                   2242: 
                   2243: You cannot mix using locals and using the return stack (according to the
                   2244: standard; Gforth has no problem).  However, they solve the same
                   2245: problems, so this shouldn't be an issue.
                   2246: 
1.141     anton    2247: @quotation Assignment
1.48      anton    2248: Can you rewrite any of the definitions you wrote until now in a better
                   2249: way using the return stack?
1.141     anton    2250: @end quotation
1.48      anton    2251: 
1.66      anton    2252: Reference: @ref{Return stack}.
                   2253: 
1.48      anton    2254: 
                   2255: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
                   2256: @section Memory
1.66      anton    2257: @cindex memory access/allocation tutorial
1.48      anton    2258: 
                   2259: You can create a global variable @code{v} with
                   2260: 
                   2261: @example
                   2262: variable v ( -- addr )
                   2263: @end example
                   2264: 
                   2265: @code{v} pushes the address of a cell in memory on the stack.  This cell
                   2266: was reserved by @code{variable}.  You can use @code{!} (store) to store
                   2267: values into this cell and @code{@@} (fetch) to load the value from the
                   2268: stack into memory:
                   2269: 
                   2270: @example
                   2271: v .
                   2272: 5 v ! .s
1.50      anton    2273: v @@ .
1.48      anton    2274: @end example
                   2275: 
1.65      anton    2276: You can see a raw dump of memory with @code{dump}:
                   2277: 
                   2278: @example
                   2279: v 1 cells .s dump
                   2280: @end example
                   2281: 
                   2282: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
                   2283: generally, address units (aus)) that @code{n1 cells} occupy.  You can
                   2284: also reserve more memory:
1.48      anton    2285: 
                   2286: @example
                   2287: create v2 20 cells allot
1.65      anton    2288: v2 20 cells dump
1.48      anton    2289: @end example
                   2290: 
1.65      anton    2291: creates a word @code{v2} and reserves 20 uninitialized cells; the
                   2292: address pushed by @code{v2} points to the start of these 20 cells.  You
                   2293: can use address arithmetic to access these cells:
1.48      anton    2294: 
                   2295: @example
                   2296: 3 v2 5 cells + !
1.65      anton    2297: v2 20 cells dump
1.48      anton    2298: @end example
                   2299: 
                   2300: You can reserve and initialize memory with @code{,}:
                   2301: 
                   2302: @example
                   2303: create v3
                   2304:   5 , 4 , 3 , 2 , 1 ,
1.50      anton    2305: v3 @@ .
                   2306: v3 cell+ @@ .
                   2307: v3 2 cells + @@ .
1.65      anton    2308: v3 5 cells dump
1.48      anton    2309: @end example
                   2310: 
1.141     anton    2311: @quotation Assignment
1.48      anton    2312: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
                   2313: @code{u} cells, with the first of these cells at @code{addr}, the next
                   2314: one at @code{addr cell+} etc.
1.141     anton    2315: @end quotation
1.48      anton    2316: 
                   2317: You can also reserve memory without creating a new word:
                   2318: 
                   2319: @example
1.60      anton    2320: here 10 cells allot .
                   2321: here .
1.48      anton    2322: @end example
                   2323: 
                   2324: @code{Here} pushes the start address of the memory area.  You should
                   2325: store it somewhere, or you will have a hard time finding the memory area
                   2326: again.
                   2327: 
                   2328: @code{Allot} manages dictionary memory.  The dictionary memory contains
                   2329: the system's data structures for words etc. on Gforth and most other
                   2330: Forth systems.  It is managed like a stack: You can free the memory that
                   2331: you have just @code{allot}ed with
                   2332: 
                   2333: @example
                   2334: -10 cells allot
1.60      anton    2335: here .
1.48      anton    2336: @end example
                   2337: 
                   2338: Note that you cannot do this if you have created a new word in the
                   2339: meantime (because then your @code{allot}ed memory is no longer on the
                   2340: top of the dictionary ``stack'').
                   2341: 
                   2342: Alternatively, you can use @code{allocate} and @code{free} which allow
                   2343: freeing memory in any order:
                   2344: 
                   2345: @example
                   2346: 10 cells allocate throw .s
                   2347: 20 cells allocate throw .s
                   2348: swap
                   2349: free throw
                   2350: free throw
                   2351: @end example
                   2352: 
                   2353: The @code{throw}s deal with errors (e.g., out of memory).
                   2354: 
1.65      anton    2355: And there is also a
                   2356: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
                   2357: garbage collector}, which eliminates the need to @code{free} memory
                   2358: explicitly.
1.48      anton    2359: 
1.66      anton    2360: Reference: @ref{Memory}.
                   2361: 
1.48      anton    2362: 
                   2363: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
                   2364: @section Characters and Strings
1.66      anton    2365: @cindex strings tutorial
                   2366: @cindex characters tutorial
1.48      anton    2367: 
                   2368: On the stack characters take up a cell, like numbers.  In memory they
                   2369: have their own size (one 8-bit byte on most systems), and therefore
                   2370: require their own words for memory access:
                   2371: 
                   2372: @example
                   2373: create v4 
                   2374:   104 c, 97 c, 108 c, 108 c, 111 c,
1.50      anton    2375: v4 4 chars + c@@ .
1.65      anton    2376: v4 5 chars dump
1.48      anton    2377: @end example
                   2378: 
                   2379: The preferred representation of strings on the stack is @code{addr
                   2380: u-count}, where @code{addr} is the address of the first character and
                   2381: @code{u-count} is the number of characters in the string.
                   2382: 
                   2383: @example
                   2384: v4 5 type
                   2385: @end example
                   2386: 
                   2387: You get a string constant with
                   2388: 
                   2389: @example
                   2390: s" hello, world" .s
                   2391: type
                   2392: @end example
                   2393: 
                   2394: Make sure you have a space between @code{s"} and the string; @code{s"}
                   2395: is a normal Forth word and must be delimited with white space (try what
                   2396: happens when you remove the space).
                   2397: 
                   2398: However, this interpretive use of @code{s"} is quite restricted: the
                   2399: string exists only until the next call of @code{s"} (some Forth systems
                   2400: keep more than one of these strings, but usually they still have a
1.62      crook    2401: limited lifetime).
1.48      anton    2402: 
                   2403: @example
                   2404: s" hello," s" world" .s
                   2405: type
                   2406: type
                   2407: @end example
                   2408: 
1.62      crook    2409: You can also use @code{s"} in a definition, and the resulting
                   2410: strings then live forever (well, for as long as the definition):
1.48      anton    2411: 
                   2412: @example
                   2413: : foo s" hello," s" world" ;
                   2414: foo .s
                   2415: type
                   2416: type
                   2417: @end example
                   2418: 
1.141     anton    2419: @quotation Assignment
1.48      anton    2420: @code{Emit ( c -- )} types @code{c} as character (not a number).
                   2421: Implement @code{type ( addr u -- )}.
1.141     anton    2422: @end quotation
1.48      anton    2423: 
1.66      anton    2424: Reference: @ref{Memory Blocks}.
                   2425: 
                   2426: 
1.84      pazsan   2427: @node Alignment Tutorial, Files Tutorial, Characters and Strings Tutorial, Tutorial
1.48      anton    2428: @section Alignment
1.66      anton    2429: @cindex alignment tutorial
                   2430: @cindex memory alignment tutorial
1.48      anton    2431: 
                   2432: On many processors cells have to be aligned in memory, if you want to
                   2433: access them with @code{@@} and @code{!} (and even if the processor does
1.62      crook    2434: not require alignment, access to aligned cells is faster).
1.48      anton    2435: 
                   2436: @code{Create} aligns @code{here} (i.e., the place where the next
                   2437: allocation will occur, and that the @code{create}d word points to).
                   2438: Likewise, the memory produced by @code{allocate} starts at an aligned
                   2439: address.  Adding a number of @code{cells} to an aligned address produces
                   2440: another aligned address.
                   2441: 
                   2442: However, address arithmetic involving @code{char+} and @code{chars} can
                   2443: create an address that is not cell-aligned.  @code{Aligned ( addr --
                   2444: a-addr )} produces the next aligned address:
                   2445: 
                   2446: @example
1.50      anton    2447: v3 char+ aligned .s @@ .
                   2448: v3 char+ .s @@ .
1.48      anton    2449: @end example
                   2450: 
                   2451: Similarly, @code{align} advances @code{here} to the next aligned
                   2452: address:
                   2453: 
                   2454: @example
                   2455: create v5 97 c,
                   2456: here .
                   2457: align here .
                   2458: 1000 ,
                   2459: @end example
                   2460: 
                   2461: Note that you should use aligned addresses even if your processor does
                   2462: not require them, if you want your program to be portable.
                   2463: 
1.66      anton    2464: Reference: @ref{Address arithmetic}.
                   2465: 
1.48      anton    2466: 
1.84      pazsan   2467: @node Files Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Alignment Tutorial, Tutorial
                   2468: @section Files
                   2469: @cindex files tutorial
                   2470: 
                   2471: This section gives a short introduction into how to use files inside
                   2472: Forth. It's broken up into five easy steps:
                   2473: 
                   2474: @enumerate 1
                   2475: @item Opened an ASCII text file for input
                   2476: @item Opened a file for output
                   2477: @item Read input file until string matched (or some other condition matched)
                   2478: @item Wrote some lines from input ( modified or not) to output
                   2479: @item Closed the files.
                   2480: @end enumerate
                   2481: 
                   2482: @subsection Open file for input
                   2483: 
                   2484: @example
                   2485: s" foo.in"  r/o open-file throw Value fd-in
                   2486: @end example
                   2487: 
                   2488: @subsection Create file for output
                   2489: 
                   2490: @example
                   2491: s" foo.out" w/o create-file throw Value fd-out
                   2492: @end example
                   2493: 
                   2494: The available file modes are r/o for read-only access, r/w for
                   2495: read-write access, and w/o for write-only access. You could open both
                   2496: files with r/w, too, if you like. All file words return error codes; for
                   2497: most applications, it's best to pass there error codes with @code{throw}
                   2498: to the outer error handler.
                   2499: 
                   2500: If you want words for opening and assigning, define them as follows:
                   2501: 
                   2502: @example
                   2503: 0 Value fd-in
                   2504: 0 Value fd-out
                   2505: : open-input ( addr u -- )  r/o open-file throw to fd-in ;
                   2506: : open-output ( addr u -- )  w/o create-file throw to fd-out ;
                   2507: @end example
                   2508: 
                   2509: Usage example:
                   2510: 
                   2511: @example
                   2512: s" foo.in" open-input
                   2513: s" foo.out" open-output
                   2514: @end example
                   2515: 
                   2516: @subsection Scan file for a particular line
                   2517: 
                   2518: @example
                   2519: 256 Constant max-line
                   2520: Create line-buffer  max-line 2 + allot
                   2521: 
                   2522: : scan-file ( addr u -- )
                   2523:   begin
                   2524:       line-buffer max-line fd-in read-line throw
                   2525:   while
                   2526:          >r 2dup line-buffer r> compare 0=
                   2527:      until
                   2528:   else
                   2529:      drop
                   2530:   then
                   2531:   2drop ;
                   2532: @end example
                   2533: 
                   2534: @code{read-line ( addr u1 fd -- u2 flag ior )} reads up to u1 bytes into
1.94      anton    2535: the buffer at addr, and returns the number of bytes read, a flag that is
                   2536: false when the end of file is reached, and an error code.
1.84      pazsan   2537: 
                   2538: @code{compare ( addr1 u1 addr2 u2 -- n )} compares two strings and
                   2539: returns zero if both strings are equal. It returns a positive number if
                   2540: the first string is lexically greater, a negative if the second string
                   2541: is lexically greater.
                   2542: 
                   2543: We haven't seen this loop here; it has two exits. Since the @code{while}
                   2544: exits with the number of bytes read on the stack, we have to clean up
                   2545: that separately; that's after the @code{else}.
                   2546: 
                   2547: Usage example:
                   2548: 
                   2549: @example
                   2550: s" The text I search is here" scan-file
                   2551: @end example
                   2552: 
                   2553: @subsection Copy input to output
                   2554: 
                   2555: @example
                   2556: : copy-file ( -- )
                   2557:   begin
                   2558:       line-buffer max-line fd-in read-line throw
                   2559:   while
                   2560:       line-buffer swap fd-out write-file throw
                   2561:   repeat ;
                   2562: @end example
                   2563: 
                   2564: @subsection Close files
                   2565: 
                   2566: @example
                   2567: fd-in close-file throw
                   2568: fd-out close-file throw
                   2569: @end example
                   2570: 
                   2571: Likewise, you can put that into definitions, too:
                   2572: 
                   2573: @example
                   2574: : close-input ( -- )  fd-in close-file throw ;
                   2575: : close-output ( -- )  fd-out close-file throw ;
                   2576: @end example
                   2577: 
1.141     anton    2578: @quotation Assignment
1.84      pazsan   2579: How could you modify @code{copy-file} so that it copies until a second line is
                   2580: matched? Can you write a program that extracts a section of a text file,
                   2581: given the line that starts and the line that terminates that section?
1.141     anton    2582: @end quotation
1.84      pazsan   2583: 
                   2584: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Files Tutorial, Tutorial
1.48      anton    2585: @section Interpretation and Compilation Semantics and Immediacy
1.66      anton    2586: @cindex semantics tutorial
                   2587: @cindex interpretation semantics tutorial
                   2588: @cindex compilation semantics tutorial
                   2589: @cindex immediate, tutorial
1.48      anton    2590: 
                   2591: When a word is compiled, it behaves differently from being interpreted.
                   2592: E.g., consider @code{+}:
                   2593: 
                   2594: @example
                   2595: 1 2 + .
                   2596: : foo + ;
                   2597: @end example
                   2598: 
                   2599: These two behaviours are known as compilation and interpretation
                   2600: semantics.  For normal words (e.g., @code{+}), the compilation semantics
                   2601: is to append the interpretation semantics to the currently defined word
                   2602: (@code{foo} in the example above).  I.e., when @code{foo} is executed
                   2603: later, the interpretation semantics of @code{+} (i.e., adding two
                   2604: numbers) will be performed.
                   2605: 
                   2606: However, there are words with non-default compilation semantics, e.g.,
                   2607: the control-flow words like @code{if}.  You can use @code{immediate} to
                   2608: change the compilation semantics of the last defined word to be equal to
                   2609: the interpretation semantics:
                   2610: 
                   2611: @example
                   2612: : [FOO] ( -- )
                   2613:  5 . ; immediate
                   2614: 
                   2615: [FOO]
                   2616: : bar ( -- )
                   2617:   [FOO] ;
                   2618: bar
                   2619: see bar
                   2620: @end example
                   2621: 
                   2622: Two conventions to mark words with non-default compilation semnatics are
                   2623: names with brackets (more frequently used) and to write them all in
                   2624: upper case (less frequently used).
                   2625: 
                   2626: In Gforth (and many other systems) you can also remove the
                   2627: interpretation semantics with @code{compile-only} (the compilation
                   2628: semantics is derived from the original interpretation semantics):
                   2629: 
                   2630: @example
                   2631: : flip ( -- )
                   2632:  6 . ; compile-only \ but not immediate
                   2633: flip
                   2634: 
                   2635: : flop ( -- )
                   2636:  flip ;
                   2637: flop
                   2638: @end example
                   2639: 
                   2640: In this example the interpretation semantics of @code{flop} is equal to
                   2641: the original interpretation semantics of @code{flip}.
                   2642: 
                   2643: The text interpreter has two states: in interpret state, it performs the
                   2644: interpretation semantics of words it encounters; in compile state, it
                   2645: performs the compilation semantics of these words.
                   2646: 
                   2647: Among other things, @code{:} switches into compile state, and @code{;}
                   2648: switches back to interpret state.  They contain the factors @code{]}
                   2649: (switch to compile state) and @code{[} (switch to interpret state), that
                   2650: do nothing but switch the state.
                   2651: 
                   2652: @example
                   2653: : xxx ( -- )
                   2654:   [ 5 . ]
                   2655: ;
                   2656: 
                   2657: xxx
                   2658: see xxx
                   2659: @end example
                   2660: 
                   2661: These brackets are also the source of the naming convention mentioned
                   2662: above.
                   2663: 
1.66      anton    2664: Reference: @ref{Interpretation and Compilation Semantics}.
                   2665: 
1.48      anton    2666: 
                   2667: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
                   2668: @section Execution Tokens
1.66      anton    2669: @cindex execution tokens tutorial
                   2670: @cindex XT tutorial
1.48      anton    2671: 
                   2672: @code{' word} gives you the execution token (XT) of a word.  The XT is a
                   2673: cell representing the interpretation semantics of a word.  You can
                   2674: execute this semantics with @code{execute}:
                   2675: 
                   2676: @example
                   2677: ' + .s
                   2678: 1 2 rot execute .
                   2679: @end example
                   2680: 
                   2681: The XT is similar to a function pointer in C.  However, parameter
                   2682: passing through the stack makes it a little more flexible:
                   2683: 
                   2684: @example
                   2685: : map-array ( ... addr u xt -- ... )
1.50      anton    2686: \ executes xt ( ... x -- ... ) for every element of the array starting
                   2687: \ at addr and containing u elements
1.48      anton    2688:   @{ xt @}
                   2689:   cells over + swap ?do
1.50      anton    2690:     i @@ xt execute
1.48      anton    2691:   1 cells +loop ;
                   2692: 
                   2693: create a 3 , 4 , 2 , -1 , 4 ,
                   2694: a 5 ' . map-array .s
                   2695: 0 a 5 ' + map-array .
                   2696: s" max-n" environment? drop .s
                   2697: a 5 ' min map-array .
                   2698: @end example
                   2699: 
                   2700: You can use map-array with the XTs of words that consume one element
                   2701: more than they produce.  In theory you can also use it with other XTs,
                   2702: but the stack effect then depends on the size of the array, which is
                   2703: hard to understand.
                   2704: 
1.51      pazsan   2705: Since XTs are cell-sized, you can store them in memory and manipulate
                   2706: them on the stack like other cells.  You can also compile the XT into a
1.48      anton    2707: word with @code{compile,}:
                   2708: 
                   2709: @example
                   2710: : foo1 ( n1 n2 -- n )
                   2711:    [ ' + compile, ] ;
                   2712: see foo
                   2713: @end example
                   2714: 
                   2715: This is non-standard, because @code{compile,} has no compilation
                   2716: semantics in the standard, but it works in good Forth systems.  For the
                   2717: broken ones, use
                   2718: 
                   2719: @example
                   2720: : [compile,] compile, ; immediate
                   2721: 
                   2722: : foo1 ( n1 n2 -- n )
                   2723:    [ ' + ] [compile,] ;
                   2724: see foo
                   2725: @end example
                   2726: 
                   2727: @code{'} is a word with default compilation semantics; it parses the
                   2728: next word when its interpretation semantics are executed, not during
                   2729: compilation:
                   2730: 
                   2731: @example
                   2732: : foo ( -- xt )
                   2733:   ' ;
                   2734: see foo
                   2735: : bar ( ... "word" -- ... )
                   2736:   ' execute ;
                   2737: see bar
1.60      anton    2738: 1 2 bar + .
1.48      anton    2739: @end example
                   2740: 
                   2741: You often want to parse a word during compilation and compile its XT so
                   2742: it will be pushed on the stack at run-time.  @code{[']} does this:
                   2743: 
                   2744: @example
                   2745: : xt-+ ( -- xt )
                   2746:   ['] + ;
                   2747: see xt-+
                   2748: 1 2 xt-+ execute .
                   2749: @end example
                   2750: 
                   2751: Many programmers tend to see @code{'} and the word it parses as one
                   2752: unit, and expect it to behave like @code{[']} when compiled, and are
                   2753: confused by the actual behaviour.  If you are, just remember that the
                   2754: Forth system just takes @code{'} as one unit and has no idea that it is
                   2755: a parsing word (attempts to convenience programmers in this issue have
                   2756: usually resulted in even worse pitfalls, see
1.66      anton    2757: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
                   2758: @code{State}-smartness---Why it is evil and How to Exorcise it}).
1.48      anton    2759: 
                   2760: Note that the state of the interpreter does not come into play when
1.51      pazsan   2761: creating and executing XTs.  I.e., even when you execute @code{'} in
1.48      anton    2762: compile state, it still gives you the interpretation semantics.  And
                   2763: whatever that state is, @code{execute} performs the semantics
1.66      anton    2764: represented by the XT (i.e., for XTs produced with @code{'} the
                   2765: interpretation semantics).
                   2766: 
                   2767: Reference: @ref{Tokens for Words}.
1.48      anton    2768: 
                   2769: 
                   2770: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
                   2771: @section Exceptions
1.66      anton    2772: @cindex exceptions tutorial
1.48      anton    2773: 
                   2774: @code{throw ( n -- )} causes an exception unless n is zero.
                   2775: 
                   2776: @example
                   2777: 100 throw .s
                   2778: 0 throw .s
                   2779: @end example
                   2780: 
                   2781: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
                   2782: it catches exceptions and pushes the number of the exception on the
                   2783: stack (or 0, if the xt executed without exception).  If there was an
                   2784: exception, the stacks have the same depth as when entering @code{catch}:
                   2785: 
                   2786: @example
                   2787: .s
                   2788: 3 0 ' / catch .s
                   2789: 3 2 ' / catch .s
                   2790: @end example
                   2791: 
1.141     anton    2792: @quotation Assignment
1.48      anton    2793: Try the same with @code{execute} instead of @code{catch}.
1.141     anton    2794: @end quotation
1.48      anton    2795: 
                   2796: @code{Throw} always jumps to the dynamically next enclosing
                   2797: @code{catch}, even if it has to leave several call levels to achieve
                   2798: this:
                   2799: 
                   2800: @example
                   2801: : foo 100 throw ;
                   2802: : foo1 foo ." after foo" ;
1.51      pazsan   2803: : bar ['] foo1 catch ;
1.60      anton    2804: bar .
1.48      anton    2805: @end example
                   2806: 
                   2807: It is often important to restore a value upon leaving a definition, even
                   2808: if the definition is left through an exception.  You can ensure this
                   2809: like this:
                   2810: 
                   2811: @example
                   2812: : ...
                   2813:    save-x
1.51      pazsan   2814:    ['] word-changing-x catch ( ... n )
1.48      anton    2815:    restore-x
                   2816:    ( ... n ) throw ;
                   2817: @end example
                   2818: 
1.55      anton    2819: Gforth provides an alternative syntax in addition to @code{catch}:
1.48      anton    2820: @code{try ... recover ... endtry}.  If the code between @code{try} and
                   2821: @code{recover} has an exception, the stack depths are restored, the
                   2822: exception number is pushed on the stack, and the code between
                   2823: @code{recover} and @code{endtry} is performed.  E.g., the definition for
                   2824: @code{catch} is
                   2825: 
                   2826: @example
                   2827: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
                   2828:   try
                   2829:     execute 0
                   2830:   recover
                   2831:     nip
                   2832:   endtry ;
                   2833: @end example
                   2834: 
                   2835: The equivalent to the restoration code above is
                   2836: 
                   2837: @example
                   2838: : ...
                   2839:   save-x
                   2840:   try
1.92      anton    2841:     word-changing-x 0
                   2842:   recover endtry
1.48      anton    2843:   restore-x
                   2844:   throw ;
                   2845: @end example
                   2846: 
1.92      anton    2847: This works if @code{word-changing-x} does not change the stack depth,
                   2848: otherwise you should add some code between @code{recover} and
                   2849: @code{endtry} to balance the stack.
1.48      anton    2850: 
1.66      anton    2851: Reference: @ref{Exception Handling}.
                   2852: 
1.48      anton    2853: 
                   2854: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
                   2855: @section Defining Words
1.66      anton    2856: @cindex defining words tutorial
                   2857: @cindex does> tutorial
                   2858: @cindex create...does> tutorial
                   2859: 
                   2860: @c before semantics?
1.48      anton    2861: 
                   2862: @code{:}, @code{create}, and @code{variable} are definition words: They
                   2863: define other words.  @code{Constant} is another definition word:
                   2864: 
                   2865: @example
                   2866: 5 constant foo
                   2867: foo .
                   2868: @end example
                   2869: 
                   2870: You can also use the prefixes @code{2} (double-cell) and @code{f}
                   2871: (floating point) with @code{variable} and @code{constant}.
                   2872: 
                   2873: You can also define your own defining words.  E.g.:
                   2874: 
                   2875: @example
                   2876: : variable ( "name" -- )
                   2877:   create 0 , ;
                   2878: @end example
                   2879: 
                   2880: You can also define defining words that create words that do something
                   2881: other than just producing their address:
                   2882: 
                   2883: @example
                   2884: : constant ( n "name" -- )
                   2885:   create ,
                   2886: does> ( -- n )
1.50      anton    2887:   ( addr ) @@ ;
1.48      anton    2888: 
                   2889: 5 constant foo
                   2890: foo .
                   2891: @end example
                   2892: 
                   2893: The definition of @code{constant} above ends at the @code{does>}; i.e.,
                   2894: @code{does>} replaces @code{;}, but it also does something else: It
                   2895: changes the last defined word such that it pushes the address of the
                   2896: body of the word and then performs the code after the @code{does>}
                   2897: whenever it is called.
                   2898: 
                   2899: In the example above, @code{constant} uses @code{,} to store 5 into the
                   2900: body of @code{foo}.  When @code{foo} executes, it pushes the address of
                   2901: the body onto the stack, then (in the code after the @code{does>})
                   2902: fetches the 5 from there.
                   2903: 
                   2904: The stack comment near the @code{does>} reflects the stack effect of the
                   2905: defined word, not the stack effect of the code after the @code{does>}
                   2906: (the difference is that the code expects the address of the body that
                   2907: the stack comment does not show).
                   2908: 
                   2909: You can use these definition words to do factoring in cases that involve
                   2910: (other) definition words.  E.g., a field offset is always added to an
                   2911: address.  Instead of defining
                   2912: 
                   2913: @example
                   2914: 2 cells constant offset-field1
                   2915: @end example
                   2916: 
                   2917: and using this like
                   2918: 
                   2919: @example
                   2920: ( addr ) offset-field1 +
                   2921: @end example
                   2922: 
                   2923: you can define a definition word
                   2924: 
                   2925: @example
                   2926: : simple-field ( n "name" -- )
                   2927:   create ,
                   2928: does> ( n1 -- n1+n )
1.50      anton    2929:   ( addr ) @@ + ;
1.48      anton    2930: @end example
1.21      crook    2931: 
1.48      anton    2932: Definition and use of field offsets now look like this:
1.21      crook    2933: 
1.48      anton    2934: @example
                   2935: 2 cells simple-field field1
1.60      anton    2936: create mystruct 4 cells allot
                   2937: mystruct .s field1 .s drop
1.48      anton    2938: @end example
1.21      crook    2939: 
1.48      anton    2940: If you want to do something with the word without performing the code
                   2941: after the @code{does>}, you can access the body of a @code{create}d word
                   2942: with @code{>body ( xt -- addr )}:
1.21      crook    2943: 
1.48      anton    2944: @example
                   2945: : value ( n "name" -- )
                   2946:   create ,
                   2947: does> ( -- n1 )
1.50      anton    2948:   @@ ;
1.48      anton    2949: : to ( n "name" -- )
                   2950:   ' >body ! ;
1.21      crook    2951: 
1.48      anton    2952: 5 value foo
                   2953: foo .
                   2954: 7 to foo
                   2955: foo .
                   2956: @end example
1.21      crook    2957: 
1.141     anton    2958: @quotation Assignment
1.48      anton    2959: Define @code{defer ( "name" -- )}, which creates a word that stores an
                   2960: XT (at the start the XT of @code{abort}), and upon execution
                   2961: @code{execute}s the XT.  Define @code{is ( xt "name" -- )} that stores
                   2962: @code{xt} into @code{name}, a word defined with @code{defer}.  Indirect
                   2963: recursion is one application of @code{defer}.
1.141     anton    2964: @end quotation
1.29      crook    2965: 
1.66      anton    2966: Reference: @ref{User-defined Defining Words}.
                   2967: 
                   2968: 
1.48      anton    2969: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
                   2970: @section Arrays and Records
1.66      anton    2971: @cindex arrays tutorial
                   2972: @cindex records tutorial
                   2973: @cindex structs tutorial
1.29      crook    2974: 
1.48      anton    2975: Forth has no standard words for defining data structures such as arrays
                   2976: and records (structs in C terminology), but you can build them yourself
                   2977: based on address arithmetic.  You can also define words for defining
                   2978: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
1.29      crook    2979: 
1.48      anton    2980: One of the first projects a Forth newcomer sets out upon when learning
                   2981: about defining words is an array defining word (possibly for
                   2982: n-dimensional arrays).  Go ahead and do it, I did it, too; you will
                   2983: learn something from it.  However, don't be disappointed when you later
                   2984: learn that you have little use for these words (inappropriate use would
                   2985: be even worse).  I have not yet found a set of useful array words yet;
                   2986: the needs are just too diverse, and named, global arrays (the result of
                   2987: naive use of defining words) are often not flexible enough (e.g.,
1.66      anton    2988: consider how to pass them as parameters).  Another such project is a set
                   2989: of words to help dealing with strings.
1.29      crook    2990: 
1.48      anton    2991: On the other hand, there is a useful set of record words, and it has
                   2992: been defined in @file{compat/struct.fs}; these words are predefined in
                   2993: Gforth.  They are explained in depth elsewhere in this manual (see
                   2994: @pxref{Structures}).  The @code{simple-field} example above is
                   2995: simplified variant of fields in this package.
1.21      crook    2996: 
                   2997: 
1.48      anton    2998: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
                   2999: @section @code{POSTPONE}
1.66      anton    3000: @cindex postpone tutorial
1.21      crook    3001: 
1.48      anton    3002: You can compile the compilation semantics (instead of compiling the
                   3003: interpretation semantics) of a word with @code{POSTPONE}:
1.21      crook    3004: 
1.48      anton    3005: @example
                   3006: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
1.51      pazsan   3007:  POSTPONE + ; immediate
1.48      anton    3008: : foo ( n1 n2 -- n )
                   3009:  MY-+ ;
                   3010: 1 2 foo .
                   3011: see foo
                   3012: @end example
1.21      crook    3013: 
1.48      anton    3014: During the definition of @code{foo} the text interpreter performs the
                   3015: compilation semantics of @code{MY-+}, which performs the compilation
                   3016: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
                   3017: 
                   3018: This example also displays separate stack comments for the compilation
                   3019: semantics and for the stack effect of the compiled code.  For words with
                   3020: default compilation semantics these stack effects are usually not
                   3021: displayed; the stack effect of the compilation semantics is always
                   3022: @code{( -- )} for these words, the stack effect for the compiled code is
                   3023: the stack effect of the interpretation semantics.
                   3024: 
                   3025: Note that the state of the interpreter does not come into play when
                   3026: performing the compilation semantics in this way.  You can also perform
                   3027: it interpretively, e.g.:
                   3028: 
                   3029: @example
                   3030: : foo2 ( n1 n2 -- n )
                   3031:  [ MY-+ ] ;
                   3032: 1 2 foo .
                   3033: see foo
                   3034: @end example
1.21      crook    3035: 
1.48      anton    3036: However, there are some broken Forth systems where this does not always
1.62      crook    3037: work, and therefore this practice was been declared non-standard in
1.48      anton    3038: 1999.
                   3039: @c !! repair.fs
                   3040: 
                   3041: Here is another example for using @code{POSTPONE}:
1.44      crook    3042: 
1.48      anton    3043: @example
                   3044: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
                   3045:  POSTPONE negate POSTPONE + ; immediate compile-only
                   3046: : bar ( n1 n2 -- n )
                   3047:   MY-- ;
                   3048: 2 1 bar .
                   3049: see bar
                   3050: @end example
1.21      crook    3051: 
1.48      anton    3052: You can define @code{ENDIF} in this way:
1.21      crook    3053: 
1.48      anton    3054: @example
                   3055: : ENDIF ( Compilation: orig -- )
                   3056:   POSTPONE then ; immediate
                   3057: @end example
1.21      crook    3058: 
1.141     anton    3059: @quotation Assignment
1.48      anton    3060: Write @code{MY-2DUP} that has compilation semantics equivalent to
                   3061: @code{2dup}, but compiles @code{over over}.
1.141     anton    3062: @end quotation
1.29      crook    3063: 
1.66      anton    3064: @c !! @xref{Macros} for reference
                   3065: 
                   3066: 
1.48      anton    3067: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
                   3068: @section @code{Literal}
1.66      anton    3069: @cindex literal tutorial
1.29      crook    3070: 
1.48      anton    3071: You cannot @code{POSTPONE} numbers:
1.21      crook    3072: 
1.48      anton    3073: @example
                   3074: : [FOO] POSTPONE 500 ; immediate
1.21      crook    3075: @end example
                   3076: 
1.48      anton    3077: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
1.29      crook    3078: 
1.48      anton    3079: @example
                   3080: : [FOO] ( compilation: --; run-time: -- n )
                   3081:   500 POSTPONE literal ; immediate
1.29      crook    3082: 
1.60      anton    3083: : flip [FOO] ;
1.48      anton    3084: flip .
                   3085: see flip
                   3086: @end example
1.29      crook    3087: 
1.48      anton    3088: @code{LITERAL} consumes a number at compile-time (when it's compilation
                   3089: semantics are executed) and pushes it at run-time (when the code it
                   3090: compiled is executed).  A frequent use of @code{LITERAL} is to compile a
                   3091: number computed at compile time into the current word:
1.29      crook    3092: 
1.48      anton    3093: @example
                   3094: : bar ( -- n )
                   3095:   [ 2 2 + ] literal ;
                   3096: see bar
                   3097: @end example
1.29      crook    3098: 
1.141     anton    3099: @quotation Assignment
1.48      anton    3100: Write @code{]L} which allows writing the example above as @code{: bar (
                   3101: -- n ) [ 2 2 + ]L ;}
1.141     anton    3102: @end quotation
1.48      anton    3103: 
1.66      anton    3104: @c !! @xref{Macros} for reference
                   3105: 
1.48      anton    3106: 
                   3107: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
                   3108: @section Advanced macros
1.66      anton    3109: @cindex macros, advanced tutorial
                   3110: @cindex run-time code generation, tutorial
1.48      anton    3111: 
1.66      anton    3112: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
                   3113: Execution Tokens}.  It frequently performs @code{execute}, a relatively
                   3114: expensive operation in some Forth implementations.  You can use
1.48      anton    3115: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
                   3116: and produce a word that contains the word to be performed directly:
                   3117: 
                   3118: @c use ]] ... [[
                   3119: @example
                   3120: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
                   3121: \ at run-time, execute xt ( ... x -- ... ) for each element of the
                   3122: \ array beginning at addr and containing u elements
                   3123:   @{ xt @}
                   3124:   POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
1.50      anton    3125:     POSTPONE i POSTPONE @@ xt compile,
1.48      anton    3126:   1 cells POSTPONE literal POSTPONE +loop ;
                   3127: 
                   3128: : sum-array ( addr u -- n )
                   3129:  0 rot rot [ ' + compile-map-array ] ;
                   3130: see sum-array
                   3131: a 5 sum-array .
                   3132: @end example
                   3133: 
                   3134: You can use the full power of Forth for generating the code; here's an
                   3135: example where the code is generated in a loop:
                   3136: 
                   3137: @example
                   3138: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
                   3139: \ n2=n1+(addr1)*n, addr2=addr1+cell
1.50      anton    3140:   POSTPONE tuck POSTPONE @@
1.48      anton    3141:   POSTPONE literal POSTPONE * POSTPONE +
                   3142:   POSTPONE swap POSTPONE cell+ ;
                   3143: 
                   3144: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
1.51      pazsan   3145: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
1.48      anton    3146:   0 postpone literal postpone swap
                   3147:   [ ' compile-vmul-step compile-map-array ]
                   3148:   postpone drop ;
                   3149: see compile-vmul
                   3150: 
                   3151: : a-vmul ( addr -- n )
1.51      pazsan   3152: \ n=a*v, where v is a vector that's as long as a and starts at addr
1.48      anton    3153:  [ a 5 compile-vmul ] ;
                   3154: see a-vmul
                   3155: a a-vmul .
                   3156: @end example
                   3157: 
                   3158: This example uses @code{compile-map-array} to show off, but you could
1.66      anton    3159: also use @code{map-array} instead (try it now!).
1.48      anton    3160: 
                   3161: You can use this technique for efficient multiplication of large
                   3162: matrices.  In matrix multiplication, you multiply every line of one
                   3163: matrix with every column of the other matrix.  You can generate the code
                   3164: for one line once, and use it for every column.  The only downside of
                   3165: this technique is that it is cumbersome to recover the memory consumed
                   3166: by the generated code when you are done (and in more complicated cases
                   3167: it is not possible portably).
                   3168: 
1.66      anton    3169: @c !! @xref{Macros} for reference
                   3170: 
                   3171: 
1.48      anton    3172: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
                   3173: @section Compilation Tokens
1.66      anton    3174: @cindex compilation tokens, tutorial
                   3175: @cindex CT, tutorial
1.48      anton    3176: 
                   3177: This section is Gforth-specific.  You can skip it.
                   3178: 
                   3179: @code{' word compile,} compiles the interpretation semantics.  For words
                   3180: with default compilation semantics this is the same as performing the
                   3181: compilation semantics.  To represent the compilation semantics of other
                   3182: words (e.g., words like @code{if} that have no interpretation
                   3183: semantics), Gforth has the concept of a compilation token (CT,
                   3184: consisting of two cells), and words @code{comp'} and @code{[comp']}.
                   3185: You can perform the compilation semantics represented by a CT with
                   3186: @code{execute}:
1.29      crook    3187: 
1.48      anton    3188: @example
                   3189: : foo2 ( n1 n2 -- n )
                   3190:    [ comp' + execute ] ;
                   3191: see foo
                   3192: @end example
1.29      crook    3193: 
1.48      anton    3194: You can compile the compilation semantics represented by a CT with
                   3195: @code{postpone,}:
1.30      anton    3196: 
1.48      anton    3197: @example
                   3198: : foo3 ( -- )
                   3199:   [ comp' + postpone, ] ;
                   3200: see foo3
                   3201: @end example
1.30      anton    3202: 
1.51      pazsan   3203: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
1.48      anton    3204: @code{comp'} is particularly useful for words that have no
                   3205: interpretation semantics:
1.29      crook    3206: 
1.30      anton    3207: @example
1.48      anton    3208: ' if
1.60      anton    3209: comp' if .s 2drop
1.30      anton    3210: @end example
                   3211: 
1.66      anton    3212: Reference: @ref{Tokens for Words}.
                   3213: 
1.29      crook    3214: 
1.48      anton    3215: @node Wordlists and Search Order Tutorial,  , Compilation Tokens Tutorial, Tutorial
                   3216: @section Wordlists and Search Order
1.66      anton    3217: @cindex wordlists tutorial
                   3218: @cindex search order, tutorial
1.48      anton    3219: 
                   3220: The dictionary is not just a memory area that allows you to allocate
                   3221: memory with @code{allot}, it also contains the Forth words, arranged in
                   3222: several wordlists.  When searching for a word in a wordlist,
                   3223: conceptually you start searching at the youngest and proceed towards
                   3224: older words (in reality most systems nowadays use hash-tables); i.e., if
                   3225: you define a word with the same name as an older word, the new word
                   3226: shadows the older word.
                   3227: 
                   3228: Which wordlists are searched in which order is determined by the search
                   3229: order.  You can display the search order with @code{order}.  It displays
                   3230: first the search order, starting with the wordlist searched first, then
                   3231: it displays the wordlist that will contain newly defined words.
1.21      crook    3232: 
1.48      anton    3233: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
1.21      crook    3234: 
1.48      anton    3235: @example
                   3236: wordlist constant mywords
                   3237: @end example
1.21      crook    3238: 
1.48      anton    3239: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
                   3240: defined words (the @emph{current} wordlist):
1.21      crook    3241: 
1.48      anton    3242: @example
                   3243: mywords set-current
                   3244: order
                   3245: @end example
1.26      crook    3246: 
1.48      anton    3247: Gforth does not display a name for the wordlist in @code{mywords}
                   3248: because this wordlist was created anonymously with @code{wordlist}.
1.21      crook    3249: 
1.48      anton    3250: You can get the current wordlist with @code{get-current ( -- wid)}.  If
                   3251: you want to put something into a specific wordlist without overall
                   3252: effect on the current wordlist, this typically looks like this:
1.21      crook    3253: 
1.48      anton    3254: @example
                   3255: get-current mywords set-current ( wid )
                   3256: create someword
                   3257: ( wid ) set-current
                   3258: @end example
1.21      crook    3259: 
1.48      anton    3260: You can write the search order with @code{set-order ( wid1 .. widn n --
                   3261: )} and read it with @code{get-order ( -- wid1 .. widn n )}.  The first
                   3262: searched wordlist is topmost.
1.21      crook    3263: 
1.48      anton    3264: @example
                   3265: get-order mywords swap 1+ set-order
                   3266: order
                   3267: @end example
1.21      crook    3268: 
1.48      anton    3269: Yes, the order of wordlists in the output of @code{order} is reversed
                   3270: from stack comments and the output of @code{.s} and thus unintuitive.
1.21      crook    3271: 
1.141     anton    3272: @quotation Assignment
1.48      anton    3273: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
                   3274: wordlist to the search order.  Define @code{previous ( -- )}, which
                   3275: removes the first searched wordlist from the search order.  Experiment
                   3276: with boundary conditions (you will see some crashes or situations that
                   3277: are hard or impossible to leave).
1.141     anton    3278: @end quotation
1.21      crook    3279: 
1.48      anton    3280: The search order is a powerful foundation for providing features similar
                   3281: to Modula-2 modules and C++ namespaces.  However, trying to modularize
                   3282: programs in this way has disadvantages for debugging and reuse/factoring
                   3283: that overcome the advantages in my experience (I don't do huge projects,
1.55      anton    3284: though).  These disadvantages are not so clear in other
1.82      anton    3285: languages/programming environments, because these languages are not so
1.48      anton    3286: strong in debugging and reuse.
1.21      crook    3287: 
1.66      anton    3288: @c !! example
                   3289: 
                   3290: Reference: @ref{Word Lists}.
1.21      crook    3291: 
1.29      crook    3292: @c ******************************************************************
1.48      anton    3293: @node Introduction, Words, Tutorial, Top
1.29      crook    3294: @comment node-name,     next,           previous, up
                   3295: @chapter An Introduction to ANS Forth
                   3296: @cindex Forth - an introduction
1.21      crook    3297: 
1.83      anton    3298: The difference of this chapter from the Tutorial (@pxref{Tutorial}) is
                   3299: that it is slower-paced in its examples, but uses them to dive deep into
                   3300: explaining Forth internals (not covered by the Tutorial).  Apart from
                   3301: that, this chapter covers far less material.  It is suitable for reading
                   3302: without using a computer.
                   3303: 
1.29      crook    3304: The primary purpose of this manual is to document Gforth. However, since
                   3305: Forth is not a widely-known language and there is a lack of up-to-date
                   3306: teaching material, it seems worthwhile to provide some introductory
1.49      anton    3307: material.  For other sources of Forth-related
                   3308: information, see @ref{Forth-related information}.
1.21      crook    3309: 
1.29      crook    3310: The examples in this section should work on any ANS Forth; the
                   3311: output shown was produced using Gforth. Each example attempts to
                   3312: reproduce the exact output that Gforth produces. If you try out the
                   3313: examples (and you should), what you should type is shown @kbd{like this}
                   3314: and Gforth's response is shown @code{like this}. The single exception is
1.30      anton    3315: that, where the example shows @key{RET} it means that you should
1.29      crook    3316: press the ``carriage return'' key. Unfortunately, some output formats for
                   3317: this manual cannot show the difference between @kbd{this} and
                   3318: @code{this} which will make trying out the examples harder (but not
                   3319: impossible).
1.21      crook    3320: 
1.29      crook    3321: Forth is an unusual language. It provides an interactive development
                   3322: environment which includes both an interpreter and compiler. Forth
                   3323: programming style encourages you to break a problem down into many
                   3324: @cindex factoring
                   3325: small fragments (@dfn{factoring}), and then to develop and test each
                   3326: fragment interactively. Forth advocates assert that breaking the
                   3327: edit-compile-test cycle used by conventional programming languages can
                   3328: lead to great productivity improvements.
1.21      crook    3329: 
1.29      crook    3330: @menu
1.67      anton    3331: * Introducing the Text Interpreter::  
                   3332: * Stacks and Postfix notation::  
                   3333: * Your first definition::       
                   3334: * How does that work?::         
                   3335: * Forth is written in Forth::   
                   3336: * Review - elements of a Forth system::  
                   3337: * Where to go next::            
                   3338: * Exercises::                   
1.29      crook    3339: @end menu
1.21      crook    3340: 
1.29      crook    3341: @comment ----------------------------------------------
                   3342: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
                   3343: @section Introducing the Text Interpreter
                   3344: @cindex text interpreter
                   3345: @cindex outer interpreter
1.21      crook    3346: 
1.30      anton    3347: @c IMO this is too detailed and the pace is too slow for
                   3348: @c an introduction.  If you know German, take a look at
                   3349: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html 
                   3350: @c to see how I do it - anton 
                   3351: 
1.44      crook    3352: @c nac-> Where I have accepted your comments 100% and modified the text
                   3353: @c accordingly, I have deleted your comments. Elsewhere I have added a
                   3354: @c response like this to attempt to rationalise what I have done. Of
                   3355: @c course, this is a very clumsy mechanism for something that would be
                   3356: @c done far more efficiently over a beer. Please delete any dialogue
                   3357: @c you consider closed.
                   3358: 
1.29      crook    3359: When you invoke the Forth image, you will see a startup banner printed
                   3360: and nothing else (if you have Gforth installed on your system, try
1.30      anton    3361: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
1.29      crook    3362: its command line interpreter, which is called the @dfn{Text Interpreter}
                   3363: (also known as the @dfn{Outer Interpreter}).  (You will learn a lot
1.49      anton    3364: about the text interpreter as you read through this chapter, for more
                   3365: detail @pxref{The Text Interpreter}).
1.21      crook    3366: 
1.29      crook    3367: Although it's not obvious, Forth is actually waiting for your
1.30      anton    3368: input. Type a number and press the @key{RET} key:
1.21      crook    3369: 
1.26      crook    3370: @example
1.30      anton    3371: @kbd{45@key{RET}}  ok
1.26      crook    3372: @end example
1.21      crook    3373: 
1.29      crook    3374: Rather than give you a prompt to invite you to input something, the text
                   3375: interpreter prints a status message @i{after} it has processed a line
                   3376: of input. The status message in this case (``@code{ ok}'' followed by
                   3377: carriage-return) indicates that the text interpreter was able to process
                   3378: all of your input successfully. Now type something illegal:
                   3379: 
                   3380: @example
1.30      anton    3381: @kbd{qwer341@key{RET}}
1.134     anton    3382: *the terminal*:2: Undefined word
                   3383: >>>qwer341<<<
                   3384: Backtrace:
                   3385: $2A95B42A20 throw 
                   3386: $2A95B57FB8 no.extensions 
1.29      crook    3387: @end example
1.23      crook    3388: 
1.134     anton    3389: The exact text, other than the ``Undefined word'' may differ slightly
                   3390: on your system, but the effect is the same; when the text interpreter
1.29      crook    3391: detects an error, it discards any remaining text on a line, resets
1.134     anton    3392: certain internal state and prints an error message. For a detailed
                   3393: description of error messages see @ref{Error messages}.
1.23      crook    3394: 
1.29      crook    3395: The text interpreter waits for you to press carriage-return, and then
                   3396: processes your input line. Starting at the beginning of the line, it
                   3397: breaks the line into groups of characters separated by spaces. For each
                   3398: group of characters in turn, it makes two attempts to do something:
1.23      crook    3399: 
1.29      crook    3400: @itemize @bullet
                   3401: @item
1.44      crook    3402: @cindex name dictionary
1.29      crook    3403: It tries to treat it as a command. It does this by searching a @dfn{name
                   3404: dictionary}. If the group of characters matches an entry in the name
                   3405: dictionary, the name dictionary provides the text interpreter with
                   3406: information that allows the text interpreter perform some actions. In
                   3407: Forth jargon, we say that the group
                   3408: @cindex word
                   3409: @cindex definition
                   3410: @cindex execution token
                   3411: @cindex xt
                   3412: of characters names a @dfn{word}, that the dictionary search returns an
                   3413: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
                   3414: word, and that the text interpreter executes the xt. Often, the terms
                   3415: @dfn{word} and @dfn{definition} are used interchangeably.
                   3416: @item
                   3417: If the text interpreter fails to find a match in the name dictionary, it
                   3418: tries to treat the group of characters as a number in the current number
                   3419: base (when you start up Forth, the current number base is base 10). If
                   3420: the group of characters legitimately represents a number, the text
                   3421: interpreter pushes the number onto a stack (we'll learn more about that
                   3422: in the next section).
                   3423: @end itemize
1.23      crook    3424: 
1.29      crook    3425: If the text interpreter is unable to do either of these things with any
                   3426: group of characters, it discards the group of characters and the rest of
                   3427: the line, then prints an error message. If the text interpreter reaches
                   3428: the end of the line without error, it prints the status message ``@code{ ok}''
                   3429: followed by carriage-return.
1.21      crook    3430: 
1.29      crook    3431: This is the simplest command we can give to the text interpreter:
1.23      crook    3432: 
                   3433: @example
1.30      anton    3434: @key{RET}  ok
1.23      crook    3435: @end example
1.21      crook    3436: 
1.29      crook    3437: The text interpreter did everything we asked it to do (nothing) without
                   3438: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
                   3439: command:
1.21      crook    3440: 
1.23      crook    3441: @example
1.30      anton    3442: @kbd{12 dup fred dup@key{RET}}
1.134     anton    3443: *the terminal*:3: Undefined word
                   3444: 12 dup >>>fred<<< dup
                   3445: Backtrace:
                   3446: $2A95B42A20 throw 
                   3447: $2A95B57FB8 no.extensions 
1.23      crook    3448: @end example
1.21      crook    3449: 
1.29      crook    3450: When you press the carriage-return key, the text interpreter starts to
                   3451: work its way along the line:
1.21      crook    3452: 
1.29      crook    3453: @itemize @bullet
                   3454: @item
                   3455: When it gets to the space after the @code{2}, it takes the group of
                   3456: characters @code{12} and looks them up in the name
                   3457: dictionary@footnote{We can't tell if it found them or not, but assume
                   3458: for now that it did not}. There is no match for this group of characters
                   3459: in the name dictionary, so it tries to treat them as a number. It is
                   3460: able to do this successfully, so it puts the number, 12, ``on the stack''
                   3461: (whatever that means).
                   3462: @item
                   3463: The text interpreter resumes scanning the line and gets the next group
                   3464: of characters, @code{dup}. It looks it up in the name dictionary and
                   3465: (you'll have to take my word for this) finds it, and executes the word
                   3466: @code{dup} (whatever that means).
                   3467: @item
                   3468: Once again, the text interpreter resumes scanning the line and gets the
                   3469: group of characters @code{fred}. It looks them up in the name
                   3470: dictionary, but can't find them. It tries to treat them as a number, but
                   3471: they don't represent any legal number.
                   3472: @end itemize
1.21      crook    3473: 
1.29      crook    3474: At this point, the text interpreter gives up and prints an error
                   3475: message. The error message shows exactly how far the text interpreter
                   3476: got in processing the line. In particular, it shows that the text
                   3477: interpreter made no attempt to do anything with the final character
                   3478: group, @code{dup}, even though we have good reason to believe that the
                   3479: text interpreter would have no problem looking that word up and
                   3480: executing it a second time.
1.21      crook    3481: 
                   3482: 
1.29      crook    3483: @comment ----------------------------------------------
                   3484: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
                   3485: @section Stacks, postfix notation and parameter passing
                   3486: @cindex text interpreter
                   3487: @cindex outer interpreter
1.21      crook    3488: 
1.29      crook    3489: In procedural programming languages (like C and Pascal), the
                   3490: building-block of programs is the @dfn{function} or @dfn{procedure}. These
                   3491: functions or procedures are called with @dfn{explicit parameters}. For
                   3492: example, in C we might write:
1.21      crook    3493: 
1.23      crook    3494: @example
1.29      crook    3495: total = total + new_volume(length,height,depth);
1.23      crook    3496: @end example
1.21      crook    3497: 
1.23      crook    3498: @noindent
1.29      crook    3499: where new_volume is a function-call to another piece of code, and total,
                   3500: length, height and depth are all variables. length, height and depth are
                   3501: parameters to the function-call.
1.21      crook    3502: 
1.29      crook    3503: In Forth, the equivalent of the function or procedure is the
                   3504: @dfn{definition} and parameters are implicitly passed between
                   3505: definitions using a shared stack that is visible to the
                   3506: programmer. Although Forth does support variables, the existence of the
                   3507: stack means that they are used far less often than in most other
                   3508: programming languages. When the text interpreter encounters a number, it
                   3509: will place (@dfn{push}) it on the stack. There are several stacks (the
1.30      anton    3510: actual number is implementation-dependent ...) and the particular stack
1.29      crook    3511: used for any operation is implied unambiguously by the operation being
                   3512: performed. The stack used for all integer operations is called the @dfn{data
                   3513: stack} and, since this is the stack used most commonly, references to
                   3514: ``the data stack'' are often abbreviated to ``the stack''.
1.21      crook    3515: 
1.29      crook    3516: The stacks have a last-in, first-out (LIFO) organisation. If you type:
1.21      crook    3517: 
1.23      crook    3518: @example
1.30      anton    3519: @kbd{1 2 3@key{RET}}  ok
1.23      crook    3520: @end example
1.21      crook    3521: 
1.29      crook    3522: Then this instructs the text interpreter to placed three numbers on the
                   3523: (data) stack. An analogy for the behaviour of the stack is to take a
                   3524: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
                   3525: the table. The 3 was the last card onto the pile (``last-in'') and if
                   3526: you take a card off the pile then, unless you're prepared to fiddle a
                   3527: bit, the card that you take off will be the 3 (``first-out''). The
                   3528: number that will be first-out of the stack is called the @dfn{top of
                   3529: stack}, which
                   3530: @cindex TOS definition
                   3531: is often abbreviated to @dfn{TOS}.
1.21      crook    3532: 
1.29      crook    3533: To understand how parameters are passed in Forth, consider the
                   3534: behaviour of the definition @code{+} (pronounced ``plus''). You will not
                   3535: be surprised to learn that this definition performs addition. More
                   3536: precisely, it adds two number together and produces a result. Where does
                   3537: it get the two numbers from? It takes the top two numbers off the
                   3538: stack. Where does it place the result? On the stack. You can act-out the
                   3539: behaviour of @code{+} with your playing cards like this:
1.21      crook    3540: 
                   3541: @itemize @bullet
                   3542: @item
1.29      crook    3543: Pick up two cards from the stack on the table
1.21      crook    3544: @item
1.29      crook    3545: Stare at them intently and ask yourself ``what @i{is} the sum of these two
                   3546: numbers''
1.21      crook    3547: @item
1.29      crook    3548: Decide that the answer is 5
1.21      crook    3549: @item
1.29      crook    3550: Shuffle the two cards back into the pack and find a 5
1.21      crook    3551: @item
1.29      crook    3552: Put a 5 on the remaining ace that's on the table.
1.21      crook    3553: @end itemize
                   3554: 
1.29      crook    3555: If you don't have a pack of cards handy but you do have Forth running,
                   3556: you can use the definition @code{.s} to show the current state of the stack,
                   3557: without affecting the stack. Type:
1.21      crook    3558: 
                   3559: @example
1.124     anton    3560: @kbd{clearstacks 1 2 3@key{RET}} ok
1.30      anton    3561: @kbd{.s@key{RET}} <3> 1 2 3  ok
1.23      crook    3562: @end example
                   3563: 
1.124     anton    3564: The text interpreter looks up the word @code{clearstacks} and executes
                   3565: it; it tidies up the stacks and removes any entries that may have been
1.29      crook    3566: left on it by earlier examples. The text interpreter pushes each of the
                   3567: three numbers in turn onto the stack. Finally, the text interpreter
                   3568: looks up the word @code{.s} and executes it. The effect of executing
                   3569: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
                   3570: followed by a list of all the items on the stack; the item on the far
                   3571: right-hand side is the TOS.
1.21      crook    3572: 
1.29      crook    3573: You can now type:
1.21      crook    3574: 
1.29      crook    3575: @example
1.30      anton    3576: @kbd{+ .s@key{RET}} <2> 1 5  ok
1.29      crook    3577: @end example
1.21      crook    3578: 
1.29      crook    3579: @noindent
                   3580: which is correct; there are now 2 items on the stack and the result of
                   3581: the addition is 5.
1.23      crook    3582: 
1.29      crook    3583: If you're playing with cards, try doing a second addition: pick up the
                   3584: two cards, work out that their sum is 6, shuffle them into the pack,
                   3585: look for a 6 and place that on the table. You now have just one item on
                   3586: the stack. What happens if you try to do a third addition? Pick up the
                   3587: first card, pick up the second card -- ah! There is no second card. This
                   3588: is called a @dfn{stack underflow} and consitutes an error. If you try to
1.95      anton    3589: do the same thing with Forth it often reports an error (probably a Stack
1.29      crook    3590: Underflow or an Invalid Memory Address error).
1.23      crook    3591: 
1.29      crook    3592: The opposite situation to a stack underflow is a @dfn{stack overflow},
                   3593: which simply accepts that there is a finite amount of storage space
                   3594: reserved for the stack. To stretch the playing card analogy, if you had
                   3595: enough packs of cards and you piled the cards up on the table, you would
                   3596: eventually be unable to add another card; you'd hit the ceiling. Gforth
                   3597: allows you to set the maximum size of the stacks. In general, the only
                   3598: time that you will get a stack overflow is because a definition has a
                   3599: bug in it and is generating data on the stack uncontrollably.
1.23      crook    3600: 
1.29      crook    3601: There's one final use for the playing card analogy. If you model your
                   3602: stack using a pack of playing cards, the maximum number of items on
                   3603: your stack will be 52 (I assume you didn't use the Joker). The maximum
                   3604: @i{value} of any item on the stack is 13 (the King). In fact, the only
                   3605: possible numbers are positive integer numbers 1 through 13; you can't
                   3606: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
                   3607: think about some of the cards, you can accommodate different
                   3608: numbers. For example, you could think of the Jack as representing 0,
                   3609: the Queen as representing -1 and the King as representing -2. Your
1.45      crook    3610: @i{range} remains unchanged (you can still only represent a total of 13
1.29      crook    3611: numbers) but the numbers that you can represent are -2 through 10.
1.28      crook    3612: 
1.29      crook    3613: In that analogy, the limit was the amount of information that a single
                   3614: stack entry could hold, and Forth has a similar limit. In Forth, the
                   3615: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
                   3616: implementation dependent and affects the maximum value that a stack
                   3617: entry can hold. A Standard Forth provides a cell size of at least
                   3618: 16-bits, and most desktop systems use a cell size of 32-bits.
1.21      crook    3619: 
1.29      crook    3620: Forth does not do any type checking for you, so you are free to
                   3621: manipulate and combine stack items in any way you wish. A convenient way
                   3622: of treating stack items is as 2's complement signed integers, and that
                   3623: is what Standard words like @code{+} do. Therefore you can type:
1.21      crook    3624: 
1.29      crook    3625: @example
1.30      anton    3626: @kbd{-5 12 + .s@key{RET}} <1> 7  ok
1.29      crook    3627: @end example
1.21      crook    3628: 
1.29      crook    3629: If you use numbers and definitions like @code{+} in order to turn Forth
                   3630: into a great big pocket calculator, you will realise that it's rather
                   3631: different from a normal calculator. Rather than typing 2 + 3 = you had
                   3632: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
                   3633: result). The terminology used to describe this difference is to say that
                   3634: your calculator uses @dfn{Infix Notation} (parameters and operators are
                   3635: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
                   3636: operators are separate), also called @dfn{Reverse Polish Notation}.
1.21      crook    3637: 
1.29      crook    3638: Whilst postfix notation might look confusing to begin with, it has
                   3639: several important advantages:
1.21      crook    3640: 
1.23      crook    3641: @itemize @bullet
                   3642: @item
1.29      crook    3643: it is unambiguous
1.23      crook    3644: @item
1.29      crook    3645: it is more concise
1.23      crook    3646: @item
1.29      crook    3647: it fits naturally with a stack-based system
1.23      crook    3648: @end itemize
1.21      crook    3649: 
1.29      crook    3650: To examine these claims in more detail, consider these sums:
1.21      crook    3651: 
1.29      crook    3652: @example
                   3653: 6 + 5 * 4 =
                   3654: 4 * 5 + 6 =
                   3655: @end example
1.21      crook    3656: 
1.29      crook    3657: If you're just learning maths or your maths is very rusty, you will
                   3658: probably come up with the answer 44 for the first and 26 for the
                   3659: second. If you are a bit of a whizz at maths you will remember the
                   3660: @i{convention} that multiplication takes precendence over addition, and
                   3661: you'd come up with the answer 26 both times. To explain the answer 26
                   3662: to someone who got the answer 44, you'd probably rewrite the first sum
                   3663: like this:
1.21      crook    3664: 
1.29      crook    3665: @example
                   3666: 6 + (5 * 4) =
                   3667: @end example
1.21      crook    3668: 
1.29      crook    3669: If what you really wanted was to perform the addition before the
                   3670: multiplication, you would have to use parentheses to force it.
1.21      crook    3671: 
1.29      crook    3672: If you did the first two sums on a pocket calculator you would probably
                   3673: get the right answers, unless you were very cautious and entered them using
                   3674: these keystroke sequences:
1.21      crook    3675: 
1.29      crook    3676: 6 + 5 = * 4 =
                   3677: 4 * 5 = + 6 =
1.21      crook    3678: 
1.29      crook    3679: Postfix notation is unambiguous because the order that the operators
                   3680: are applied is always explicit; that also means that parentheses are
                   3681: never required. The operators are @i{active} (the act of quoting the
                   3682: operator makes the operation occur) which removes the need for ``=''.
1.28      crook    3683: 
1.29      crook    3684: The sum 6 + 5 * 4 can be written (in postfix notation) in two
                   3685: equivalent ways:
1.26      crook    3686: 
                   3687: @example
1.29      crook    3688: 6 5 4 * +      or:
                   3689: 5 4 * 6 +
1.26      crook    3690: @end example
1.23      crook    3691: 
1.29      crook    3692: An important thing that you should notice about this notation is that
                   3693: the @i{order} of the numbers does not change; if you want to subtract
                   3694: 2 from 10 you type @code{10 2 -}.
1.1       anton    3695: 
1.29      crook    3696: The reason that Forth uses postfix notation is very simple to explain: it
                   3697: makes the implementation extremely simple, and it follows naturally from
                   3698: using the stack as a mechanism for passing parameters. Another way of
                   3699: thinking about this is to realise that all Forth definitions are
                   3700: @i{active}; they execute as they are encountered by the text
                   3701: interpreter. The result of this is that the syntax of Forth is trivially
                   3702: simple.
1.1       anton    3703: 
                   3704: 
                   3705: 
1.29      crook    3706: @comment ----------------------------------------------
                   3707: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
                   3708: @section Your first Forth definition
                   3709: @cindex first definition
1.1       anton    3710: 
1.29      crook    3711: Until now, the examples we've seen have been trivial; we've just been
                   3712: using Forth as a bigger-than-pocket calculator. Also, each calculation
                   3713: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
                   3714: again@footnote{That's not quite true. If you press the up-arrow key on
                   3715: your keyboard you should be able to scroll back to any earlier command,
                   3716: edit it and re-enter it.} In this section we'll see how to add new
                   3717: words to Forth's vocabulary.
1.1       anton    3718: 
1.29      crook    3719: The easiest way to create a new word is to use a @dfn{colon
                   3720: definition}. We'll define a few and try them out before worrying too
                   3721: much about how they work. Try typing in these examples; be careful to
                   3722: copy the spaces accurately:
1.1       anton    3723: 
1.29      crook    3724: @example
                   3725: : add-two 2 + . ;
                   3726: : greet ." Hello and welcome" ;
                   3727: : demo 5 add-two ;
                   3728: @end example
1.1       anton    3729: 
1.29      crook    3730: @noindent
                   3731: Now try them out:
1.1       anton    3732: 
1.29      crook    3733: @example
1.30      anton    3734: @kbd{greet@key{RET}} Hello and welcome  ok
                   3735: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome  ok
                   3736: @kbd{4 add-two@key{RET}} 6  ok
                   3737: @kbd{demo@key{RET}} 7  ok
                   3738: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11  ok
1.29      crook    3739: @end example
1.1       anton    3740: 
1.29      crook    3741: The first new thing that we've introduced here is the pair of words
                   3742: @code{:} and @code{;}. These are used to start and terminate a new
                   3743: definition, respectively. The first word after the @code{:} is the name
                   3744: for the new definition.
1.1       anton    3745: 
1.29      crook    3746: As you can see from the examples, a definition is built up of words that
                   3747: have already been defined; Forth makes no distinction between
                   3748: definitions that existed when you started the system up, and those that
                   3749: you define yourself.
1.1       anton    3750: 
1.29      crook    3751: The examples also introduce the words @code{.} (dot), @code{."}
                   3752: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
                   3753: the stack and displays it. It's like @code{.s} except that it only
                   3754: displays the top item of the stack and it is destructive; after it has
                   3755: executed, the number is no longer on the stack. There is always one
                   3756: space printed after the number, and no spaces before it. Dot-quote
                   3757: defines a string (a sequence of characters) that will be printed when
                   3758: the word is executed. The string can contain any printable characters
                   3759: except @code{"}. A @code{"} has a special function; it is not a Forth
                   3760: word but it acts as a delimiter (the way that delimiters work is
                   3761: described in the next section). Finally, @code{dup} duplicates the value
                   3762: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
1.1       anton    3763: 
1.29      crook    3764: We already know that the text interpreter searches through the
                   3765: dictionary to locate names. If you've followed the examples earlier, you
                   3766: will already have a definition called @code{add-two}. Lets try modifying
                   3767: it by typing in a new definition:
1.1       anton    3768: 
1.29      crook    3769: @example
1.30      anton    3770: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two  ok
1.29      crook    3771: @end example
1.5       anton    3772: 
1.29      crook    3773: Forth recognised that we were defining a word that already exists, and
                   3774: printed a message to warn us of that fact. Let's try out the new
                   3775: definition:
1.5       anton    3776: 
1.29      crook    3777: @example
1.30      anton    3778: @kbd{9 add-two@key{RET}} 9 + 2 =11  ok
1.29      crook    3779: @end example
1.1       anton    3780: 
1.29      crook    3781: @noindent
                   3782: All that we've actually done here, though, is to create a new
                   3783: definition, with a particular name. The fact that there was already a
                   3784: definition with the same name did not make any difference to the way
                   3785: that the new definition was created (except that Forth printed a warning
                   3786: message). The old definition of add-two still exists (try @code{demo}
                   3787: again to see that this is true). Any new definition will use the new
                   3788: definition of @code{add-two}, but old definitions continue to use the
                   3789: version that already existed at the time that they were @code{compiled}.
1.1       anton    3790: 
1.29      crook    3791: Before you go on to the next section, try defining and redefining some
                   3792: words of your own.
1.1       anton    3793: 
1.29      crook    3794: @comment ----------------------------------------------
                   3795: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
                   3796: @section How does that work?
                   3797: @cindex parsing words
1.1       anton    3798: 
1.30      anton    3799: @c That's pretty deep (IMO way too deep) for an introduction. - anton
                   3800: 
                   3801: @c Is it a good idea to talk about the interpretation semantics of a
                   3802: @c number? We don't have an xt to go along with it. - anton
                   3803: 
                   3804: @c Now that I have eliminated execution semantics, I wonder if it would not
                   3805: @c be better to keep them (or add run-time semantics), to make it easier to
                   3806: @c explain what compilation semantics usually does. - anton
                   3807: 
1.44      crook    3808: @c nac-> I removed the term ``default compilation sematics'' from the
                   3809: @c introductory chapter. Removing ``execution semantics'' was making
                   3810: @c everything simpler to explain, then I think the use of this term made
                   3811: @c everything more complex again. I replaced it with ``default
                   3812: @c semantics'' (which is used elsewhere in the manual) by which I mean
                   3813: @c ``a definition that has neither the immediate nor the compile-only
1.83      anton    3814: @c flag set''.
                   3815: 
                   3816: @c anton: I have eliminated default semantics (except in one place where it
                   3817: @c means "default interpretation and compilation semantics"), because it
                   3818: @c makes no sense in the presence of combined words.  I reverted to
                   3819: @c "execution semantics" where necessary.
                   3820: 
                   3821: @c nac-> I reworded big chunks of the ``how does that work''
1.44      crook    3822: @c section (and, unusually for me, I think I even made it shorter!).  See
                   3823: @c what you think -- I know I have not addressed your primary concern
                   3824: @c that it is too heavy-going for an introduction. From what I understood
                   3825: @c of your course notes it looks as though they might be a good framework. 
                   3826: @c Things that I've tried to capture here are some things that came as a
                   3827: @c great revelation here when I first understood them. Also, I like the
                   3828: @c fact that a very simple code example shows up almost all of the issues
                   3829: @c that you need to understand to see how Forth works. That's unique and
                   3830: @c worthwhile to emphasise.
                   3831: 
1.83      anton    3832: @c anton: I think it's a good idea to present the details, especially those
                   3833: @c that you found to be a revelation, and probably the tutorial tries to be
                   3834: @c too superficial and does not get some of the things across that make
                   3835: @c Forth special.  I do believe that most of the time these things should
                   3836: @c be discussed at the end of a section or in separate sections instead of
                   3837: @c in the middle of a section (e.g., the stuff you added in "User-defined
                   3838: @c defining words" leads in a completely different direction from the rest
                   3839: @c of the section).
                   3840: 
1.29      crook    3841: Now we're going to take another look at the definition of @code{add-two}
                   3842: from the previous section. From our knowledge of the way that the text
                   3843: interpreter works, we would have expected this result when we tried to
                   3844: define @code{add-two}:
1.21      crook    3845: 
1.29      crook    3846: @example
1.44      crook    3847: @kbd{: add-two 2 + . ;@key{RET}}
1.134     anton    3848: *the terminal*:4: Undefined word
                   3849: : >>>add-two<<< 2 + . ;
1.29      crook    3850: @end example
1.28      crook    3851: 
1.29      crook    3852: The reason that this didn't happen is bound up in the way that @code{:}
                   3853: works. The word @code{:} does two special things. The first special
                   3854: thing that it does prevents the text interpreter from ever seeing the
                   3855: characters @code{add-two}. The text interpreter uses a variable called
                   3856: @cindex modifying >IN
1.44      crook    3857: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
1.29      crook    3858: input line. When it encounters the word @code{:} it behaves in exactly
                   3859: the same way as it does for any other word; it looks it up in the name
                   3860: dictionary, finds its xt and executes it. When @code{:} executes, it
                   3861: looks at the input buffer, finds the word @code{add-two} and advances the
                   3862: value of @code{>IN} to point past it. It then does some other stuff
                   3863: associated with creating the new definition (including creating an entry
                   3864: for @code{add-two} in the name dictionary). When the execution of @code{:}
                   3865: completes, control returns to the text interpreter, which is oblivious
                   3866: to the fact that it has been tricked into ignoring part of the input
                   3867: line.
1.21      crook    3868: 
1.29      crook    3869: @cindex parsing words
                   3870: Words like @code{:} -- words that advance the value of @code{>IN} and so
                   3871: prevent the text interpreter from acting on the whole of the input line
                   3872: -- are called @dfn{parsing words}.
1.21      crook    3873: 
1.29      crook    3874: @cindex @code{state} - effect on the text interpreter
                   3875: @cindex text interpreter - effect of state
                   3876: The second special thing that @code{:} does is change the value of a
                   3877: variable called @code{state}, which affects the way that the text
                   3878: interpreter behaves. When Gforth starts up, @code{state} has the value
                   3879: 0, and the text interpreter is said to be @dfn{interpreting}. During a
                   3880: colon definition (started with @code{:}), @code{state} is set to -1 and
1.44      crook    3881: the text interpreter is said to be @dfn{compiling}.
                   3882: 
                   3883: In this example, the text interpreter is compiling when it processes the
                   3884: string ``@code{2 + . ;}''. It still breaks the string down into
                   3885: character sequences in the same way. However, instead of pushing the
                   3886: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
                   3887: into the definition of @code{add-two} that will make the number @code{2} get
                   3888: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
                   3889: the behaviours of @code{+} and @code{.} are also compiled into the
                   3890: definition.
                   3891: 
                   3892: One category of words don't get compiled. These so-called @dfn{immediate
                   3893: words} get executed (performed @i{now}) regardless of whether the text
                   3894: interpreter is interpreting or compiling. The word @code{;} is an
                   3895: immediate word. Rather than being compiled into the definition, it
                   3896: executes. Its effect is to terminate the current definition, which
                   3897: includes changing the value of @code{state} back to 0.
                   3898: 
                   3899: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
                   3900: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
                   3901: definition.
1.28      crook    3902: 
1.30      anton    3903: In Forth, every word or number can be described in terms of two
1.29      crook    3904: properties:
1.28      crook    3905: 
                   3906: @itemize @bullet
                   3907: @item
1.29      crook    3908: @cindex interpretation semantics
1.44      crook    3909: Its @dfn{interpretation semantics} describe how it will behave when the
                   3910: text interpreter encounters it in @dfn{interpret} state. The
                   3911: interpretation semantics of a word are represented by an @dfn{execution
                   3912: token}.
1.28      crook    3913: @item
1.29      crook    3914: @cindex compilation semantics
1.44      crook    3915: Its @dfn{compilation semantics} describe how it will behave when the
                   3916: text interpreter encounters it in @dfn{compile} state. The compilation
                   3917: semantics of a word are represented in an implementation-dependent way;
                   3918: Gforth uses a @dfn{compilation token}.
1.29      crook    3919: @end itemize
                   3920: 
                   3921: @noindent
                   3922: Numbers are always treated in a fixed way:
                   3923: 
                   3924: @itemize @bullet
1.28      crook    3925: @item
1.44      crook    3926: When the number is @dfn{interpreted}, its behaviour is to push the
                   3927: number onto the stack.
1.28      crook    3928: @item
1.30      anton    3929: When the number is @dfn{compiled}, a piece of code is appended to the
                   3930: current definition that pushes the number when it runs. (In other words,
                   3931: the compilation semantics of a number are to postpone its interpretation
                   3932: semantics until the run-time of the definition that it is being compiled
                   3933: into.)
1.29      crook    3934: @end itemize
                   3935: 
1.44      crook    3936: Words don't behave in such a regular way, but most have @i{default
                   3937: semantics} which means that they behave like this:
1.29      crook    3938: 
                   3939: @itemize @bullet
1.28      crook    3940: @item
1.30      anton    3941: The @dfn{interpretation semantics} of the word are to do something useful.
                   3942: @item
1.29      crook    3943: The @dfn{compilation semantics} of the word are to append its
1.30      anton    3944: @dfn{interpretation semantics} to the current definition (so that its
                   3945: run-time behaviour is to do something useful).
1.28      crook    3946: @end itemize
                   3947: 
1.30      anton    3948: @cindex immediate words
1.44      crook    3949: The actual behaviour of any particular word can be controlled by using
                   3950: the words @code{immediate} and @code{compile-only} when the word is
                   3951: defined. These words set flags in the name dictionary entry of the most
                   3952: recently defined word, and these flags are retrieved by the text
                   3953: interpreter when it finds the word in the name dictionary.
                   3954: 
                   3955: A word that is marked as @dfn{immediate} has compilation semantics that
                   3956: are identical to its interpretation semantics. In other words, it
                   3957: behaves like this:
1.29      crook    3958: 
                   3959: @itemize @bullet
                   3960: @item
1.30      anton    3961: The @dfn{interpretation semantics} of the word are to do something useful.
1.29      crook    3962: @item
1.30      anton    3963: The @dfn{compilation semantics} of the word are to do something useful
                   3964: (and actually the same thing); i.e., it is executed during compilation.
1.29      crook    3965: @end itemize
1.28      crook    3966: 
1.44      crook    3967: Marking a word as @dfn{compile-only} prohibits the text interpreter from
                   3968: performing the interpretation semantics of the word directly; an attempt
                   3969: to do so will generate an error. It is never necessary to use
                   3970: @code{compile-only} (and it is not even part of ANS Forth, though it is
                   3971: provided by many implementations) but it is good etiquette to apply it
                   3972: to a word that will not behave correctly (and might have unexpected
                   3973: side-effects) in interpret state. For example, it is only legal to use
                   3974: the conditional word @code{IF} within a definition. If you forget this
                   3975: and try to use it elsewhere, the fact that (in Gforth) it is marked as
                   3976: @code{compile-only} allows the text interpreter to generate a helpful
                   3977: error message rather than subjecting you to the consequences of your
                   3978: folly.
                   3979: 
1.29      crook    3980: This example shows the difference between an immediate and a
                   3981: non-immediate word:
1.28      crook    3982: 
1.29      crook    3983: @example
                   3984: : show-state state @@ . ;
                   3985: : show-state-now show-state ; immediate
                   3986: : word1 show-state ;
                   3987: : word2 show-state-now ;
1.28      crook    3988: @end example
1.23      crook    3989: 
1.29      crook    3990: The word @code{immediate} after the definition of @code{show-state-now}
                   3991: makes that word an immediate word. These definitions introduce a new
                   3992: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
                   3993: variable, and leaves it on the stack. Therefore, the behaviour of
                   3994: @code{show-state} is to print a number that represents the current value
                   3995: of @code{state}.
1.28      crook    3996: 
1.29      crook    3997: When you execute @code{word1}, it prints the number 0, indicating that
                   3998: the system is interpreting. When the text interpreter compiled the
                   3999: definition of @code{word1}, it encountered @code{show-state} whose
1.30      anton    4000: compilation semantics are to append its interpretation semantics to the
1.29      crook    4001: current definition. When you execute @code{word1}, it performs the
1.30      anton    4002: interpretation semantics of @code{show-state}.  At the time that @code{word1}
1.29      crook    4003: (and therefore @code{show-state}) are executed, the system is
                   4004: interpreting.
1.28      crook    4005: 
1.30      anton    4006: When you pressed @key{RET} after entering the definition of @code{word2},
1.29      crook    4007: you should have seen the number -1 printed, followed by ``@code{
                   4008: ok}''. When the text interpreter compiled the definition of
                   4009: @code{word2}, it encountered @code{show-state-now}, an immediate word,
1.30      anton    4010: whose compilation semantics are therefore to perform its interpretation
1.29      crook    4011: semantics. It is executed straight away (even before the text
                   4012: interpreter has moved on to process another group of characters; the
                   4013: @code{;} in this example). The effect of executing it are to display the
                   4014: value of @code{state} @i{at the time that the definition of}
                   4015: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
                   4016: system is compiling at this time. If you execute @code{word2} it does
                   4017: nothing at all.
1.28      crook    4018: 
1.29      crook    4019: @cindex @code{."}, how it works
                   4020: Before leaving the subject of immediate words, consider the behaviour of
                   4021: @code{."} in the definition of @code{greet}, in the previous
                   4022: section. This word is both a parsing word and an immediate word. Notice
                   4023: that there is a space between @code{."} and the start of the text
                   4024: @code{Hello and welcome}, but that there is no space between the last
                   4025: letter of @code{welcome} and the @code{"} character. The reason for this
                   4026: is that @code{."} is a Forth word; it must have a space after it so that
                   4027: the text interpreter can identify it. The @code{"} is not a Forth word;
                   4028: it is a @dfn{delimiter}. The examples earlier show that, when the string
                   4029: is displayed, there is neither a space before the @code{H} nor after the
                   4030: @code{e}. Since @code{."} is an immediate word, it executes at the time
                   4031: that @code{greet} is defined. When it executes, its behaviour is to
                   4032: search forward in the input line looking for the delimiter. When it
                   4033: finds the delimiter, it updates @code{>IN} to point past the
                   4034: delimiter. It also compiles some magic code into the definition of
                   4035: @code{greet}; the xt of a run-time routine that prints a text string. It
                   4036: compiles the string @code{Hello and welcome} into memory so that it is
                   4037: available to be printed later. When the text interpreter gains control,
                   4038: the next word it finds in the input stream is @code{;} and so it
                   4039: terminates the definition of @code{greet}.
1.28      crook    4040: 
                   4041: 
                   4042: @comment ----------------------------------------------
1.29      crook    4043: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
                   4044: @section Forth is written in Forth
                   4045: @cindex structure of Forth programs
                   4046: 
                   4047: When you start up a Forth compiler, a large number of definitions
                   4048: already exist. In Forth, you develop a new application using bottom-up
                   4049: programming techniques to create new definitions that are defined in
                   4050: terms of existing definitions. As you create each definition you can
                   4051: test and debug it interactively.
                   4052: 
                   4053: If you have tried out the examples in this section, you will probably
                   4054: have typed them in by hand; when you leave Gforth, your definitions will
                   4055: be lost. You can avoid this by using a text editor to enter Forth source
                   4056: code into a file, and then loading code from the file using
1.49      anton    4057: @code{include} (@pxref{Forth source files}). A Forth source file is
1.29      crook    4058: processed by the text interpreter, just as though you had typed it in by
                   4059: hand@footnote{Actually, there are some subtle differences -- see
                   4060: @ref{The Text Interpreter}.}.
                   4061: 
                   4062: Gforth also supports the traditional Forth alternative to using text
1.49      anton    4063: files for program entry (@pxref{Blocks}).
1.28      crook    4064: 
1.29      crook    4065: In common with many, if not most, Forth compilers, most of Gforth is
                   4066: actually written in Forth. All of the @file{.fs} files in the
                   4067: installation directory@footnote{For example,
1.30      anton    4068: @file{/usr/local/share/gforth...}} are Forth source files, which you can
1.29      crook    4069: study to see examples of Forth programming.
1.28      crook    4070: 
1.29      crook    4071: Gforth maintains a history file that records every line that you type to
                   4072: the text interpreter. This file is preserved between sessions, and is
                   4073: used to provide a command-line recall facility. If you enter long
                   4074: definitions by hand, you can use a text editor to paste them out of the
                   4075: history file into a Forth source file for reuse at a later time
1.49      anton    4076: (for more information @pxref{Command-line editing}).
1.28      crook    4077: 
                   4078: 
                   4079: @comment ----------------------------------------------
1.29      crook    4080: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
                   4081: @section Review - elements of a Forth system
                   4082: @cindex elements of a Forth system
1.28      crook    4083: 
1.29      crook    4084: To summarise this chapter:
1.28      crook    4085: 
                   4086: @itemize @bullet
                   4087: @item
1.29      crook    4088: Forth programs use @dfn{factoring} to break a problem down into small
                   4089: fragments called @dfn{words} or @dfn{definitions}.
                   4090: @item
                   4091: Forth program development is an interactive process.
                   4092: @item
                   4093: The main command loop that accepts input, and controls both
                   4094: interpretation and compilation, is called the @dfn{text interpreter}
                   4095: (also known as the @dfn{outer interpreter}).
                   4096: @item
                   4097: Forth has a very simple syntax, consisting of words and numbers
                   4098: separated by spaces or carriage-return characters. Any additional syntax
                   4099: is imposed by @dfn{parsing words}.
                   4100: @item
                   4101: Forth uses a stack to pass parameters between words. As a result, it
                   4102: uses postfix notation.
                   4103: @item
                   4104: To use a word that has previously been defined, the text interpreter
                   4105: searches for the word in the @dfn{name dictionary}.
                   4106: @item
1.30      anton    4107: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
1.28      crook    4108: @item
1.29      crook    4109: The text interpreter uses the value of @code{state} to select between
                   4110: the use of the @dfn{interpretation semantics} and the  @dfn{compilation
                   4111: semantics} of a word that it encounters.
1.28      crook    4112: @item
1.30      anton    4113: The relationship between the @dfn{interpretation semantics} and
                   4114: @dfn{compilation semantics} for a word
1.29      crook    4115: depend upon the way in which the word was defined (for example, whether
                   4116: it is an @dfn{immediate} word).
1.28      crook    4117: @item
1.29      crook    4118: Forth definitions can be implemented in Forth (called @dfn{high-level
                   4119: definitions}) or in some other way (usually a lower-level language and
                   4120: as a result often called @dfn{low-level definitions}, @dfn{code
                   4121: definitions} or @dfn{primitives}).
1.28      crook    4122: @item
1.29      crook    4123: Many Forth systems are implemented mainly in Forth.
1.28      crook    4124: @end itemize
                   4125: 
                   4126: 
1.29      crook    4127: @comment ----------------------------------------------
1.48      anton    4128: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
1.29      crook    4129: @section Where To Go Next
                   4130: @cindex where to go next
1.28      crook    4131: 
1.29      crook    4132: Amazing as it may seem, if you have read (and understood) this far, you
                   4133: know almost all the fundamentals about the inner workings of a Forth
                   4134: system. You certainly know enough to be able to read and understand the
                   4135: rest of this manual and the ANS Forth document, to learn more about the
                   4136: facilities that Forth in general and Gforth in particular provide. Even
                   4137: scarier, you know almost enough to implement your own Forth system.
1.30      anton    4138: However, that's not a good idea just yet... better to try writing some
1.29      crook    4139: programs in Gforth.
1.28      crook    4140: 
1.29      crook    4141: Forth has such a rich vocabulary that it can be hard to know where to
                   4142: start in learning it. This section suggests a few sets of words that are
                   4143: enough to write small but useful programs. Use the word index in this
                   4144: document to learn more about each word, then try it out and try to write
                   4145: small definitions using it. Start by experimenting with these words:
1.28      crook    4146: 
                   4147: @itemize @bullet
                   4148: @item
1.29      crook    4149: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
                   4150: @item
                   4151: Comparison: @code{MIN MAX =}
                   4152: @item
                   4153: Logic: @code{AND OR XOR NOT}
                   4154: @item
                   4155: Stack manipulation: @code{DUP DROP SWAP OVER}
1.28      crook    4156: @item
1.29      crook    4157: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
1.28      crook    4158: @item
1.29      crook    4159: Input/Output: @code{. ." EMIT CR KEY}
1.28      crook    4160: @item
1.29      crook    4161: Defining words: @code{: ; CREATE}
1.28      crook    4162: @item
1.29      crook    4163: Memory allocation words: @code{ALLOT ,}
1.28      crook    4164: @item
1.29      crook    4165: Tools: @code{SEE WORDS .S MARKER}
                   4166: @end itemize
                   4167: 
                   4168: When you have mastered those, go on to:
                   4169: 
                   4170: @itemize @bullet
1.28      crook    4171: @item
1.29      crook    4172: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
1.28      crook    4173: @item
1.29      crook    4174: Memory access: @code{@@ !}
1.28      crook    4175: @end itemize
1.23      crook    4176: 
1.29      crook    4177: When you have mastered these, there's nothing for it but to read through
                   4178: the whole of this manual and find out what you've missed.
                   4179: 
                   4180: @comment ----------------------------------------------
1.48      anton    4181: @node Exercises,  , Where to go next, Introduction
1.29      crook    4182: @section Exercises
                   4183: @cindex exercises
                   4184: 
                   4185: TODO: provide a set of programming excercises linked into the stuff done
                   4186: already and into other sections of the manual. Provide solutions to all
                   4187: the exercises in a .fs file in the distribution.
                   4188: 
                   4189: @c Get some inspiration from Starting Forth and Kelly&Spies.
                   4190: 
                   4191: @c excercises:
                   4192: @c 1. take inches and convert to feet and inches.
                   4193: @c 2. take temperature and convert from fahrenheight to celcius;
                   4194: @c    may need to care about symmetric vs floored??
                   4195: @c 3. take input line and do character substitution
                   4196: @c    to encipher or decipher
                   4197: @c 4. as above but work on a file for in and out
                   4198: @c 5. take input line and convert to pig-latin 
                   4199: @c
                   4200: @c thing of sets of things to exercise then come up with
                   4201: @c problems that need those things.
                   4202: 
                   4203: 
1.26      crook    4204: @c ******************************************************************
1.29      crook    4205: @node Words, Error messages, Introduction, Top
1.1       anton    4206: @chapter Forth Words
1.26      crook    4207: @cindex words
1.1       anton    4208: 
                   4209: @menu
                   4210: * Notation::                    
1.65      anton    4211: * Case insensitivity::          
                   4212: * Comments::                    
                   4213: * Boolean Flags::               
1.1       anton    4214: * Arithmetic::                  
                   4215: * Stack Manipulation::          
1.5       anton    4216: * Memory::                      
1.1       anton    4217: * Control Structures::          
                   4218: * Defining Words::              
1.65      anton    4219: * Interpretation and Compilation Semantics::  
1.47      crook    4220: * Tokens for Words::            
1.81      anton    4221: * Compiling words::             
1.65      anton    4222: * The Text Interpreter::        
1.111     anton    4223: * The Input Stream::            
1.65      anton    4224: * Word Lists::                  
                   4225: * Environmental Queries::       
1.12      anton    4226: * Files::                       
                   4227: * Blocks::                      
                   4228: * Other I/O::                   
1.121     anton    4229: * OS command line arguments::   
1.78      anton    4230: * Locals::                      
                   4231: * Structures::                  
                   4232: * Object-oriented Forth::       
1.12      anton    4233: * Programming Tools::           
                   4234: * Assembler and Code Words::    
                   4235: * Threading Words::             
1.65      anton    4236: * Passing Commands to the OS::  
                   4237: * Keeping track of Time::       
                   4238: * Miscellaneous Words::         
1.1       anton    4239: @end menu
                   4240: 
1.65      anton    4241: @node Notation, Case insensitivity, Words, Words
1.1       anton    4242: @section Notation
                   4243: @cindex notation of glossary entries
                   4244: @cindex format of glossary entries
                   4245: @cindex glossary notation format
                   4246: @cindex word glossary entry format
                   4247: 
                   4248: The Forth words are described in this section in the glossary notation
1.67      anton    4249: that has become a de-facto standard for Forth texts:
1.1       anton    4250: 
                   4251: @format
1.29      crook    4252: @i{word}     @i{Stack effect}   @i{wordset}   @i{pronunciation}
1.1       anton    4253: @end format
1.29      crook    4254: @i{Description}
1.1       anton    4255: 
                   4256: @table @var
                   4257: @item word
1.28      crook    4258: The name of the word.
1.1       anton    4259: 
                   4260: @item Stack effect
                   4261: @cindex stack effect
1.29      crook    4262: The stack effect is written in the notation @code{@i{before} --
                   4263: @i{after}}, where @i{before} and @i{after} describe the top of
1.1       anton    4264: stack entries before and after the execution of the word. The rest of
                   4265: the stack is not touched by the word. The top of stack is rightmost,
                   4266: i.e., a stack sequence is written as it is typed in. Note that Gforth
                   4267: uses a separate floating point stack, but a unified stack
1.29      crook    4268: notation. Also, return stack effects are not shown in @i{stack
                   4269: effect}, but in @i{Description}. The name of a stack item describes
1.1       anton    4270: the type and/or the function of the item. See below for a discussion of
                   4271: the types.
                   4272: 
                   4273: All words have two stack effects: A compile-time stack effect and a
                   4274: run-time stack effect. The compile-time stack-effect of most words is
1.29      crook    4275: @i{ -- }. If the compile-time stack-effect of a word deviates from
1.1       anton    4276: this standard behaviour, or the word does other unusual things at
                   4277: compile time, both stack effects are shown; otherwise only the run-time
                   4278: stack effect is shown.
                   4279: 
                   4280: @cindex pronounciation of words
                   4281: @item pronunciation
                   4282: How the word is pronounced.
                   4283: 
                   4284: @cindex wordset
1.67      anton    4285: @cindex environment wordset
1.1       anton    4286: @item wordset
1.21      crook    4287: The ANS Forth standard is divided into several word sets. A standard
                   4288: system need not support all of them. Therefore, in theory, the fewer
                   4289: word sets your program uses the more portable it will be. However, we
                   4290: suspect that most ANS Forth systems on personal machines will feature
1.26      crook    4291: all word sets. Words that are not defined in ANS Forth have
1.21      crook    4292: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
1.1       anton    4293: describes words that will work in future releases of Gforth;
                   4294: @code{gforth-internal} words are more volatile. Environmental query
                   4295: strings are also displayed like words; you can recognize them by the
1.21      crook    4296: @code{environment} in the word set field.
1.1       anton    4297: 
                   4298: @item Description
                   4299: A description of the behaviour of the word.
                   4300: @end table
                   4301: 
                   4302: @cindex types of stack items
                   4303: @cindex stack item types
                   4304: The type of a stack item is specified by the character(s) the name
                   4305: starts with:
                   4306: 
                   4307: @table @code
                   4308: @item f
                   4309: @cindex @code{f}, stack item type
                   4310: Boolean flags, i.e. @code{false} or @code{true}.
                   4311: @item c
                   4312: @cindex @code{c}, stack item type
                   4313: Char
                   4314: @item w
                   4315: @cindex @code{w}, stack item type
                   4316: Cell, can contain an integer or an address
                   4317: @item n
                   4318: @cindex @code{n}, stack item type
                   4319: signed integer
                   4320: @item u
                   4321: @cindex @code{u}, stack item type
                   4322: unsigned integer
                   4323: @item d
                   4324: @cindex @code{d}, stack item type
                   4325: double sized signed integer
                   4326: @item ud
                   4327: @cindex @code{ud}, stack item type
                   4328: double sized unsigned integer
                   4329: @item r
                   4330: @cindex @code{r}, stack item type
                   4331: Float (on the FP stack)
1.21      crook    4332: @item a-
1.1       anton    4333: @cindex @code{a_}, stack item type
                   4334: Cell-aligned address
1.21      crook    4335: @item c-
1.1       anton    4336: @cindex @code{c_}, stack item type
                   4337: Char-aligned address (note that a Char may have two bytes in Windows NT)
1.21      crook    4338: @item f-
1.1       anton    4339: @cindex @code{f_}, stack item type
                   4340: Float-aligned address
1.21      crook    4341: @item df-
1.1       anton    4342: @cindex @code{df_}, stack item type
                   4343: Address aligned for IEEE double precision float
1.21      crook    4344: @item sf-
1.1       anton    4345: @cindex @code{sf_}, stack item type
                   4346: Address aligned for IEEE single precision float
                   4347: @item xt
                   4348: @cindex @code{xt}, stack item type
                   4349: Execution token, same size as Cell
                   4350: @item wid
                   4351: @cindex @code{wid}, stack item type
1.21      crook    4352: Word list ID, same size as Cell
1.68      anton    4353: @item ior, wior
                   4354: @cindex ior type description
                   4355: @cindex wior type description
                   4356: I/O result code, cell-sized.  In Gforth, you can @code{throw} iors.
1.1       anton    4357: @item f83name
                   4358: @cindex @code{f83name}, stack item type
                   4359: Pointer to a name structure
                   4360: @item "
                   4361: @cindex @code{"}, stack item type
1.12      anton    4362: string in the input stream (not on the stack). The terminating character
                   4363: is a blank by default. If it is not a blank, it is shown in @code{<>}
1.1       anton    4364: quotes.
                   4365: @end table
                   4366: 
1.65      anton    4367: @comment ----------------------------------------------
                   4368: @node Case insensitivity, Comments, Notation, Words
                   4369: @section Case insensitivity
                   4370: @cindex case sensitivity
                   4371: @cindex upper and lower case
                   4372: 
                   4373: Gforth is case-insensitive; you can enter definitions and invoke
                   4374: Standard words using upper, lower or mixed case (however,
                   4375: @pxref{core-idef, Implementation-defined options, Implementation-defined
                   4376: options}).
                   4377: 
                   4378: ANS Forth only @i{requires} implementations to recognise Standard words
                   4379: when they are typed entirely in upper case. Therefore, a Standard
                   4380: program must use upper case for all Standard words. You can use whatever
                   4381: case you like for words that you define, but in a Standard program you
                   4382: have to use the words in the same case that you defined them.
                   4383: 
                   4384: Gforth supports case sensitivity through @code{table}s (case-sensitive
                   4385: wordlists, @pxref{Word Lists}).
                   4386: 
                   4387: Two people have asked how to convert Gforth to be case-sensitive; while
                   4388: we think this is a bad idea, you can change all wordlists into tables
                   4389: like this:
                   4390: 
                   4391: @example
                   4392: ' table-find forth-wordlist wordlist-map @ !
                   4393: @end example
                   4394: 
                   4395: Note that you now have to type the predefined words in the same case
                   4396: that we defined them, which are varying.  You may want to convert them
                   4397: to your favourite case before doing this operation (I won't explain how,
                   4398: because if you are even contemplating doing this, you'd better have
                   4399: enough knowledge of Forth systems to know this already).
                   4400: 
                   4401: @node Comments, Boolean Flags, Case insensitivity, Words
1.21      crook    4402: @section Comments
1.26      crook    4403: @cindex comments
1.21      crook    4404: 
1.29      crook    4405: Forth supports two styles of comment; the traditional @i{in-line} comment,
                   4406: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
1.21      crook    4407: 
1.44      crook    4408: 
1.23      crook    4409: doc-(
1.21      crook    4410: doc-\
1.23      crook    4411: doc-\G
1.21      crook    4412: 
1.44      crook    4413: 
1.21      crook    4414: @node Boolean Flags, Arithmetic, Comments, Words
                   4415: @section Boolean Flags
1.26      crook    4416: @cindex Boolean flags
1.21      crook    4417: 
                   4418: A Boolean flag is cell-sized. A cell with all bits clear represents the
                   4419: flag @code{false} and a flag with all bits set represents the flag
1.26      crook    4420: @code{true}. Words that check a flag (for example, @code{IF}) will treat
1.29      crook    4421: a cell that has @i{any} bit set as @code{true}.
1.67      anton    4422: @c on and off to Memory? 
                   4423: @c true and false to "Bitwise operations" or "Numeric comparison"?
1.44      crook    4424: 
1.21      crook    4425: doc-true
                   4426: doc-false
1.29      crook    4427: doc-on
                   4428: doc-off
1.21      crook    4429: 
1.44      crook    4430: 
1.21      crook    4431: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
1.1       anton    4432: @section Arithmetic
                   4433: @cindex arithmetic words
                   4434: 
                   4435: @cindex division with potentially negative operands
                   4436: Forth arithmetic is not checked, i.e., you will not hear about integer
                   4437: overflow on addition or multiplication, you may hear about division by
                   4438: zero if you are lucky. The operator is written after the operands, but
                   4439: the operands are still in the original order. I.e., the infix @code{2-1}
                   4440: corresponds to @code{2 1 -}. Forth offers a variety of division
                   4441: operators. If you perform division with potentially negative operands,
                   4442: you do not want to use @code{/} or @code{/mod} with its undefined
                   4443: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
                   4444: former, @pxref{Mixed precision}).
1.26      crook    4445: @comment TODO discuss the different division forms and the std approach
1.1       anton    4446: 
                   4447: @menu
                   4448: * Single precision::            
1.67      anton    4449: * Double precision::            Double-cell integer arithmetic
1.1       anton    4450: * Bitwise operations::          
1.67      anton    4451: * Numeric comparison::          
1.29      crook    4452: * Mixed precision::             Operations with single and double-cell integers
1.1       anton    4453: * Floating Point::              
                   4454: @end menu
                   4455: 
1.67      anton    4456: @node Single precision, Double precision, Arithmetic, Arithmetic
1.1       anton    4457: @subsection Single precision
                   4458: @cindex single precision arithmetic words
                   4459: 
1.67      anton    4460: @c !! cell undefined
                   4461: 
                   4462: By default, numbers in Forth are single-precision integers that are one
1.26      crook    4463: cell in size. They can be signed or unsigned, depending upon how you
1.49      anton    4464: treat them. For the rules used by the text interpreter for recognising
                   4465: single-precision integers see @ref{Number Conversion}.
1.21      crook    4466: 
1.67      anton    4467: These words are all defined for signed operands, but some of them also
                   4468: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
                   4469: @code{*}.
1.44      crook    4470: 
1.1       anton    4471: doc-+
1.21      crook    4472: doc-1+
1.128     anton    4473: doc-under+
1.1       anton    4474: doc--
1.21      crook    4475: doc-1-
1.1       anton    4476: doc-*
                   4477: doc-/
                   4478: doc-mod
                   4479: doc-/mod
                   4480: doc-negate
                   4481: doc-abs
                   4482: doc-min
                   4483: doc-max
1.27      crook    4484: doc-floored
1.1       anton    4485: 
1.44      crook    4486: 
1.67      anton    4487: @node Double precision, Bitwise operations, Single precision, Arithmetic
1.21      crook    4488: @subsection Double precision
                   4489: @cindex double precision arithmetic words
                   4490: 
1.49      anton    4491: For the rules used by the text interpreter for
                   4492: recognising double-precision integers, see @ref{Number Conversion}.
1.21      crook    4493: 
                   4494: A double precision number is represented by a cell pair, with the most
1.67      anton    4495: significant cell at the TOS. It is trivial to convert an unsigned single
                   4496: to a double: simply push a @code{0} onto the TOS. Since numbers are
                   4497: represented by Gforth using 2's complement arithmetic, converting a
                   4498: signed single to a (signed) double requires sign-extension across the
                   4499: most significant cell. This can be achieved using @code{s>d}. The moral
                   4500: of the story is that you cannot convert a number without knowing whether
                   4501: it represents an unsigned or a signed number.
1.21      crook    4502: 
1.67      anton    4503: These words are all defined for signed operands, but some of them also
                   4504: work for unsigned numbers: @code{d+}, @code{d-}.
1.44      crook    4505: 
1.21      crook    4506: doc-s>d
1.67      anton    4507: doc-d>s
1.21      crook    4508: doc-d+
                   4509: doc-d-
                   4510: doc-dnegate
                   4511: doc-dabs
                   4512: doc-dmin
                   4513: doc-dmax
                   4514: 
1.44      crook    4515: 
1.67      anton    4516: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
                   4517: @subsection Bitwise operations
                   4518: @cindex bitwise operation words
                   4519: 
                   4520: 
                   4521: doc-and
                   4522: doc-or
                   4523: doc-xor
                   4524: doc-invert
                   4525: doc-lshift
                   4526: doc-rshift
                   4527: doc-2*
                   4528: doc-d2*
                   4529: doc-2/
                   4530: doc-d2/
                   4531: 
                   4532: 
                   4533: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
1.21      crook    4534: @subsection Numeric comparison
                   4535: @cindex numeric comparison words
                   4536: 
1.67      anton    4537: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
                   4538: d0= d0<>}) work for for both signed and unsigned numbers.
1.44      crook    4539: 
1.28      crook    4540: doc-<
                   4541: doc-<=
                   4542: doc-<>
                   4543: doc-=
                   4544: doc->
                   4545: doc->=
                   4546: 
1.21      crook    4547: doc-0<
1.23      crook    4548: doc-0<=
1.21      crook    4549: doc-0<>
                   4550: doc-0=
1.23      crook    4551: doc-0>
                   4552: doc-0>=
1.28      crook    4553: 
                   4554: doc-u<
                   4555: doc-u<=
1.44      crook    4556: @c u<> and u= exist but are the same as <> and =
1.31      anton    4557: @c doc-u<>
                   4558: @c doc-u=
1.28      crook    4559: doc-u>
                   4560: doc-u>=
                   4561: 
                   4562: doc-within
                   4563: 
                   4564: doc-d<
                   4565: doc-d<=
                   4566: doc-d<>
                   4567: doc-d=
                   4568: doc-d>
                   4569: doc-d>=
1.23      crook    4570: 
1.21      crook    4571: doc-d0<
1.23      crook    4572: doc-d0<=
                   4573: doc-d0<>
1.21      crook    4574: doc-d0=
1.23      crook    4575: doc-d0>
                   4576: doc-d0>=
                   4577: 
1.21      crook    4578: doc-du<
1.28      crook    4579: doc-du<=
1.44      crook    4580: @c du<> and du= exist but are the same as d<> and d=
1.31      anton    4581: @c doc-du<>
                   4582: @c doc-du=
1.28      crook    4583: doc-du>
                   4584: doc-du>=
1.1       anton    4585: 
1.44      crook    4586: 
1.21      crook    4587: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
1.1       anton    4588: @subsection Mixed precision
                   4589: @cindex mixed precision arithmetic words
                   4590: 
1.44      crook    4591: 
1.1       anton    4592: doc-m+
                   4593: doc-*/
                   4594: doc-*/mod
                   4595: doc-m*
                   4596: doc-um*
                   4597: doc-m*/
                   4598: doc-um/mod
                   4599: doc-fm/mod
                   4600: doc-sm/rem
                   4601: 
1.44      crook    4602: 
1.21      crook    4603: @node Floating Point,  , Mixed precision, Arithmetic
1.1       anton    4604: @subsection Floating Point
                   4605: @cindex floating point arithmetic words
                   4606: 
1.49      anton    4607: For the rules used by the text interpreter for
                   4608: recognising floating-point numbers see @ref{Number Conversion}.
1.1       anton    4609: 
1.67      anton    4610: Gforth has a separate floating point stack, but the documentation uses
                   4611: the unified notation.@footnote{It's easy to generate the separate
                   4612: notation from that by just separating the floating-point numbers out:
                   4613: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
                   4614: r3 )}.}
1.1       anton    4615: 
                   4616: @cindex floating-point arithmetic, pitfalls
                   4617: Floating point numbers have a number of unpleasant surprises for the
                   4618: unwary (e.g., floating point addition is not associative) and even a few
                   4619: for the wary. You should not use them unless you know what you are doing
                   4620: or you don't care that the results you get are totally bogus. If you
                   4621: want to learn about the problems of floating point numbers (and how to
1.66      anton    4622: avoid them), you might start with @cite{David Goldberg,
                   4623: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
                   4624: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
                   4625: Surveys 23(1):5@minus{}48, March 1991}.
1.1       anton    4626: 
1.44      crook    4627: 
1.21      crook    4628: doc-d>f
                   4629: doc-f>d
1.1       anton    4630: doc-f+
                   4631: doc-f-
                   4632: doc-f*
                   4633: doc-f/
                   4634: doc-fnegate
                   4635: doc-fabs
                   4636: doc-fmax
                   4637: doc-fmin
                   4638: doc-floor
                   4639: doc-fround
                   4640: doc-f**
                   4641: doc-fsqrt
                   4642: doc-fexp
                   4643: doc-fexpm1
                   4644: doc-fln
                   4645: doc-flnp1
                   4646: doc-flog
                   4647: doc-falog
1.32      anton    4648: doc-f2*
                   4649: doc-f2/
                   4650: doc-1/f
                   4651: doc-precision
                   4652: doc-set-precision
                   4653: 
                   4654: @cindex angles in trigonometric operations
                   4655: @cindex trigonometric operations
                   4656: Angles in floating point operations are given in radians (a full circle
                   4657: has 2 pi radians).
                   4658: 
1.1       anton    4659: doc-fsin
                   4660: doc-fcos
                   4661: doc-fsincos
                   4662: doc-ftan
                   4663: doc-fasin
                   4664: doc-facos
                   4665: doc-fatan
                   4666: doc-fatan2
                   4667: doc-fsinh
                   4668: doc-fcosh
                   4669: doc-ftanh
                   4670: doc-fasinh
                   4671: doc-facosh
                   4672: doc-fatanh
1.21      crook    4673: doc-pi
1.28      crook    4674: 
1.32      anton    4675: @cindex equality of floats
                   4676: @cindex floating-point comparisons
1.31      anton    4677: One particular problem with floating-point arithmetic is that comparison
                   4678: for equality often fails when you would expect it to succeed.  For this
                   4679: reason approximate equality is often preferred (but you still have to
1.67      anton    4680: know what you are doing).  Also note that IEEE NaNs may compare
1.68      anton    4681: differently from what you might expect.  The comparison words are:
1.31      anton    4682: 
                   4683: doc-f~rel
                   4684: doc-f~abs
1.68      anton    4685: doc-f~
1.31      anton    4686: doc-f=
                   4687: doc-f<>
                   4688: 
                   4689: doc-f<
                   4690: doc-f<=
                   4691: doc-f>
                   4692: doc-f>=
                   4693: 
1.21      crook    4694: doc-f0<
1.28      crook    4695: doc-f0<=
                   4696: doc-f0<>
1.21      crook    4697: doc-f0=
1.28      crook    4698: doc-f0>
                   4699: doc-f0>=
                   4700: 
1.1       anton    4701: 
                   4702: @node Stack Manipulation, Memory, Arithmetic, Words
                   4703: @section Stack Manipulation
                   4704: @cindex stack manipulation words
                   4705: 
                   4706: @cindex floating-point stack in the standard
1.21      crook    4707: Gforth maintains a number of separate stacks:
                   4708: 
1.29      crook    4709: @cindex data stack
                   4710: @cindex parameter stack
1.21      crook    4711: @itemize @bullet
                   4712: @item
1.29      crook    4713: A data stack (also known as the @dfn{parameter stack}) -- for
                   4714: characters, cells, addresses, and double cells.
1.21      crook    4715: 
1.29      crook    4716: @cindex floating-point stack
1.21      crook    4717: @item
1.44      crook    4718: A floating point stack -- for holding floating point (FP) numbers.
1.21      crook    4719: 
1.29      crook    4720: @cindex return stack
1.21      crook    4721: @item
1.44      crook    4722: A return stack -- for holding the return addresses of colon
1.32      anton    4723: definitions and other (non-FP) data.
1.21      crook    4724: 
1.29      crook    4725: @cindex locals stack
1.21      crook    4726: @item
1.44      crook    4727: A locals stack -- for holding local variables.
1.21      crook    4728: @end itemize
                   4729: 
1.1       anton    4730: @menu
                   4731: * Data stack::                  
                   4732: * Floating point stack::        
                   4733: * Return stack::                
                   4734: * Locals stack::                
                   4735: * Stack pointer manipulation::  
                   4736: @end menu
                   4737: 
                   4738: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
                   4739: @subsection Data stack
                   4740: @cindex data stack manipulation words
                   4741: @cindex stack manipulations words, data stack
                   4742: 
1.44      crook    4743: 
1.1       anton    4744: doc-drop
                   4745: doc-nip
                   4746: doc-dup
                   4747: doc-over
                   4748: doc-tuck
                   4749: doc-swap
1.21      crook    4750: doc-pick
1.1       anton    4751: doc-rot
                   4752: doc--rot
                   4753: doc-?dup
                   4754: doc-roll
                   4755: doc-2drop
                   4756: doc-2nip
                   4757: doc-2dup
                   4758: doc-2over
                   4759: doc-2tuck
                   4760: doc-2swap
                   4761: doc-2rot
                   4762: 
1.44      crook    4763: 
1.1       anton    4764: @node Floating point stack, Return stack, Data stack, Stack Manipulation
                   4765: @subsection Floating point stack
                   4766: @cindex floating-point stack manipulation words
                   4767: @cindex stack manipulation words, floating-point stack
                   4768: 
1.32      anton    4769: Whilst every sane Forth has a separate floating-point stack, it is not
                   4770: strictly required; an ANS Forth system could theoretically keep
                   4771: floating-point numbers on the data stack. As an additional difficulty,
                   4772: you don't know how many cells a floating-point number takes. It is
                   4773: reportedly possible to write words in a way that they work also for a
                   4774: unified stack model, but we do not recommend trying it. Instead, just
                   4775: say that your program has an environmental dependency on a separate
                   4776: floating-point stack.
                   4777: 
                   4778: doc-floating-stack
                   4779: 
1.1       anton    4780: doc-fdrop
                   4781: doc-fnip
                   4782: doc-fdup
                   4783: doc-fover
                   4784: doc-ftuck
                   4785: doc-fswap
1.21      crook    4786: doc-fpick
1.1       anton    4787: doc-frot
                   4788: 
1.44      crook    4789: 
1.1       anton    4790: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
                   4791: @subsection Return stack
                   4792: @cindex return stack manipulation words
                   4793: @cindex stack manipulation words, return stack
                   4794: 
1.32      anton    4795: @cindex return stack and locals
                   4796: @cindex locals and return stack
                   4797: A Forth system is allowed to keep local variables on the
                   4798: return stack. This is reasonable, as local variables usually eliminate
                   4799: the need to use the return stack explicitly. So, if you want to produce
                   4800: a standard compliant program and you are using local variables in a
                   4801: word, forget about return stack manipulations in that word (refer to the
                   4802: standard document for the exact rules).
                   4803: 
1.1       anton    4804: doc->r
                   4805: doc-r>
                   4806: doc-r@
                   4807: doc-rdrop
                   4808: doc-2>r
                   4809: doc-2r>
                   4810: doc-2r@
                   4811: doc-2rdrop
                   4812: 
1.44      crook    4813: 
1.1       anton    4814: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
                   4815: @subsection Locals stack
                   4816: 
1.78      anton    4817: Gforth uses an extra locals stack.  It is described, along with the
                   4818: reasons for its existence, in @ref{Locals implementation}.
1.21      crook    4819: 
1.1       anton    4820: @node Stack pointer manipulation,  , Locals stack, Stack Manipulation
                   4821: @subsection Stack pointer manipulation
                   4822: @cindex stack pointer manipulation words
                   4823: 
1.44      crook    4824: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
1.21      crook    4825: doc-sp0
1.1       anton    4826: doc-sp@
                   4827: doc-sp!
1.21      crook    4828: doc-fp0
1.1       anton    4829: doc-fp@
                   4830: doc-fp!
1.21      crook    4831: doc-rp0
1.1       anton    4832: doc-rp@
                   4833: doc-rp!
1.21      crook    4834: doc-lp0
1.1       anton    4835: doc-lp@
                   4836: doc-lp!
                   4837: 
1.44      crook    4838: 
1.1       anton    4839: @node Memory, Control Structures, Stack Manipulation, Words
                   4840: @section Memory
1.26      crook    4841: @cindex memory words
1.1       anton    4842: 
1.32      anton    4843: @menu
                   4844: * Memory model::                
                   4845: * Dictionary allocation::       
                   4846: * Heap Allocation::             
                   4847: * Memory Access::               
                   4848: * Address arithmetic::          
                   4849: * Memory Blocks::               
                   4850: @end menu
                   4851: 
1.67      anton    4852: In addition to the standard Forth memory allocation words, there is also
                   4853: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
                   4854: garbage collector}.
                   4855: 
1.32      anton    4856: @node Memory model, Dictionary allocation, Memory, Memory
                   4857: @subsection ANS Forth and Gforth memory models
                   4858: 
                   4859: @c The ANS Forth description is a mess (e.g., is the heap part of
                   4860: @c the dictionary?), so let's not stick to closely with it.
                   4861: 
1.67      anton    4862: ANS Forth considers a Forth system as consisting of several address
                   4863: spaces, of which only @dfn{data space} is managed and accessible with
                   4864: the memory words.  Memory not necessarily in data space includes the
                   4865: stacks, the code (called code space) and the headers (called name
                   4866: space). In Gforth everything is in data space, but the code for the
                   4867: primitives is usually read-only.
1.32      anton    4868: 
                   4869: Data space is divided into a number of areas: The (data space portion of
                   4870: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
                   4871: refer to the search data structure embodied in word lists and headers,
                   4872: because it is used for looking up names, just as you would in a
                   4873: conventional dictionary.}, the heap, and a number of system-allocated
                   4874: buffers.
                   4875: 
1.68      anton    4876: @cindex address arithmetic restrictions, ANS vs. Gforth
                   4877: @cindex contiguous regions, ANS vs. Gforth
1.32      anton    4878: In ANS Forth data space is also divided into contiguous regions.  You
                   4879: can only use address arithmetic within a contiguous region, not between
                   4880: them.  Usually each allocation gives you one contiguous region, but the
1.33      anton    4881: dictionary allocation words have additional rules (@pxref{Dictionary
1.32      anton    4882: allocation}).
                   4883: 
                   4884: Gforth provides one big address space, and address arithmetic can be
                   4885: performed between any addresses. However, in the dictionary headers or
                   4886: code are interleaved with data, so almost the only contiguous data space
                   4887: regions there are those described by ANS Forth as contiguous; but you
                   4888: can be sure that the dictionary is allocated towards increasing
                   4889: addresses even between contiguous regions.  The memory order of
                   4890: allocations in the heap is platform-dependent (and possibly different
                   4891: from one run to the next).
                   4892: 
1.27      crook    4893: 
1.32      anton    4894: @node Dictionary allocation, Heap Allocation, Memory model, Memory
                   4895: @subsection Dictionary allocation
1.27      crook    4896: @cindex reserving data space
                   4897: @cindex data space - reserving some
                   4898: 
1.32      anton    4899: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
                   4900: you want to deallocate X, you also deallocate everything
                   4901: allocated after X.
                   4902: 
1.68      anton    4903: @cindex contiguous regions in dictionary allocation
1.32      anton    4904: The allocations using the words below are contiguous and grow the region
                   4905: towards increasing addresses.  Other words that allocate dictionary
                   4906: memory of any kind (i.e., defining words including @code{:noname}) end
                   4907: the contiguous region and start a new one.
                   4908: 
                   4909: In ANS Forth only @code{create}d words are guaranteed to produce an
                   4910: address that is the start of the following contiguous region.  In
                   4911: particular, the cell allocated by @code{variable} is not guaranteed to
                   4912: be contiguous with following @code{allot}ed memory.
                   4913: 
                   4914: You can deallocate memory by using @code{allot} with a negative argument
                   4915: (with some restrictions, see @code{allot}). For larger deallocations use
                   4916: @code{marker}.
1.27      crook    4917: 
1.29      crook    4918: 
1.27      crook    4919: doc-here
                   4920: doc-unused
                   4921: doc-allot
                   4922: doc-c,
1.29      crook    4923: doc-f,
1.27      crook    4924: doc-,
                   4925: doc-2,
                   4926: 
1.32      anton    4927: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
                   4928: course you should allocate memory in an aligned way, too. I.e., before
                   4929: allocating allocating a cell, @code{here} must be cell-aligned, etc.
                   4930: The words below align @code{here} if it is not already.  Basically it is
                   4931: only already aligned for a type, if the last allocation was a multiple
                   4932: of the size of this type and if @code{here} was aligned for this type
                   4933: before.
                   4934: 
                   4935: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
                   4936: ANS Forth (@code{maxalign}ed in Gforth).
                   4937: 
                   4938: doc-align
                   4939: doc-falign
                   4940: doc-sfalign
                   4941: doc-dfalign
                   4942: doc-maxalign
                   4943: doc-cfalign
                   4944: 
                   4945: 
                   4946: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
                   4947: @subsection Heap allocation
                   4948: @cindex heap allocation
                   4949: @cindex dynamic allocation of memory
                   4950: @cindex memory-allocation word set
                   4951: 
1.68      anton    4952: @cindex contiguous regions and heap allocation
1.32      anton    4953: Heap allocation supports deallocation of allocated memory in any
                   4954: order. Dictionary allocation is not affected by it (i.e., it does not
                   4955: end a contiguous region). In Gforth, these words are implemented using
                   4956: the standard C library calls malloc(), free() and resize().
                   4957: 
1.68      anton    4958: The memory region produced by one invocation of @code{allocate} or
                   4959: @code{resize} is internally contiguous.  There is no contiguity between
                   4960: such a region and any other region (including others allocated from the
                   4961: heap).
                   4962: 
1.32      anton    4963: doc-allocate
                   4964: doc-free
                   4965: doc-resize
                   4966: 
1.27      crook    4967: 
1.32      anton    4968: @node Memory Access, Address arithmetic, Heap Allocation, Memory
1.1       anton    4969: @subsection Memory Access
                   4970: @cindex memory access words
                   4971: 
                   4972: doc-@
                   4973: doc-!
                   4974: doc-+!
                   4975: doc-c@
                   4976: doc-c!
                   4977: doc-2@
                   4978: doc-2!
                   4979: doc-f@
                   4980: doc-f!
                   4981: doc-sf@
                   4982: doc-sf!
                   4983: doc-df@
                   4984: doc-df!
1.144     anton    4985: doc-sw@
                   4986: doc-uw@
                   4987: doc-w!
                   4988: doc-sl@
                   4989: doc-ul@
                   4990: doc-l!
1.68      anton    4991: 
1.32      anton    4992: @node Address arithmetic, Memory Blocks, Memory Access, Memory
                   4993: @subsection Address arithmetic
1.1       anton    4994: @cindex address arithmetic words
                   4995: 
1.67      anton    4996: Address arithmetic is the foundation on which you can build data
                   4997: structures like arrays, records (@pxref{Structures}) and objects
                   4998: (@pxref{Object-oriented Forth}).
1.32      anton    4999: 
1.68      anton    5000: @cindex address unit
                   5001: @cindex au (address unit)
1.1       anton    5002: ANS Forth does not specify the sizes of the data types. Instead, it
                   5003: offers a number of words for computing sizes and doing address
1.29      crook    5004: arithmetic. Address arithmetic is performed in terms of address units
                   5005: (aus); on most systems the address unit is one byte. Note that a
                   5006: character may have more than one au, so @code{chars} is no noop (on
1.68      anton    5007: platforms where it is a noop, it compiles to nothing).
1.1       anton    5008: 
1.67      anton    5009: The basic address arithmetic words are @code{+} and @code{-}.  E.g., if
                   5010: you have the address of a cell, perform @code{1 cells +}, and you will
                   5011: have the address of the next cell.
                   5012: 
1.68      anton    5013: @cindex contiguous regions and address arithmetic
1.67      anton    5014: In ANS Forth you can perform address arithmetic only within a contiguous
                   5015: region, i.e., if you have an address into one region, you can only add
                   5016: and subtract such that the result is still within the region; you can
                   5017: only subtract or compare addresses from within the same contiguous
                   5018: region.  Reasons: several contiguous regions can be arranged in memory
                   5019: in any way; on segmented systems addresses may have unusual
                   5020: representations, such that address arithmetic only works within a
                   5021: region.  Gforth provides a few more guarantees (linear address space,
                   5022: dictionary grows upwards), but in general I have found it easy to stay
                   5023: within contiguous regions (exception: computing and comparing to the
                   5024: address just beyond the end of an array).
                   5025: 
1.1       anton    5026: @cindex alignment of addresses for types
                   5027: ANS Forth also defines words for aligning addresses for specific
                   5028: types. Many computers require that accesses to specific data types
                   5029: must only occur at specific addresses; e.g., that cells may only be
                   5030: accessed at addresses divisible by 4. Even if a machine allows unaligned
                   5031: accesses, it can usually perform aligned accesses faster. 
                   5032: 
                   5033: For the performance-conscious: alignment operations are usually only
                   5034: necessary during the definition of a data structure, not during the
                   5035: (more frequent) accesses to it.
                   5036: 
                   5037: ANS Forth defines no words for character-aligning addresses. This is not
                   5038: an oversight, but reflects the fact that addresses that are not
                   5039: char-aligned have no use in the standard and therefore will not be
                   5040: created.
                   5041: 
                   5042: @cindex @code{CREATE} and alignment
1.29      crook    5043: ANS Forth guarantees that addresses returned by @code{CREATE}d words
1.1       anton    5044: are cell-aligned; in addition, Gforth guarantees that these addresses
                   5045: are aligned for all purposes.
                   5046: 
1.26      crook    5047: Note that the ANS Forth word @code{char} has nothing to do with address
                   5048: arithmetic.
1.1       anton    5049: 
1.44      crook    5050: 
1.1       anton    5051: doc-chars
                   5052: doc-char+
                   5053: doc-cells
                   5054: doc-cell+
                   5055: doc-cell
                   5056: doc-aligned
                   5057: doc-floats
                   5058: doc-float+
                   5059: doc-float
                   5060: doc-faligned
                   5061: doc-sfloats
                   5062: doc-sfloat+
                   5063: doc-sfaligned
                   5064: doc-dfloats
                   5065: doc-dfloat+
                   5066: doc-dfaligned
                   5067: doc-maxaligned
                   5068: doc-cfaligned
                   5069: doc-address-unit-bits
1.145   ! anton    5070: doc-/w
        !          5071: doc-/l
1.44      crook    5072: 
1.32      anton    5073: @node Memory Blocks,  , Address arithmetic, Memory
1.1       anton    5074: @subsection Memory Blocks
                   5075: @cindex memory block words
1.27      crook    5076: @cindex character strings - moving and copying
                   5077: 
1.49      anton    5078: Memory blocks often represent character strings; For ways of storing
                   5079: character strings in memory see @ref{String Formats}.  For other
                   5080: string-processing words see @ref{Displaying characters and strings}.
1.1       anton    5081: 
1.67      anton    5082: A few of these words work on address unit blocks.  In that case, you
                   5083: usually have to insert @code{CHARS} before the word when working on
                   5084: character strings.  Most words work on character blocks, and expect a
                   5085: char-aligned address.
                   5086: 
                   5087: When copying characters between overlapping memory regions, use
                   5088: @code{chars move} or choose carefully between @code{cmove} and
                   5089: @code{cmove>}.
1.44      crook    5090: 
1.1       anton    5091: doc-move
                   5092: doc-erase
                   5093: doc-cmove
                   5094: doc-cmove>
                   5095: doc-fill
                   5096: doc-blank
1.21      crook    5097: doc-compare
1.111     anton    5098: doc-str=
                   5099: doc-str<
                   5100: doc-string-prefix?
1.21      crook    5101: doc-search
1.27      crook    5102: doc--trailing
                   5103: doc-/string
1.82      anton    5104: doc-bounds
1.141     anton    5105: doc-pad
1.111     anton    5106: 
1.27      crook    5107: @comment TODO examples
                   5108: 
1.1       anton    5109: 
1.26      crook    5110: @node Control Structures, Defining Words, Memory, Words
1.1       anton    5111: @section Control Structures
                   5112: @cindex control structures
                   5113: 
1.33      anton    5114: Control structures in Forth cannot be used interpretively, only in a
                   5115: colon definition@footnote{To be precise, they have no interpretation
                   5116: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
                   5117: not like this limitation, but have not seen a satisfying way around it
                   5118: yet, although many schemes have been proposed.
1.1       anton    5119: 
                   5120: @menu
1.33      anton    5121: * Selection::                   IF ... ELSE ... ENDIF
                   5122: * Simple Loops::                BEGIN ...
1.29      crook    5123: * Counted Loops::               DO
1.67      anton    5124: * Arbitrary control structures::  
                   5125: * Calls and returns::           
1.1       anton    5126: * Exception Handling::          
                   5127: @end menu
                   5128: 
                   5129: @node Selection, Simple Loops, Control Structures, Control Structures
                   5130: @subsection Selection
                   5131: @cindex selection control structures
                   5132: @cindex control structures for selection
                   5133: 
                   5134: @cindex @code{IF} control structure
                   5135: @example
1.29      crook    5136: @i{flag}
1.1       anton    5137: IF
1.29      crook    5138:   @i{code}
1.1       anton    5139: ENDIF
                   5140: @end example
1.21      crook    5141: @noindent
1.33      anton    5142: 
1.44      crook    5143: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
                   5144: with any bit set represents truth) @i{code} is executed.
1.33      anton    5145: 
1.1       anton    5146: @example
1.29      crook    5147: @i{flag}
1.1       anton    5148: IF
1.29      crook    5149:   @i{code1}
1.1       anton    5150: ELSE
1.29      crook    5151:   @i{code2}
1.1       anton    5152: ENDIF
                   5153: @end example
                   5154: 
1.44      crook    5155: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
                   5156: executed.
1.33      anton    5157: 
1.1       anton    5158: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
                   5159: standard, and @code{ENDIF} is not, although it is quite popular. We
                   5160: recommend using @code{ENDIF}, because it is less confusing for people
                   5161: who also know other languages (and is not prone to reinforcing negative
                   5162: prejudices against Forth in these people). Adding @code{ENDIF} to a
                   5163: system that only supplies @code{THEN} is simple:
                   5164: @example
1.82      anton    5165: : ENDIF   POSTPONE then ; immediate
1.1       anton    5166: @end example
                   5167: 
                   5168: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
                   5169: (adv.)}  has the following meanings:
                   5170: @quotation
                   5171: ... 2b: following next after in order ... 3d: as a necessary consequence
                   5172: (if you were there, then you saw them).
                   5173: @end quotation
                   5174: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
                   5175: and many other programming languages has the meaning 3d.]
                   5176: 
1.21      crook    5177: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
1.1       anton    5178: you can avoid using @code{?dup}. Using these alternatives is also more
1.26      crook    5179: efficient than using @code{?dup}. Definitions in ANS Forth
1.1       anton    5180: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
                   5181: @file{compat/control.fs}.
                   5182: 
                   5183: @cindex @code{CASE} control structure
                   5184: @example
1.29      crook    5185: @i{n}
1.1       anton    5186: CASE
1.29      crook    5187:   @i{n1} OF @i{code1} ENDOF
                   5188:   @i{n2} OF @i{code2} ENDOF
1.1       anton    5189:   @dots{}
1.68      anton    5190:   ( n ) @i{default-code} ( n )
1.131     anton    5191: ENDCASE ( )
1.1       anton    5192: @end example
                   5193: 
1.131     anton    5194: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}.  If
                   5195: no @i{ni} matches, the optional @i{default-code} is executed. The
                   5196: optional default case can be added by simply writing the code after
                   5197: the last @code{ENDOF}. It may use @i{n}, which is on top of the stack,
                   5198: but must not consume it.  The value @i{n} is consumed by this
                   5199: construction (either by a OF that matches, or by the ENDCASE, if no OF
                   5200: matches).
1.1       anton    5201: 
1.69      anton    5202: @progstyle
1.131     anton    5203: To keep the code understandable, you should ensure that you change the
                   5204: stack in the same way (wrt. number and types of stack items consumed
                   5205: and pushed) on all paths through a selection construct.
1.69      anton    5206: 
1.1       anton    5207: @node Simple Loops, Counted Loops, Selection, Control Structures
                   5208: @subsection Simple Loops
                   5209: @cindex simple loops
                   5210: @cindex loops without count 
                   5211: 
                   5212: @cindex @code{WHILE} loop
                   5213: @example
                   5214: BEGIN
1.29      crook    5215:   @i{code1}
                   5216:   @i{flag}
1.1       anton    5217: WHILE
1.29      crook    5218:   @i{code2}
1.1       anton    5219: REPEAT
                   5220: @end example
                   5221: 
1.29      crook    5222: @i{code1} is executed and @i{flag} is computed. If it is true,
                   5223: @i{code2} is executed and the loop is restarted; If @i{flag} is
1.1       anton    5224: false, execution continues after the @code{REPEAT}.
                   5225: 
                   5226: @cindex @code{UNTIL} loop
                   5227: @example
                   5228: BEGIN
1.29      crook    5229:   @i{code}
                   5230:   @i{flag}
1.1       anton    5231: UNTIL
                   5232: @end example
                   5233: 
1.29      crook    5234: @i{code} is executed. The loop is restarted if @code{flag} is false.
1.1       anton    5235: 
1.69      anton    5236: @progstyle
                   5237: To keep the code understandable, a complete iteration of the loop should
                   5238: not change the number and types of the items on the stacks.
                   5239: 
1.1       anton    5240: @cindex endless loop
                   5241: @cindex loops, endless
                   5242: @example
                   5243: BEGIN
1.29      crook    5244:   @i{code}
1.1       anton    5245: AGAIN
                   5246: @end example
                   5247: 
                   5248: This is an endless loop.
                   5249: 
                   5250: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
                   5251: @subsection Counted Loops
                   5252: @cindex counted loops
                   5253: @cindex loops, counted
                   5254: @cindex @code{DO} loops
                   5255: 
                   5256: The basic counted loop is:
                   5257: @example
1.29      crook    5258: @i{limit} @i{start}
1.1       anton    5259: ?DO
1.29      crook    5260:   @i{body}
1.1       anton    5261: LOOP
                   5262: @end example
                   5263: 
1.29      crook    5264: This performs one iteration for every integer, starting from @i{start}
                   5265: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
1.21      crook    5266: accessed with @code{i}. For example, the loop:
1.1       anton    5267: @example
                   5268: 10 0 ?DO
                   5269:   i .
                   5270: LOOP
                   5271: @end example
1.21      crook    5272: @noindent
                   5273: prints @code{0 1 2 3 4 5 6 7 8 9}
                   5274: 
1.1       anton    5275: The index of the innermost loop can be accessed with @code{i}, the index
                   5276: of the next loop with @code{j}, and the index of the third loop with
                   5277: @code{k}.
                   5278: 
1.44      crook    5279: 
1.1       anton    5280: doc-i
                   5281: doc-j
                   5282: doc-k
                   5283: 
1.44      crook    5284: 
1.1       anton    5285: The loop control data are kept on the return stack, so there are some
1.21      crook    5286: restrictions on mixing return stack accesses and counted loop words. In
                   5287: particuler, if you put values on the return stack outside the loop, you
                   5288: cannot read them inside the loop@footnote{well, not in a way that is
                   5289: portable.}. If you put values on the return stack within a loop, you
                   5290: have to remove them before the end of the loop and before accessing the
                   5291: index of the loop.
1.1       anton    5292: 
                   5293: There are several variations on the counted loop:
                   5294: 
1.21      crook    5295: @itemize @bullet
                   5296: @item
                   5297: @code{LEAVE} leaves the innermost counted loop immediately; execution
                   5298: continues after the associated @code{LOOP} or @code{NEXT}. For example:
                   5299: 
                   5300: @example
                   5301: 10 0 ?DO  i DUP . 3 = IF LEAVE THEN LOOP
                   5302: @end example
                   5303: prints @code{0 1 2 3}
                   5304: 
1.1       anton    5305: 
1.21      crook    5306: @item
                   5307: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
                   5308: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
                   5309: return stack so @code{EXIT} can get to its return address. For example:
                   5310: 
                   5311: @example
                   5312: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
                   5313: @end example
                   5314: prints @code{0 1 2 3}
                   5315: 
                   5316: 
                   5317: @item
1.29      crook    5318: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
1.1       anton    5319: (and @code{LOOP} iterates until they become equal by wrap-around
                   5320: arithmetic). This behaviour is usually not what you want. Therefore,
                   5321: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1.29      crook    5322: @code{?DO}), which do not enter the loop if @i{start} is greater than
                   5323: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1.1       anton    5324: unsigned loop parameters.
                   5325: 
1.21      crook    5326: @item
                   5327: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
                   5328: the loop, independent of the loop parameters. Do not use @code{DO}, even
                   5329: if you know that the loop is entered in any case. Such knowledge tends
                   5330: to become invalid during maintenance of a program, and then the
                   5331: @code{DO} will make trouble.
                   5332: 
                   5333: @item
1.29      crook    5334: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
                   5335: index by @i{n} instead of by 1. The loop is terminated when the border
                   5336: between @i{limit-1} and @i{limit} is crossed. E.g.:
1.1       anton    5337: 
1.21      crook    5338: @example
                   5339: 4 0 +DO  i .  2 +LOOP
                   5340: @end example
                   5341: @noindent
                   5342: prints @code{0 2}
                   5343: 
                   5344: @example
                   5345: 4 1 +DO  i .  2 +LOOP
                   5346: @end example
                   5347: @noindent
                   5348: prints @code{1 3}
1.1       anton    5349: 
1.68      anton    5350: @item
1.1       anton    5351: @cindex negative increment for counted loops
                   5352: @cindex counted loops with negative increment
1.29      crook    5353: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
1.1       anton    5354: 
1.21      crook    5355: @example
                   5356: -1 0 ?DO  i .  -1 +LOOP
                   5357: @end example
                   5358: @noindent
                   5359: prints @code{0 -1}
1.1       anton    5360: 
1.21      crook    5361: @example
                   5362: 0 0 ?DO  i .  -1 +LOOP
                   5363: @end example
                   5364: prints nothing.
1.1       anton    5365: 
1.29      crook    5366: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
                   5367: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
                   5368: index by @i{u} each iteration. The loop is terminated when the border
                   5369: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
1.1       anton    5370: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
                   5371: 
1.21      crook    5372: @example
                   5373: -2 0 -DO  i .  1 -LOOP
                   5374: @end example
                   5375: @noindent
                   5376: prints @code{0 -1}
1.1       anton    5377: 
1.21      crook    5378: @example
                   5379: -1 0 -DO  i .  1 -LOOP
                   5380: @end example
                   5381: @noindent
                   5382: prints @code{0}
                   5383: 
                   5384: @example
                   5385: 0 0 -DO  i .  1 -LOOP
                   5386: @end example
                   5387: @noindent
                   5388: prints nothing.
1.1       anton    5389: 
1.21      crook    5390: @end itemize
1.1       anton    5391: 
                   5392: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
1.26      crook    5393: @code{-LOOP} are not defined in ANS Forth. However, an implementation
                   5394: for these words that uses only standard words is provided in
                   5395: @file{compat/loops.fs}.
1.1       anton    5396: 
                   5397: 
                   5398: @cindex @code{FOR} loops
1.26      crook    5399: Another counted loop is:
1.1       anton    5400: @example
1.29      crook    5401: @i{n}
1.1       anton    5402: FOR
1.29      crook    5403:   @i{body}
1.1       anton    5404: NEXT
                   5405: @end example
                   5406: This is the preferred loop of native code compiler writers who are too
1.26      crook    5407: lazy to optimize @code{?DO} loops properly. This loop structure is not
1.29      crook    5408: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
                   5409: @code{i} produces values starting with @i{n} and ending with 0. Other
1.26      crook    5410: Forth systems may behave differently, even if they support @code{FOR}
                   5411: loops. To avoid problems, don't use @code{FOR} loops.
1.1       anton    5412: 
                   5413: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
                   5414: @subsection Arbitrary control structures
                   5415: @cindex control structures, user-defined
                   5416: 
                   5417: @cindex control-flow stack
                   5418: ANS Forth permits and supports using control structures in a non-nested
                   5419: way. Information about incomplete control structures is stored on the
                   5420: control-flow stack. This stack may be implemented on the Forth data
                   5421: stack, and this is what we have done in Gforth.
                   5422: 
                   5423: @cindex @code{orig}, control-flow stack item
                   5424: @cindex @code{dest}, control-flow stack item
                   5425: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
                   5426: entry represents a backward branch target. A few words are the basis for
                   5427: building any control structure possible (except control structures that
                   5428: need storage, like calls, coroutines, and backtracking).
                   5429: 
1.44      crook    5430: 
1.1       anton    5431: doc-if
                   5432: doc-ahead
                   5433: doc-then
                   5434: doc-begin
                   5435: doc-until
                   5436: doc-again
                   5437: doc-cs-pick
                   5438: doc-cs-roll
                   5439: 
1.44      crook    5440: 
1.21      crook    5441: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
                   5442: manipulate the control-flow stack in a portable way. Without them, you
                   5443: would need to know how many stack items are occupied by a control-flow
                   5444: entry (many systems use one cell. In Gforth they currently take three,
                   5445: but this may change in the future).
                   5446: 
1.1       anton    5447: Some standard control structure words are built from these words:
                   5448: 
1.44      crook    5449: 
1.1       anton    5450: doc-else
                   5451: doc-while
                   5452: doc-repeat
                   5453: 
1.44      crook    5454: 
                   5455: @noindent
1.1       anton    5456: Gforth adds some more control-structure words:
                   5457: 
1.44      crook    5458: 
1.1       anton    5459: doc-endif
                   5460: doc-?dup-if
                   5461: doc-?dup-0=-if
                   5462: 
1.44      crook    5463: 
                   5464: @noindent
1.1       anton    5465: Counted loop words constitute a separate group of words:
                   5466: 
1.44      crook    5467: 
1.1       anton    5468: doc-?do
                   5469: doc-+do
                   5470: doc-u+do
                   5471: doc--do
                   5472: doc-u-do
                   5473: doc-do
                   5474: doc-for
                   5475: doc-loop
                   5476: doc-+loop
                   5477: doc--loop
                   5478: doc-next
                   5479: doc-leave
                   5480: doc-?leave
                   5481: doc-unloop
                   5482: doc-done
                   5483: 
1.44      crook    5484: 
1.21      crook    5485: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
                   5486: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
1.1       anton    5487: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
                   5488: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
                   5489: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
                   5490: resolved (by using one of the loop-ending words or @code{DONE}).
                   5491: 
1.44      crook    5492: @noindent
1.26      crook    5493: Another group of control structure words are:
1.1       anton    5494: 
1.44      crook    5495: 
1.1       anton    5496: doc-case
                   5497: doc-endcase
                   5498: doc-of
                   5499: doc-endof
                   5500: 
1.44      crook    5501: 
1.21      crook    5502: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
                   5503: @code{CS-ROLL}.
1.1       anton    5504: 
                   5505: @subsubsection Programming Style
1.47      crook    5506: @cindex control structures programming style
                   5507: @cindex programming style, arbitrary control structures
1.1       anton    5508: 
                   5509: In order to ensure readability we recommend that you do not create
                   5510: arbitrary control structures directly, but define new control structure
                   5511: words for the control structure you want and use these words in your
1.26      crook    5512: program. For example, instead of writing:
1.1       anton    5513: 
                   5514: @example
1.26      crook    5515: BEGIN
1.1       anton    5516:   ...
1.26      crook    5517: IF [ 1 CS-ROLL ]
1.1       anton    5518:   ...
1.26      crook    5519: AGAIN THEN
1.1       anton    5520: @end example
                   5521: 
1.21      crook    5522: @noindent
1.1       anton    5523: we recommend defining control structure words, e.g.,
                   5524: 
                   5525: @example
1.26      crook    5526: : WHILE ( DEST -- ORIG DEST )
                   5527:  POSTPONE IF
                   5528:  1 CS-ROLL ; immediate
                   5529: 
                   5530: : REPEAT ( orig dest -- )
                   5531:  POSTPONE AGAIN
                   5532:  POSTPONE THEN ; immediate
1.1       anton    5533: @end example
                   5534: 
1.21      crook    5535: @noindent
1.1       anton    5536: and then using these to create the control structure:
                   5537: 
                   5538: @example
1.26      crook    5539: BEGIN
1.1       anton    5540:   ...
1.26      crook    5541: WHILE
1.1       anton    5542:   ...
1.26      crook    5543: REPEAT
1.1       anton    5544: @end example
                   5545: 
                   5546: That's much easier to read, isn't it? Of course, @code{REPEAT} and
                   5547: @code{WHILE} are predefined, so in this example it would not be
                   5548: necessary to define them.
                   5549: 
                   5550: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
                   5551: @subsection Calls and returns
                   5552: @cindex calling a definition
                   5553: @cindex returning from a definition
                   5554: 
1.3       anton    5555: @cindex recursive definitions
                   5556: A definition can be called simply be writing the name of the definition
1.26      crook    5557: to be called. Normally a definition is invisible during its own
1.3       anton    5558: definition. If you want to write a directly recursive definition, you
1.26      crook    5559: can use @code{recursive} to make the current definition visible, or
                   5560: @code{recurse} to call the current definition directly.
1.3       anton    5561: 
1.44      crook    5562: 
1.3       anton    5563: doc-recursive
                   5564: doc-recurse
                   5565: 
1.44      crook    5566: 
1.21      crook    5567: @comment TODO add example of the two recursion methods
1.12      anton    5568: @quotation
                   5569: @progstyle
                   5570: I prefer using @code{recursive} to @code{recurse}, because calling the
                   5571: definition by name is more descriptive (if the name is well-chosen) than
                   5572: the somewhat cryptic @code{recurse}.  E.g., in a quicksort
                   5573: implementation, it is much better to read (and think) ``now sort the
                   5574: partitions'' than to read ``now do a recursive call''.
                   5575: @end quotation
1.3       anton    5576: 
1.29      crook    5577: For mutual recursion, use @code{Defer}red words, like this:
1.3       anton    5578: 
                   5579: @example
1.28      crook    5580: Defer foo
1.3       anton    5581: 
                   5582: : bar ( ... -- ... )
                   5583:  ... foo ... ;
                   5584: 
                   5585: :noname ( ... -- ... )
                   5586:  ... bar ... ;
                   5587: IS foo
                   5588: @end example
                   5589: 
1.44      crook    5590: Deferred words are discussed in more detail in @ref{Deferred words}.
1.33      anton    5591: 
1.26      crook    5592: The current definition returns control to the calling definition when
1.33      anton    5593: the end of the definition is reached or @code{EXIT} is encountered.
1.1       anton    5594: 
                   5595: doc-exit
                   5596: doc-;s
                   5597: 
1.44      crook    5598: 
1.1       anton    5599: @node Exception Handling,  , Calls and returns, Control Structures
                   5600: @subsection Exception Handling
1.26      crook    5601: @cindex exceptions
1.1       anton    5602: 
1.68      anton    5603: @c quit is a very bad idea for error handling, 
                   5604: @c because it does not translate into a THROW
                   5605: @c it also does not belong into this chapter
                   5606: 
                   5607: If a word detects an error condition that it cannot handle, it can
                   5608: @code{throw} an exception.  In the simplest case, this will terminate
                   5609: your program, and report an appropriate error.
1.21      crook    5610: 
1.68      anton    5611: doc-throw
1.1       anton    5612: 
1.69      anton    5613: @code{Throw} consumes a cell-sized error number on the stack. There are
                   5614: some predefined error numbers in ANS Forth (see @file{errors.fs}).  In
                   5615: Gforth (and most other systems) you can use the iors produced by various
                   5616: words as error numbers (e.g., a typical use of @code{allocate} is
                   5617: @code{allocate throw}).  Gforth also provides the word @code{exception}
                   5618: to define your own error numbers (with decent error reporting); an ANS
                   5619: Forth version of this word (but without the error messages) is available
                   5620: in @code{compat/except.fs}.  And finally, you can use your own error
1.68      anton    5621: numbers (anything outside the range -4095..0), but won't get nice error
                   5622: messages, only numbers.  For example, try:
                   5623: 
                   5624: @example
1.69      anton    5625: -10 throw                    \ ANS defined
                   5626: -267 throw                   \ system defined
                   5627: s" my error" exception throw \ user defined
                   5628: 7 throw                      \ arbitrary number
1.68      anton    5629: @end example
                   5630: 
                   5631: doc---exception-exception
1.1       anton    5632: 
1.69      anton    5633: A common idiom to @code{THROW} a specific error if a flag is true is
                   5634: this:
                   5635: 
                   5636: @example
                   5637: @code{( flag ) 0<> @i{errno} and throw}
                   5638: @end example
                   5639: 
                   5640: Your program can provide exception handlers to catch exceptions.  An
                   5641: exception handler can be used to correct the problem, or to clean up
                   5642: some data structures and just throw the exception to the next exception
                   5643: handler.  Note that @code{throw} jumps to the dynamically innermost
                   5644: exception handler.  The system's exception handler is outermost, and just
                   5645: prints an error and restarts command-line interpretation (or, in batch
                   5646: mode (i.e., while processing the shell command line), leaves Gforth).
1.1       anton    5647: 
1.68      anton    5648: The ANS Forth way to catch exceptions is @code{catch}:
1.1       anton    5649: 
1.68      anton    5650: doc-catch
                   5651: 
                   5652: The most common use of exception handlers is to clean up the state when
                   5653: an error happens.  E.g.,
1.1       anton    5654: 
1.26      crook    5655: @example
1.68      anton    5656: base @ >r hex \ actually the hex should be inside foo, or we h
                   5657: ['] foo catch ( nerror|0 )
                   5658: r> base !
1.69      anton    5659: ( nerror|0 ) throw \ pass it on
1.26      crook    5660: @end example
1.1       anton    5661: 
1.69      anton    5662: A use of @code{catch} for handling the error @code{myerror} might look
                   5663: like this:
1.44      crook    5664: 
1.68      anton    5665: @example
1.69      anton    5666: ['] foo catch
                   5667: CASE
                   5668:   myerror OF ... ( do something about it ) ENDOF
                   5669:   dup throw \ default: pass other errors on, do nothing on non-errors
                   5670: ENDCASE
1.68      anton    5671: @end example
1.44      crook    5672: 
1.68      anton    5673: Having to wrap the code into a separate word is often cumbersome,
                   5674: therefore Gforth provides an alternative syntax:
1.1       anton    5675: 
                   5676: @example
1.69      anton    5677: TRY
1.68      anton    5678:   @i{code1}
1.69      anton    5679: RECOVER     \ optional
1.68      anton    5680:   @i{code2} \ optional
1.69      anton    5681: ENDTRY
1.1       anton    5682: @end example
                   5683: 
1.68      anton    5684: This performs @i{Code1}.  If @i{code1} completes normally, execution
                   5685: continues after the @code{endtry}.  If @i{Code1} throws, the stacks are
                   5686: reset to the state during @code{try}, the throw value is pushed on the
                   5687: data stack, and execution constinues at @i{code2}, and finally falls
1.92      anton    5688: through the @code{endtry} into the following code.
1.26      crook    5689: 
1.68      anton    5690: doc-try
                   5691: doc-recover
                   5692: doc-endtry
1.26      crook    5693: 
1.69      anton    5694: The cleanup example from above in this syntax:
1.26      crook    5695: 
1.68      anton    5696: @example
1.69      anton    5697: base @ >r TRY
1.68      anton    5698:   hex foo \ now the hex is placed correctly
1.69      anton    5699:   0       \ value for throw
1.92      anton    5700: RECOVER ENDTRY
1.68      anton    5701: r> base ! throw
1.1       anton    5702: @end example
                   5703: 
1.69      anton    5704: And here's the error handling example:
1.1       anton    5705: 
1.68      anton    5706: @example
1.69      anton    5707: TRY
1.68      anton    5708:   foo
1.69      anton    5709: RECOVER
                   5710:   CASE
                   5711:     myerror OF ... ( do something about it ) ENDOF
                   5712:     throw \ pass other errors on
                   5713:   ENDCASE
                   5714: ENDTRY
1.68      anton    5715: @end example
1.1       anton    5716: 
1.69      anton    5717: @progstyle
                   5718: As usual, you should ensure that the stack depth is statically known at
                   5719: the end: either after the @code{throw} for passing on errors, or after
                   5720: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
                   5721: selection construct for handling the error).
                   5722: 
1.68      anton    5723: There are two alternatives to @code{throw}: @code{Abort"} is conditional
                   5724: and you can provide an error message.  @code{Abort} just produces an
                   5725: ``Aborted'' error.
1.1       anton    5726: 
1.68      anton    5727: The problem with these words is that exception handlers cannot
                   5728: differentiate between different @code{abort"}s; they just look like
                   5729: @code{-2 throw} to them (the error message cannot be accessed by
                   5730: standard programs).  Similar @code{abort} looks like @code{-1 throw} to
                   5731: exception handlers.
1.44      crook    5732: 
1.68      anton    5733: doc-abort"
1.26      crook    5734: doc-abort
1.29      crook    5735: 
                   5736: 
1.44      crook    5737: 
1.29      crook    5738: @c -------------------------------------------------------------
1.47      crook    5739: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
1.29      crook    5740: @section Defining Words
                   5741: @cindex defining words
                   5742: 
1.47      crook    5743: Defining words are used to extend Forth by creating new entries in the dictionary.
                   5744: 
1.29      crook    5745: @menu
1.67      anton    5746: * CREATE::                      
1.44      crook    5747: * Variables::                   Variables and user variables
1.67      anton    5748: * Constants::                   
1.44      crook    5749: * Values::                      Initialised variables
1.67      anton    5750: * Colon Definitions::           
1.44      crook    5751: * Anonymous Definitions::       Definitions without names
1.69      anton    5752: * Supplying names::             Passing definition names as strings
1.67      anton    5753: * User-defined Defining Words::  
1.44      crook    5754: * Deferred words::              Allow forward references
1.67      anton    5755: * Aliases::                     
1.29      crook    5756: @end menu
                   5757: 
1.44      crook    5758: @node CREATE, Variables, Defining Words, Defining Words
                   5759: @subsection @code{CREATE}
1.29      crook    5760: @cindex simple defining words
                   5761: @cindex defining words, simple
                   5762: 
                   5763: Defining words are used to create new entries in the dictionary. The
                   5764: simplest defining word is @code{CREATE}. @code{CREATE} is used like
                   5765: this:
                   5766: 
                   5767: @example
                   5768: CREATE new-word1
                   5769: @end example
                   5770: 
1.69      anton    5771: @code{CREATE} is a parsing word, i.e., it takes an argument from the
                   5772: input stream (@code{new-word1} in our example).  It generates a
                   5773: dictionary entry for @code{new-word1}. When @code{new-word1} is
                   5774: executed, all that it does is leave an address on the stack. The address
                   5775: represents the value of the data space pointer (@code{HERE}) at the time
                   5776: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
                   5777: associating a name with the address of a region of memory.
1.29      crook    5778: 
1.34      anton    5779: doc-create
                   5780: 
1.69      anton    5781: Note that in ANS Forth guarantees only for @code{create} that its body
                   5782: is in dictionary data space (i.e., where @code{here}, @code{allot}
                   5783: etc. work, @pxref{Dictionary allocation}).  Also, in ANS Forth only
                   5784: @code{create}d words can be modified with @code{does>}
                   5785: (@pxref{User-defined Defining Words}).  And in ANS Forth @code{>body}
                   5786: can only be applied to @code{create}d words.
                   5787: 
1.29      crook    5788: By extending this example to reserve some memory in data space, we end
1.69      anton    5789: up with something like a @i{variable}. Here are two different ways to do
                   5790: it:
1.29      crook    5791: 
                   5792: @example
                   5793: CREATE new-word2 1 cells allot  \ reserve 1 cell - initial value undefined
                   5794: CREATE new-word3 4 ,            \ reserve 1 cell and initialise it (to 4)
                   5795: @end example
                   5796: 
                   5797: The variable can be examined and modified using @code{@@} (``fetch'') and
                   5798: @code{!} (``store'') like this:
                   5799: 
                   5800: @example
                   5801: new-word2 @@ .      \ get address, fetch from it and display
                   5802: 1234 new-word2 !   \ new value, get address, store to it
                   5803: @end example
                   5804: 
1.44      crook    5805: @cindex arrays
                   5806: A similar mechanism can be used to create arrays. For example, an
                   5807: 80-character text input buffer:
1.29      crook    5808: 
                   5809: @example
1.44      crook    5810: CREATE text-buf 80 chars allot
                   5811: 
                   5812: text-buf 0 chars c@@ \ the 1st character (offset 0)
                   5813: text-buf 3 chars c@@ \ the 4th character (offset 3)
                   5814: @end example
1.29      crook    5815: 
1.44      crook    5816: You can build arbitrarily complex data structures by allocating
1.49      anton    5817: appropriate areas of memory. For further discussions of this, and to
1.66      anton    5818: learn about some Gforth tools that make it easier,
1.49      anton    5819: @xref{Structures}.
1.44      crook    5820: 
                   5821: 
                   5822: @node Variables, Constants, CREATE, Defining Words
                   5823: @subsection Variables
                   5824: @cindex variables
                   5825: 
                   5826: The previous section showed how a sequence of commands could be used to
                   5827: generate a variable.  As a final refinement, the whole code sequence can
                   5828: be wrapped up in a defining word (pre-empting the subject of the next
                   5829: section), making it easier to create new variables:
                   5830: 
                   5831: @example
                   5832: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
                   5833: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
                   5834: 
                   5835: myvariableX foo \ variable foo starts off with an unknown value
                   5836: myvariable0 joe \ whilst joe is initialised to 0
1.29      crook    5837: 
                   5838: 45 3 * foo !   \ set foo to 135
                   5839: 1234 joe !     \ set joe to 1234
                   5840: 3 joe +!       \ increment joe by 3.. to 1237
                   5841: @end example
                   5842: 
                   5843: Not surprisingly, there is no need to define @code{myvariable}, since
1.44      crook    5844: Forth already has a definition @code{Variable}. ANS Forth does not
1.69      anton    5845: guarantee that a @code{Variable} is initialised when it is created
                   5846: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
                   5847: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
                   5848: like @code{myvariable0}). Forth also provides @code{2Variable} and
1.47      crook    5849: @code{fvariable} for double and floating-point variables, respectively
1.69      anton    5850: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
1.47      crook    5851: store a boolean, you can use @code{on} and @code{off} to toggle its
                   5852: state.
1.29      crook    5853: 
1.34      anton    5854: doc-variable
                   5855: doc-2variable
                   5856: doc-fvariable
                   5857: 
1.29      crook    5858: @cindex user variables
                   5859: @cindex user space
                   5860: The defining word @code{User} behaves in the same way as @code{Variable}.
                   5861: The difference is that it reserves space in @i{user (data) space} rather
                   5862: than normal data space. In a Forth system that has a multi-tasker, each
                   5863: task has its own set of user variables.
                   5864: 
1.34      anton    5865: doc-user
1.67      anton    5866: @c doc-udp
                   5867: @c doc-uallot
1.34      anton    5868: 
1.29      crook    5869: @comment TODO is that stuff about user variables strictly correct? Is it
                   5870: @comment just terminal tasks that have user variables?
                   5871: @comment should document tasker.fs (with some examples) elsewhere
                   5872: @comment in this manual, then expand on user space and user variables.
                   5873: 
1.44      crook    5874: @node Constants, Values, Variables, Defining Words
                   5875: @subsection Constants
                   5876: @cindex constants
                   5877: 
                   5878: @code{Constant} allows you to declare a fixed value and refer to it by
                   5879: name. For example:
1.29      crook    5880: 
                   5881: @example
                   5882: 12 Constant INCHES-PER-FOOT
                   5883: 3E+08 fconstant SPEED-O-LIGHT
                   5884: @end example
                   5885: 
                   5886: A @code{Variable} can be both read and written, so its run-time
                   5887: behaviour is to supply an address through which its current value can be
                   5888: manipulated. In contrast, the value of a @code{Constant} cannot be
                   5889: changed once it has been declared@footnote{Well, often it can be -- but
                   5890: not in a Standard, portable way. It's safer to use a @code{Value} (read
                   5891: on).} so it's not necessary to supply the address -- it is more
                   5892: efficient to return the value of the constant directly. That's exactly
                   5893: what happens; the run-time effect of a constant is to put its value on
1.49      anton    5894: the top of the stack (You can find one
                   5895: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
1.29      crook    5896: 
1.69      anton    5897: Forth also provides @code{2Constant} and @code{fconstant} for defining
1.29      crook    5898: double and floating-point constants, respectively.
                   5899: 
1.34      anton    5900: doc-constant
                   5901: doc-2constant
                   5902: doc-fconstant
                   5903: 
                   5904: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
1.44      crook    5905: @c nac-> How could that not be true in an ANS Forth? You can't define a
                   5906: @c constant, use it and then delete the definition of the constant..
1.69      anton    5907: 
                   5908: @c anton->An ANS Forth system can compile a constant to a literal; On
                   5909: @c decompilation you would see only the number, just as if it had been used
                   5910: @c in the first place.  The word will stay, of course, but it will only be
                   5911: @c used by the text interpreter (no run-time duties, except when it is 
                   5912: @c POSTPONEd or somesuch).
                   5913: 
                   5914: @c nac:
1.44      crook    5915: @c I agree that it's rather deep, but IMO it is an important difference
                   5916: @c relative to other programming languages.. often it's annoying: it
                   5917: @c certainly changes my programming style relative to C.
                   5918: 
1.69      anton    5919: @c anton: In what way?
                   5920: 
1.29      crook    5921: Constants in Forth behave differently from their equivalents in other
                   5922: programming languages. In other languages, a constant (such as an EQU in
                   5923: assembler or a #define in C) only exists at compile-time; in the
                   5924: executable program the constant has been translated into an absolute
                   5925: number and, unless you are using a symbolic debugger, it's impossible to
                   5926: know what abstract thing that number represents. In Forth a constant has
1.44      crook    5927: an entry in the header space and remains there after the code that uses
                   5928: it has been defined. In fact, it must remain in the dictionary since it
                   5929: has run-time duties to perform. For example:
1.29      crook    5930: 
                   5931: @example
                   5932: 12 Constant INCHES-PER-FOOT
                   5933: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
                   5934: @end example
                   5935: 
                   5936: @cindex in-lining of constants
                   5937: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
                   5938: associated with the constant @code{INCHES-PER-FOOT}. If you use
                   5939: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
                   5940: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
                   5941: attempt to optimise constants by in-lining them where they are used. You
                   5942: can force Gforth to in-line a constant like this:
                   5943: 
                   5944: @example
                   5945: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
                   5946: @end example
                   5947: 
                   5948: If you use @code{see} to decompile @i{this} version of
                   5949: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
1.49      anton    5950: longer present. To understand how this works, read
                   5951: @ref{Interpret/Compile states}, and @ref{Literals}.
1.29      crook    5952: 
                   5953: In-lining constants in this way might improve execution time
                   5954: fractionally, and can ensure that a constant is now only referenced at
                   5955: compile-time. However, the definition of the constant still remains in
                   5956: the dictionary. Some Forth compilers provide a mechanism for controlling
                   5957: a second dictionary for holding transient words such that this second
                   5958: dictionary can be deleted later in order to recover memory
                   5959: space. However, there is no standard way of doing this.
                   5960: 
                   5961: 
1.44      crook    5962: @node Values, Colon Definitions, Constants, Defining Words
                   5963: @subsection Values
                   5964: @cindex values
1.34      anton    5965: 
1.69      anton    5966: A @code{Value} behaves like a @code{Constant}, but it can be changed.
                   5967: @code{TO} is a parsing word that changes a @code{Values}.  In Gforth
                   5968: (not in ANS Forth) you can access (and change) a @code{value} also with
                   5969: @code{>body}.
                   5970: 
                   5971: Here are some
                   5972: examples:
1.29      crook    5973: 
                   5974: @example
1.69      anton    5975: 12 Value APPLES     \ Define APPLES with an initial value of 12
                   5976: 34 TO APPLES        \ Change the value of APPLES. TO is a parsing word
                   5977: 1 ' APPLES >body +! \ Increment APPLES.  Non-standard usage.
                   5978: APPLES              \ puts 35 on the top of the stack.
1.29      crook    5979: @end example
                   5980: 
1.44      crook    5981: doc-value
                   5982: doc-to
1.29      crook    5983: 
1.35      anton    5984: 
1.69      anton    5985: 
1.44      crook    5986: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
                   5987: @subsection Colon Definitions
                   5988: @cindex colon definitions
1.35      anton    5989: 
                   5990: @example
1.44      crook    5991: : name ( ... -- ... )
                   5992:     word1 word2 word3 ;
1.29      crook    5993: @end example
                   5994: 
1.44      crook    5995: @noindent
                   5996: Creates a word called @code{name} that, upon execution, executes
                   5997: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
1.29      crook    5998: 
1.49      anton    5999: The explanation above is somewhat superficial. For simple examples of
                   6000: colon definitions see @ref{Your first definition}.  For an in-depth
1.66      anton    6001: discussion of some of the issues involved, @xref{Interpretation and
1.49      anton    6002: Compilation Semantics}.
1.29      crook    6003: 
1.44      crook    6004: doc-:
                   6005: doc-;
1.1       anton    6006: 
1.34      anton    6007: 
1.69      anton    6008: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
1.44      crook    6009: @subsection Anonymous Definitions
                   6010: @cindex colon definitions
                   6011: @cindex defining words without name
1.34      anton    6012: 
1.44      crook    6013: Sometimes you want to define an @dfn{anonymous word}; a word without a
                   6014: name. You can do this with:
1.1       anton    6015: 
1.44      crook    6016: doc-:noname
1.1       anton    6017: 
1.44      crook    6018: This leaves the execution token for the word on the stack after the
                   6019: closing @code{;}. Here's an example in which a deferred word is
                   6020: initialised with an @code{xt} from an anonymous colon definition:
1.1       anton    6021: 
1.29      crook    6022: @example
1.44      crook    6023: Defer deferred
                   6024: :noname ( ... -- ... )
                   6025:   ... ;
                   6026: IS deferred
1.29      crook    6027: @end example
1.26      crook    6028: 
1.44      crook    6029: @noindent
                   6030: Gforth provides an alternative way of doing this, using two separate
                   6031: words:
1.27      crook    6032: 
1.44      crook    6033: doc-noname
                   6034: @cindex execution token of last defined word
1.116     anton    6035: doc-latestxt
1.1       anton    6036: 
1.44      crook    6037: @noindent
                   6038: The previous example can be rewritten using @code{noname} and
1.116     anton    6039: @code{latestxt}:
1.1       anton    6040: 
1.26      crook    6041: @example
1.44      crook    6042: Defer deferred
                   6043: noname : ( ... -- ... )
                   6044:   ... ;
1.116     anton    6045: latestxt IS deferred
1.26      crook    6046: @end example
1.1       anton    6047: 
1.29      crook    6048: @noindent
1.44      crook    6049: @code{noname} works with any defining word, not just @code{:}.
                   6050: 
1.116     anton    6051: @code{latestxt} also works when the last word was not defined as
1.71      anton    6052: @code{noname}.  It does not work for combined words, though.  It also has
                   6053: the useful property that is is valid as soon as the header for a
                   6054: definition has been built. Thus:
1.44      crook    6055: 
                   6056: @example
1.116     anton    6057: latestxt . : foo [ latestxt . ] ; ' foo .
1.44      crook    6058: @end example
1.1       anton    6059: 
1.44      crook    6060: @noindent
                   6061: prints 3 numbers; the last two are the same.
1.26      crook    6062: 
1.69      anton    6063: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
                   6064: @subsection Supplying the name of a defined word
                   6065: @cindex names for defined words
                   6066: @cindex defining words, name given in a string
                   6067: 
                   6068: By default, a defining word takes the name for the defined word from the
                   6069: input stream. Sometimes you want to supply the name from a string. You
                   6070: can do this with:
                   6071: 
                   6072: doc-nextname
                   6073: 
                   6074: For example:
                   6075: 
                   6076: @example
                   6077: s" foo" nextname create
                   6078: @end example
                   6079: 
                   6080: @noindent
                   6081: is equivalent to:
                   6082: 
                   6083: @example
                   6084: create foo
                   6085: @end example
                   6086: 
                   6087: @noindent
                   6088: @code{nextname} works with any defining word.
                   6089: 
1.1       anton    6090: 
1.69      anton    6091: @node User-defined Defining Words, Deferred words, Supplying names, Defining Words
1.26      crook    6092: @subsection User-defined Defining Words
                   6093: @cindex user-defined defining words
                   6094: @cindex defining words, user-defined
1.1       anton    6095: 
1.29      crook    6096: You can create a new defining word by wrapping defining-time code around
                   6097: an existing defining word and putting the sequence in a colon
1.69      anton    6098: definition. 
                   6099: 
                   6100: @c anton: This example is very complex and leads in a quite different
                   6101: @c direction from the CREATE-DOES> stuff that follows.  It should probably
                   6102: @c be done elsewhere, or as a subsubsection of this subsection (or as a
                   6103: @c subsection of Defining Words)
                   6104: 
                   6105: For example, suppose that you have a word @code{stats} that
1.29      crook    6106: gathers statistics about colon definitions given the @i{xt} of the
                   6107: definition, and you want every colon definition in your application to
                   6108: make a call to @code{stats}. You can define and use a new version of
                   6109: @code{:} like this:
                   6110: 
                   6111: @example
                   6112: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
                   6113:   ... ;  \ other code
                   6114: 
1.116     anton    6115: : my: : latestxt postpone literal ['] stats compile, ;
1.29      crook    6116: 
                   6117: my: foo + - ;
                   6118: @end example
                   6119: 
                   6120: When @code{foo} is defined using @code{my:} these steps occur:
                   6121: 
                   6122: @itemize @bullet
                   6123: @item
                   6124: @code{my:} is executed.
                   6125: @item
                   6126: The @code{:} within the definition (the one between @code{my:} and
1.116     anton    6127: @code{latestxt}) is executed, and does just what it always does; it parses
1.29      crook    6128: the input stream for a name, builds a dictionary header for the name
                   6129: @code{foo} and switches @code{state} from interpret to compile.
                   6130: @item
1.116     anton    6131: The word @code{latestxt} is executed. It puts the @i{xt} for the word that is
1.29      crook    6132: being defined -- @code{foo} -- onto the stack.
                   6133: @item
                   6134: The code that was produced by @code{postpone literal} is executed; this
                   6135: causes the value on the stack to be compiled as a literal in the code
                   6136: area of @code{foo}.
                   6137: @item
                   6138: The code @code{['] stats} compiles a literal into the definition of
                   6139: @code{my:}. When @code{compile,} is executed, that literal -- the
                   6140: execution token for @code{stats} -- is layed down in the code area of
                   6141: @code{foo} , following the literal@footnote{Strictly speaking, the
                   6142: mechanism that @code{compile,} uses to convert an @i{xt} into something
                   6143: in the code area is implementation-dependent. A threaded implementation
                   6144: might spit out the execution token directly whilst another
                   6145: implementation might spit out a native code sequence.}.
                   6146: @item
                   6147: At this point, the execution of @code{my:} is complete, and control
                   6148: returns to the text interpreter. The text interpreter is in compile
                   6149: state, so subsequent text @code{+ -} is compiled into the definition of
                   6150: @code{foo} and the @code{;} terminates the definition as always.
                   6151: @end itemize
                   6152: 
                   6153: You can use @code{see} to decompile a word that was defined using
                   6154: @code{my:} and see how it is different from a normal @code{:}
                   6155: definition. For example:
                   6156: 
                   6157: @example
                   6158: : bar + - ;  \ like foo but using : rather than my:
                   6159: see bar
                   6160: : bar
                   6161:   + - ;
                   6162: see foo
                   6163: : foo
                   6164:   107645672 stats + - ;
                   6165: 
1.140     anton    6166: \ use ' foo . to show that 107645672 is the xt for foo
1.29      crook    6167: @end example
                   6168: 
                   6169: You can use techniques like this to make new defining words in terms of
                   6170: @i{any} existing defining word.
1.1       anton    6171: 
                   6172: 
1.29      crook    6173: @cindex defining defining words
1.26      crook    6174: @cindex @code{CREATE} ... @code{DOES>}
                   6175: If you want the words defined with your defining words to behave
                   6176: differently from words defined with standard defining words, you can
                   6177: write your defining word like this:
1.1       anton    6178: 
                   6179: @example
1.26      crook    6180: : def-word ( "name" -- )
1.29      crook    6181:     CREATE @i{code1}
1.26      crook    6182: DOES> ( ... -- ... )
1.29      crook    6183:     @i{code2} ;
1.26      crook    6184: 
                   6185: def-word name
1.1       anton    6186: @end example
                   6187: 
1.29      crook    6188: @cindex child words
                   6189: This fragment defines a @dfn{defining word} @code{def-word} and then
                   6190: executes it.  When @code{def-word} executes, it @code{CREATE}s a new
                   6191: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
                   6192: is not executed at this time. The word @code{name} is sometimes called a
                   6193: @dfn{child} of @code{def-word}.
                   6194: 
                   6195: When you execute @code{name}, the address of the body of @code{name} is
                   6196: put on the data stack and @i{code2} is executed (the address of the body
                   6197: of @code{name} is the address @code{HERE} returns immediately after the
1.69      anton    6198: @code{CREATE}, i.e., the address a @code{create}d word returns by
                   6199: default).
                   6200: 
                   6201: @c anton:
                   6202: @c www.dictionary.com says:
                   6203: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
                   6204: @c several generations of absence, usually caused by the chance
                   6205: @c recombination of genes.  2.An individual or a part that exhibits
                   6206: @c atavism. Also called throwback.  3.The return of a trait or recurrence
                   6207: @c of previous behavior after a period of absence.
                   6208: @c
                   6209: @c Doesn't seem to fit.
1.29      crook    6210: 
1.69      anton    6211: @c @cindex atavism in child words
1.33      anton    6212: You can use @code{def-word} to define a set of child words that behave
1.69      anton    6213: similarly; they all have a common run-time behaviour determined by
                   6214: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
                   6215: body of the child word. The structure of the data is common to all
                   6216: children of @code{def-word}, but the data values are specific -- and
                   6217: private -- to each child word. When a child word is executed, the
                   6218: address of its private data area is passed as a parameter on TOS to be
                   6219: used and manipulated@footnote{It is legitimate both to read and write to
                   6220: this data area.} by @i{code2}.
1.29      crook    6221: 
                   6222: The two fragments of code that make up the defining words act (are
                   6223: executed) at two completely separate times:
1.1       anton    6224: 
1.29      crook    6225: @itemize @bullet
                   6226: @item
                   6227: At @i{define time}, the defining word executes @i{code1} to generate a
                   6228: child word
                   6229: @item
                   6230: At @i{child execution time}, when a child word is invoked, @i{code2}
                   6231: is executed, using parameters (data) that are private and specific to
                   6232: the child word.
                   6233: @end itemize
                   6234: 
1.44      crook    6235: Another way of understanding the behaviour of @code{def-word} and
                   6236: @code{name} is to say that, if you make the following definitions:
1.33      anton    6237: @example
                   6238: : def-word1 ( "name" -- )
                   6239:     CREATE @i{code1} ;
                   6240: 
                   6241: : action1 ( ... -- ... )
                   6242:     @i{code2} ;
                   6243: 
                   6244: def-word1 name1
                   6245: @end example
                   6246: 
1.44      crook    6247: @noindent
                   6248: Then using @code{name1 action1} is equivalent to using @code{name}.
1.1       anton    6249: 
1.29      crook    6250: The classic example is that you can define @code{CONSTANT} in this way:
1.26      crook    6251: 
1.1       anton    6252: @example
1.29      crook    6253: : CONSTANT ( w "name" -- )
                   6254:     CREATE ,
1.26      crook    6255: DOES> ( -- w )
                   6256:     @@ ;
1.1       anton    6257: @end example
                   6258: 
1.29      crook    6259: @comment There is a beautiful description of how this works and what
                   6260: @comment it does in the Forthwrite 100th edition.. as well as an elegant
                   6261: @comment commentary on the Counting Fruits problem.
                   6262: 
                   6263: When you create a constant with @code{5 CONSTANT five}, a set of
                   6264: define-time actions take place; first a new word @code{five} is created,
                   6265: then the value 5 is laid down in the body of @code{five} with
1.44      crook    6266: @code{,}. When @code{five} is executed, the address of the body is put on
1.29      crook    6267: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
                   6268: no code of its own; it simply contains a data field and a pointer to the
                   6269: code that follows @code{DOES>} in its defining word. That makes words
                   6270: created in this way very compact.
                   6271: 
                   6272: The final example in this section is intended to remind you that space
                   6273: reserved in @code{CREATE}d words is @i{data} space and therefore can be
                   6274: both read and written by a Standard program@footnote{Exercise: use this
                   6275: example as a starting point for your own implementation of @code{Value}
                   6276: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
                   6277: @code{[']}.}:
                   6278: 
                   6279: @example
                   6280: : foo ( "name" -- )
                   6281:     CREATE -1 ,
                   6282: DOES> ( -- )
1.33      anton    6283:     @@ . ;
1.29      crook    6284: 
                   6285: foo first-word
                   6286: foo second-word
                   6287: 
                   6288: 123 ' first-word >BODY !
                   6289: @end example
                   6290: 
                   6291: If @code{first-word} had been a @code{CREATE}d word, we could simply
                   6292: have executed it to get the address of its data field. However, since it
                   6293: was defined to have @code{DOES>} actions, its execution semantics are to
                   6294: perform those @code{DOES>} actions. To get the address of its data field
                   6295: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
                   6296: translate the xt into the address of the data field.  When you execute
                   6297: @code{first-word}, it will display @code{123}. When you execute
                   6298: @code{second-word} it will display @code{-1}.
1.26      crook    6299: 
                   6300: @cindex stack effect of @code{DOES>}-parts
                   6301: @cindex @code{DOES>}-parts, stack effect
1.29      crook    6302: In the examples above the stack comment after the @code{DOES>} specifies
1.26      crook    6303: the stack effect of the defined words, not the stack effect of the
                   6304: following code (the following code expects the address of the body on
                   6305: the top of stack, which is not reflected in the stack comment). This is
                   6306: the convention that I use and recommend (it clashes a bit with using
                   6307: locals declarations for stack effect specification, though).
1.1       anton    6308: 
1.53      anton    6309: @menu
                   6310: * CREATE..DOES> applications::  
                   6311: * CREATE..DOES> details::       
1.63      anton    6312: * Advanced does> usage example::  
1.91      anton    6313: * @code{Const-does>}::          
1.53      anton    6314: @end menu
                   6315: 
                   6316: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
1.26      crook    6317: @subsubsection Applications of @code{CREATE..DOES>}
                   6318: @cindex @code{CREATE} ... @code{DOES>}, applications
1.1       anton    6319: 
1.26      crook    6320: You may wonder how to use this feature. Here are some usage patterns:
1.1       anton    6321: 
1.26      crook    6322: @cindex factoring similar colon definitions
                   6323: When you see a sequence of code occurring several times, and you can
                   6324: identify a meaning, you will factor it out as a colon definition. When
                   6325: you see similar colon definitions, you can factor them using
                   6326: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
                   6327: that look very similar:
1.1       anton    6328: @example
1.26      crook    6329: : ori, ( reg-target reg-source n -- )
                   6330:     0 asm-reg-reg-imm ;
                   6331: : andi, ( reg-target reg-source n -- )
                   6332:     1 asm-reg-reg-imm ;
1.1       anton    6333: @end example
                   6334: 
1.26      crook    6335: @noindent
                   6336: This could be factored with:
                   6337: @example
                   6338: : reg-reg-imm ( op-code -- )
                   6339:     CREATE ,
                   6340: DOES> ( reg-target reg-source n -- )
                   6341:     @@ asm-reg-reg-imm ;
                   6342: 
                   6343: 0 reg-reg-imm ori,
                   6344: 1 reg-reg-imm andi,
                   6345: @end example
1.1       anton    6346: 
1.26      crook    6347: @cindex currying
                   6348: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
                   6349: supply a part of the parameters for a word (known as @dfn{currying} in
                   6350: the functional language community). E.g., @code{+} needs two
                   6351: parameters. Creating versions of @code{+} with one parameter fixed can
                   6352: be done like this:
1.82      anton    6353: 
1.1       anton    6354: @example
1.82      anton    6355: : curry+ ( n1 "name" -- )
1.26      crook    6356:     CREATE ,
                   6357: DOES> ( n2 -- n1+n2 )
                   6358:     @@ + ;
                   6359: 
                   6360:  3 curry+ 3+
                   6361: -2 curry+ 2-
1.1       anton    6362: @end example
                   6363: 
1.91      anton    6364: 
1.63      anton    6365: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
1.26      crook    6366: @subsubsection The gory details of @code{CREATE..DOES>}
                   6367: @cindex @code{CREATE} ... @code{DOES>}, details
1.1       anton    6368: 
1.26      crook    6369: doc-does>
1.1       anton    6370: 
1.26      crook    6371: @cindex @code{DOES>} in a separate definition
                   6372: This means that you need not use @code{CREATE} and @code{DOES>} in the
                   6373: same definition; you can put the @code{DOES>}-part in a separate
1.29      crook    6374: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
1.26      crook    6375: @example
                   6376: : does1 
                   6377: DOES> ( ... -- ... )
1.44      crook    6378:     ... ;
                   6379: 
                   6380: : does2
                   6381: DOES> ( ... -- ... )
                   6382:     ... ;
                   6383: 
                   6384: : def-word ( ... -- ... )
                   6385:     create ...
                   6386:     IF
                   6387:        does1
                   6388:     ELSE
                   6389:        does2
                   6390:     ENDIF ;
                   6391: @end example
                   6392: 
                   6393: In this example, the selection of whether to use @code{does1} or
1.69      anton    6394: @code{does2} is made at definition-time; at the time that the child word is
1.44      crook    6395: @code{CREATE}d.
                   6396: 
                   6397: @cindex @code{DOES>} in interpretation state
                   6398: In a standard program you can apply a @code{DOES>}-part only if the last
                   6399: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
                   6400: will override the behaviour of the last word defined in any case. In a
                   6401: standard program, you can use @code{DOES>} only in a colon
                   6402: definition. In Gforth, you can also use it in interpretation state, in a
                   6403: kind of one-shot mode; for example:
                   6404: @example
                   6405: CREATE name ( ... -- ... )
                   6406:   @i{initialization}
                   6407: DOES>
                   6408:   @i{code} ;
                   6409: @end example
                   6410: 
                   6411: @noindent
                   6412: is equivalent to the standard:
                   6413: @example
                   6414: :noname
                   6415: DOES>
                   6416:     @i{code} ;
                   6417: CREATE name EXECUTE ( ... -- ... )
                   6418:     @i{initialization}
                   6419: @end example
                   6420: 
1.53      anton    6421: doc->body
                   6422: 
1.91      anton    6423: @node Advanced does> usage example, @code{Const-does>}, CREATE..DOES> details, User-defined Defining Words
1.63      anton    6424: @subsubsection Advanced does> usage example
                   6425: 
                   6426: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
                   6427: for disassembling instructions, that follow a very repetetive scheme:
                   6428: 
                   6429: @example
                   6430: :noname @var{disasm-operands} s" @var{inst-name}" type ;
                   6431: @var{entry-num} cells @var{table} + !
                   6432: @end example
                   6433: 
                   6434: Of course, this inspires the idea to factor out the commonalities to
                   6435: allow a definition like
                   6436: 
                   6437: @example
                   6438: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
                   6439: @end example
                   6440: 
                   6441: The parameters @var{disasm-operands} and @var{table} are usually
1.69      anton    6442: correlated.  Moreover, before I wrote the disassembler, there already
                   6443: existed code that defines instructions like this:
1.63      anton    6444: 
                   6445: @example
                   6446: @var{entry-num} @var{inst-format} @var{inst-name}
                   6447: @end example
                   6448: 
                   6449: This code comes from the assembler and resides in
                   6450: @file{arch/mips/insts.fs}.
                   6451: 
                   6452: So I had to define the @var{inst-format} words that performed the scheme
                   6453: above when executed.  At first I chose to use run-time code-generation:
                   6454: 
                   6455: @example
                   6456: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
                   6457:   :noname Postpone @var{disasm-operands}
                   6458:   name Postpone sliteral Postpone type Postpone ;
                   6459:   swap cells @var{table} + ! ;
                   6460: @end example
                   6461: 
                   6462: Note that this supplies the other two parameters of the scheme above.
1.44      crook    6463: 
1.63      anton    6464: An alternative would have been to write this using
                   6465: @code{create}/@code{does>}:
                   6466: 
                   6467: @example
                   6468: : @var{inst-format} ( entry-num "name" -- )
                   6469:   here name string, ( entry-num c-addr ) \ parse and save "name"
                   6470:   noname create , ( entry-num )
1.116     anton    6471:   latestxt swap cells @var{table} + !
1.63      anton    6472: does> ( addr w -- )
                   6473:   \ disassemble instruction w at addr
                   6474:   @@ >r 
                   6475:   @var{disasm-operands}
                   6476:   r> count type ;
                   6477: @end example
                   6478: 
                   6479: Somehow the first solution is simpler, mainly because it's simpler to
                   6480: shift a string from definition-time to use-time with @code{sliteral}
                   6481: than with @code{string,} and friends.
                   6482: 
                   6483: I wrote a lot of words following this scheme and soon thought about
                   6484: factoring out the commonalities among them.  Note that this uses a
                   6485: two-level defining word, i.e., a word that defines ordinary defining
                   6486: words.
                   6487: 
                   6488: This time a solution involving @code{postpone} and friends seemed more
                   6489: difficult (try it as an exercise), so I decided to use a
                   6490: @code{create}/@code{does>} word; since I was already at it, I also used
                   6491: @code{create}/@code{does>} for the lower level (try using
                   6492: @code{postpone} etc. as an exercise), resulting in the following
                   6493: definition:
                   6494: 
                   6495: @example
                   6496: : define-format ( disasm-xt table-xt -- )
                   6497:     \ define an instruction format that uses disasm-xt for
                   6498:     \ disassembling and enters the defined instructions into table
                   6499:     \ table-xt
                   6500:     create 2,
                   6501: does> ( u "inst" -- )
                   6502:     \ defines an anonymous word for disassembling instruction inst,
                   6503:     \ and enters it as u-th entry into table-xt
                   6504:     2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
                   6505:     noname create 2,      \ define anonymous word
1.116     anton    6506:     execute latestxt swap ! \ enter xt of defined word into table-xt
1.63      anton    6507: does> ( addr w -- )
                   6508:     \ disassemble instruction w at addr
                   6509:     2@@ >r ( addr w disasm-xt R: c-addr )
                   6510:     execute ( R: c-addr ) \ disassemble operands
                   6511:     r> count type ; \ print name 
                   6512: @end example
                   6513: 
                   6514: Note that the tables here (in contrast to above) do the @code{cells +}
                   6515: by themselves (that's why you have to pass an xt).  This word is used in
                   6516: the following way:
                   6517: 
                   6518: @example
                   6519: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
                   6520: @end example
                   6521: 
1.71      anton    6522: As shown above, the defined instruction format is then used like this:
                   6523: 
                   6524: @example
                   6525: @var{entry-num} @var{inst-format} @var{inst-name}
                   6526: @end example
                   6527: 
1.63      anton    6528: In terms of currying, this kind of two-level defining word provides the
                   6529: parameters in three stages: first @var{disasm-operands} and @var{table},
                   6530: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
                   6531: the instruction to be disassembled.  
                   6532: 
                   6533: Of course this did not quite fit all the instruction format names used
                   6534: in @file{insts.fs}, so I had to define a few wrappers that conditioned
                   6535: the parameters into the right form.
                   6536: 
                   6537: If you have trouble following this section, don't worry.  First, this is
                   6538: involved and takes time (and probably some playing around) to
                   6539: understand; second, this is the first two-level
                   6540: @code{create}/@code{does>} word I have written in seventeen years of
                   6541: Forth; and if I did not have @file{insts.fs} to start with, I may well
                   6542: have elected to use just a one-level defining word (with some repeating
                   6543: of parameters when using the defining word). So it is not necessary to
                   6544: understand this, but it may improve your understanding of Forth.
1.44      crook    6545: 
                   6546: 
1.91      anton    6547: @node @code{Const-does>},  , Advanced does> usage example, User-defined Defining Words
                   6548: @subsubsection @code{Const-does>}
                   6549: 
                   6550: A frequent use of @code{create}...@code{does>} is for transferring some
                   6551: values from definition-time to run-time.  Gforth supports this use with
                   6552: 
                   6553: doc-const-does>
                   6554: 
                   6555: A typical use of this word is:
                   6556: 
                   6557: @example
                   6558: : curry+ ( n1 "name" -- )
                   6559: 1 0 CONST-DOES> ( n2 -- n1+n2 )
                   6560:     + ;
                   6561: 
                   6562: 3 curry+ 3+
                   6563: @end example
                   6564: 
                   6565: Here the @code{1 0} means that 1 cell and 0 floats are transferred from
                   6566: definition to run-time.
                   6567: 
                   6568: The advantages of using @code{const-does>} are:
                   6569: 
                   6570: @itemize
                   6571: 
                   6572: @item
                   6573: You don't have to deal with storing and retrieving the values, i.e.,
                   6574: your program becomes more writable and readable.
                   6575: 
                   6576: @item
                   6577: When using @code{does>}, you have to introduce a @code{@@} that cannot
                   6578: be optimized away (because you could change the data using
                   6579: @code{>body}...@code{!}); @code{const-does>} avoids this problem.
                   6580: 
                   6581: @end itemize
                   6582: 
                   6583: An ANS Forth implementation of @code{const-does>} is available in
                   6584: @file{compat/const-does.fs}.
                   6585: 
                   6586: 
1.44      crook    6587: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
                   6588: @subsection Deferred words
                   6589: @cindex deferred words
                   6590: 
                   6591: The defining word @code{Defer} allows you to define a word by name
                   6592: without defining its behaviour; the definition of its behaviour is
                   6593: deferred. Here are two situation where this can be useful:
                   6594: 
                   6595: @itemize @bullet
                   6596: @item
                   6597: Where you want to allow the behaviour of a word to be altered later, and
                   6598: for all precompiled references to the word to change when its behaviour
                   6599: is changed.
                   6600: @item
                   6601: For mutual recursion; @xref{Calls and returns}.
                   6602: @end itemize
                   6603: 
                   6604: In the following example, @code{foo} always invokes the version of
                   6605: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
                   6606: always invokes the version that prints ``@code{Hello}''. There is no way
                   6607: of getting @code{foo} to use the later version without re-ordering the
                   6608: source code and recompiling it.
                   6609: 
                   6610: @example
                   6611: : greet ." Good morning" ;
                   6612: : foo ... greet ... ;
                   6613: : greet ." Hello" ;
                   6614: : bar ... greet ... ;
                   6615: @end example
                   6616: 
                   6617: This problem can be solved by defining @code{greet} as a @code{Defer}red
                   6618: word. The behaviour of a @code{Defer}red word can be defined and
                   6619: redefined at any time by using @code{IS} to associate the xt of a
                   6620: previously-defined word with it. The previous example becomes:
                   6621: 
                   6622: @example
1.69      anton    6623: Defer greet ( -- )
1.44      crook    6624: : foo ... greet ... ;
                   6625: : bar ... greet ... ;
1.69      anton    6626: : greet1 ( -- ) ." Good morning" ;
                   6627: : greet2 ( -- ) ." Hello" ;
1.132     anton    6628: ' greet2 IS greet  \ make greet behave like greet2
1.44      crook    6629: @end example
                   6630: 
1.69      anton    6631: @progstyle
                   6632: You should write a stack comment for every deferred word, and put only
                   6633: XTs into deferred words that conform to this stack effect.  Otherwise
                   6634: it's too difficult to use the deferred word.
                   6635: 
1.44      crook    6636: A deferred word can be used to improve the statistics-gathering example
                   6637: from @ref{User-defined Defining Words}; rather than edit the
                   6638: application's source code to change every @code{:} to a @code{my:}, do
                   6639: this:
                   6640: 
                   6641: @example
                   6642: : real: : ;     \ retain access to the original
                   6643: defer :         \ redefine as a deferred word
1.132     anton    6644: ' my: IS :      \ use special version of :
1.44      crook    6645: \
                   6646: \ load application here
                   6647: \
1.132     anton    6648: ' real: IS :    \ go back to the original
1.44      crook    6649: @end example
                   6650: 
                   6651: 
1.132     anton    6652: One thing to note is that @code{IS} has special compilation semantics,
                   6653: such that it parses the name at compile time (like @code{TO}):
1.44      crook    6654: 
                   6655: @example
                   6656: : set-greet ( xt -- )
1.132     anton    6657:   IS greet ;
1.44      crook    6658: 
                   6659: ' greet1 set-greet
                   6660: @end example
                   6661: 
1.132     anton    6662: In situations where @code{IS} does not fit, use @code{defer!} instead.
                   6663: 
1.69      anton    6664: A deferred word can only inherit execution semantics from the xt
                   6665: (because that is all that an xt can represent -- for more discussion of
                   6666: this @pxref{Tokens for Words}); by default it will have default
                   6667: interpretation and compilation semantics deriving from this execution
                   6668: semantics.  However, you can change the interpretation and compilation
                   6669: semantics of the deferred word in the usual ways:
1.44      crook    6670: 
                   6671: @example
1.132     anton    6672: : bar .... ; immediate
1.44      crook    6673: Defer fred immediate
                   6674: Defer jim
                   6675: 
1.132     anton    6676: ' bar IS jim  \ jim has default semantics
                   6677: ' bar IS fred \ fred is immediate
1.44      crook    6678: @end example
                   6679: 
                   6680: doc-defer
1.132     anton    6681: doc-defer!
1.44      crook    6682: doc-is
1.132     anton    6683: doc-defer@
                   6684: doc-action-of
1.44      crook    6685: @comment TODO document these: what's defers [is]
                   6686: doc-defers
                   6687: 
                   6688: @c Use @code{words-deferred} to see a list of deferred words.
                   6689: 
1.132     anton    6690: Definitions of these words (except @code{defers}) in ANS Forth are
                   6691: provided in @file{compat/defer.fs}.
1.44      crook    6692: 
                   6693: 
1.69      anton    6694: @node Aliases,  , Deferred words, Defining Words
1.44      crook    6695: @subsection Aliases
                   6696: @cindex aliases
1.1       anton    6697: 
1.44      crook    6698: The defining word @code{Alias} allows you to define a word by name that
                   6699: has the same behaviour as some other word. Here are two situation where
                   6700: this can be useful:
1.1       anton    6701: 
1.44      crook    6702: @itemize @bullet
                   6703: @item
                   6704: When you want access to a word's definition from a different word list
                   6705: (for an example of this, see the definition of the @code{Root} word list
                   6706: in the Gforth source).
                   6707: @item
                   6708: When you want to create a synonym; a definition that can be known by
                   6709: either of two names (for example, @code{THEN} and @code{ENDIF} are
                   6710: aliases).
                   6711: @end itemize
1.1       anton    6712: 
1.69      anton    6713: Like deferred words, an alias has default compilation and interpretation
                   6714: semantics at the beginning (not the modifications of the other word),
                   6715: but you can change them in the usual ways (@code{immediate},
                   6716: @code{compile-only}). For example:
1.1       anton    6717: 
                   6718: @example
1.44      crook    6719: : foo ... ; immediate
                   6720: 
                   6721: ' foo Alias bar \ bar is not an immediate word
                   6722: ' foo Alias fooby immediate \ fooby is an immediate word
1.1       anton    6723: @end example
                   6724: 
1.44      crook    6725: Words that are aliases have the same xt, different headers in the
                   6726: dictionary, and consequently different name tokens (@pxref{Tokens for
                   6727: Words}) and possibly different immediate flags.  An alias can only have
                   6728: default or immediate compilation semantics; you can define aliases for
                   6729: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
1.1       anton    6730: 
1.44      crook    6731: doc-alias
1.1       anton    6732: 
                   6733: 
1.47      crook    6734: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
                   6735: @section Interpretation and Compilation Semantics
1.26      crook    6736: @cindex semantics, interpretation and compilation
1.1       anton    6737: 
1.71      anton    6738: @c !! state and ' are used without explanation
                   6739: @c example for immediate/compile-only? or is the tutorial enough
                   6740: 
1.26      crook    6741: @cindex interpretation semantics
1.71      anton    6742: The @dfn{interpretation semantics} of a (named) word are what the text
1.26      crook    6743: interpreter does when it encounters the word in interpret state. It also
                   6744: appears in some other contexts, e.g., the execution token returned by
1.71      anton    6745: @code{' @i{word}} identifies the interpretation semantics of @i{word}
                   6746: (in other words, @code{' @i{word} execute} is equivalent to
1.29      crook    6747: interpret-state text interpretation of @code{@i{word}}).
1.1       anton    6748: 
1.26      crook    6749: @cindex compilation semantics
1.71      anton    6750: The @dfn{compilation semantics} of a (named) word are what the text
                   6751: interpreter does when it encounters the word in compile state. It also
                   6752: appears in other contexts, e.g, @code{POSTPONE @i{word}}
                   6753: compiles@footnote{In standard terminology, ``appends to the current
                   6754: definition''.} the compilation semantics of @i{word}.
1.1       anton    6755: 
1.26      crook    6756: @cindex execution semantics
                   6757: The standard also talks about @dfn{execution semantics}. They are used
                   6758: only for defining the interpretation and compilation semantics of many
                   6759: words. By default, the interpretation semantics of a word are to
                   6760: @code{execute} its execution semantics, and the compilation semantics of
                   6761: a word are to @code{compile,} its execution semantics.@footnote{In
                   6762: standard terminology: The default interpretation semantics are its
                   6763: execution semantics; the default compilation semantics are to append its
                   6764: execution semantics to the execution semantics of the current
                   6765: definition.}
                   6766: 
1.71      anton    6767: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
                   6768: the text interpreter, ticked, or @code{postpone}d, so they have no
                   6769: interpretation or compilation semantics.  Their behaviour is represented
                   6770: by their XT (@pxref{Tokens for Words}), and we call it execution
                   6771: semantics, too.
                   6772: 
1.26      crook    6773: @comment TODO expand, make it co-operate with new sections on text interpreter.
                   6774: 
                   6775: @cindex immediate words
                   6776: @cindex compile-only words
                   6777: You can change the semantics of the most-recently defined word:
                   6778: 
1.44      crook    6779: 
1.26      crook    6780: doc-immediate
                   6781: doc-compile-only
                   6782: doc-restrict
                   6783: 
1.82      anton    6784: By convention, words with non-default compilation semantics (e.g.,
                   6785: immediate words) often have names surrounded with brackets (e.g.,
                   6786: @code{[']}, @pxref{Execution token}).
1.44      crook    6787: 
1.26      crook    6788: Note that ticking (@code{'}) a compile-only word gives an error
                   6789: (``Interpreting a compile-only word'').
1.1       anton    6790: 
1.47      crook    6791: @menu
1.67      anton    6792: * Combined words::              
1.47      crook    6793: @end menu
1.44      crook    6794: 
1.71      anton    6795: 
1.48      anton    6796: @node Combined words,  , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
1.44      crook    6797: @subsection Combined Words
                   6798: @cindex combined words
                   6799: 
                   6800: Gforth allows you to define @dfn{combined words} -- words that have an
                   6801: arbitrary combination of interpretation and compilation semantics.
                   6802: 
1.26      crook    6803: doc-interpret/compile:
1.1       anton    6804: 
1.26      crook    6805: This feature was introduced for implementing @code{TO} and @code{S"}. I
                   6806: recommend that you do not define such words, as cute as they may be:
                   6807: they make it hard to get at both parts of the word in some contexts.
                   6808: E.g., assume you want to get an execution token for the compilation
                   6809: part. Instead, define two words, one that embodies the interpretation
                   6810: part, and one that embodies the compilation part.  Once you have done
                   6811: that, you can define a combined word with @code{interpret/compile:} for
                   6812: the convenience of your users.
1.1       anton    6813: 
1.26      crook    6814: You might try to use this feature to provide an optimizing
                   6815: implementation of the default compilation semantics of a word. For
                   6816: example, by defining:
1.1       anton    6817: @example
1.26      crook    6818: :noname
                   6819:    foo bar ;
                   6820: :noname
                   6821:    POSTPONE foo POSTPONE bar ;
1.29      crook    6822: interpret/compile: opti-foobar
1.1       anton    6823: @end example
1.26      crook    6824: 
1.23      crook    6825: @noindent
1.26      crook    6826: as an optimizing version of:
                   6827: 
1.1       anton    6828: @example
1.26      crook    6829: : foobar
                   6830:     foo bar ;
1.1       anton    6831: @end example
                   6832: 
1.26      crook    6833: Unfortunately, this does not work correctly with @code{[compile]},
                   6834: because @code{[compile]} assumes that the compilation semantics of all
                   6835: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
1.29      crook    6836: opti-foobar} would compile compilation semantics, whereas
                   6837: @code{[compile] foobar} would compile interpretation semantics.
1.1       anton    6838: 
1.26      crook    6839: @cindex state-smart words (are a bad idea)
1.82      anton    6840: @anchor{state-smartness}
1.29      crook    6841: Some people try to use @dfn{state-smart} words to emulate the feature provided
1.26      crook    6842: by @code{interpret/compile:} (words are state-smart if they check
                   6843: @code{STATE} during execution). E.g., they would try to code
                   6844: @code{foobar} like this:
1.1       anton    6845: 
1.26      crook    6846: @example
                   6847: : foobar
                   6848:   STATE @@
                   6849:   IF ( compilation state )
                   6850:     POSTPONE foo POSTPONE bar
                   6851:   ELSE
                   6852:     foo bar
                   6853:   ENDIF ; immediate
                   6854: @end example
1.1       anton    6855: 
1.26      crook    6856: Although this works if @code{foobar} is only processed by the text
                   6857: interpreter, it does not work in other contexts (like @code{'} or
                   6858: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
                   6859: for a state-smart word, not for the interpretation semantics of the
                   6860: original @code{foobar}; when you execute this execution token (directly
                   6861: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
                   6862: state, the result will not be what you expected (i.e., it will not
                   6863: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
                   6864: write them@footnote{For a more detailed discussion of this topic, see
1.66      anton    6865: M. Anton Ertl,
                   6866: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
                   6867: it is Evil and How to Exorcise it}}, EuroForth '98.}!
1.1       anton    6868: 
1.26      crook    6869: @cindex defining words with arbitrary semantics combinations
                   6870: It is also possible to write defining words that define words with
                   6871: arbitrary combinations of interpretation and compilation semantics. In
                   6872: general, they look like this:
1.1       anton    6873: 
1.26      crook    6874: @example
                   6875: : def-word
                   6876:     create-interpret/compile
1.29      crook    6877:     @i{code1}
1.26      crook    6878: interpretation>
1.29      crook    6879:     @i{code2}
1.26      crook    6880: <interpretation
                   6881: compilation>
1.29      crook    6882:     @i{code3}
1.26      crook    6883: <compilation ;
                   6884: @end example
1.1       anton    6885: 
1.29      crook    6886: For a @i{word} defined with @code{def-word}, the interpretation
                   6887: semantics are to push the address of the body of @i{word} and perform
                   6888: @i{code2}, and the compilation semantics are to push the address of
                   6889: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
1.26      crook    6890: can also be defined like this (except that the defined constants don't
                   6891: behave correctly when @code{[compile]}d):
1.1       anton    6892: 
1.26      crook    6893: @example
                   6894: : constant ( n "name" -- )
                   6895:     create-interpret/compile
                   6896:     ,
                   6897: interpretation> ( -- n )
                   6898:     @@
                   6899: <interpretation
                   6900: compilation> ( compilation. -- ; run-time. -- n )
                   6901:     @@ postpone literal
                   6902: <compilation ;
                   6903: @end example
1.1       anton    6904: 
1.44      crook    6905: 
1.26      crook    6906: doc-create-interpret/compile
                   6907: doc-interpretation>
                   6908: doc-<interpretation
                   6909: doc-compilation>
                   6910: doc-<compilation
1.1       anton    6911: 
1.44      crook    6912: 
1.29      crook    6913: Words defined with @code{interpret/compile:} and
1.26      crook    6914: @code{create-interpret/compile} have an extended header structure that
                   6915: differs from other words; however, unless you try to access them with
                   6916: plain address arithmetic, you should not notice this. Words for
                   6917: accessing the header structure usually know how to deal with this; e.g.,
1.29      crook    6918: @code{'} @i{word} @code{>body} also gives you the body of a word created
                   6919: with @code{create-interpret/compile}.
1.1       anton    6920: 
1.44      crook    6921: 
1.47      crook    6922: @c -------------------------------------------------------------
1.81      anton    6923: @node Tokens for Words, Compiling words, Interpretation and Compilation Semantics, Words
1.47      crook    6924: @section Tokens for Words
                   6925: @cindex tokens for words
                   6926: 
                   6927: This section describes the creation and use of tokens that represent
                   6928: words.
                   6929: 
1.71      anton    6930: @menu
                   6931: * Execution token::             represents execution/interpretation semantics
                   6932: * Compilation token::           represents compilation semantics
                   6933: * Name token::                  represents named words
                   6934: @end menu
1.47      crook    6935: 
1.71      anton    6936: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
                   6937: @subsection Execution token
1.47      crook    6938: 
                   6939: @cindex xt
                   6940: @cindex execution token
1.71      anton    6941: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
                   6942: You can use @code{execute} to invoke this behaviour.
1.47      crook    6943: 
1.71      anton    6944: @cindex tick (')
                   6945: You can use @code{'} to get an execution token that represents the
                   6946: interpretation semantics of a named word:
1.47      crook    6947: 
                   6948: @example
1.97      anton    6949: 5 ' .   ( n xt ) 
                   6950: execute ( )      \ execute the xt (i.e., ".")
1.71      anton    6951: @end example
1.47      crook    6952: 
1.71      anton    6953: doc-'
                   6954: 
                   6955: @code{'} parses at run-time; there is also a word @code{[']} that parses
                   6956: when it is compiled, and compiles the resulting XT:
                   6957: 
                   6958: @example
                   6959: : foo ['] . execute ;
                   6960: 5 foo
                   6961: : bar ' execute ; \ by contrast,
                   6962: 5 bar .           \ ' parses "." when bar executes
                   6963: @end example
                   6964: 
                   6965: doc-[']
                   6966: 
                   6967: If you want the execution token of @i{word}, write @code{['] @i{word}}
                   6968: in compiled code and @code{' @i{word}} in interpreted code.  Gforth's
                   6969: @code{'} and @code{[']} behave somewhat unusually by complaining about
                   6970: compile-only words (because these words have no interpretation
                   6971: semantics).  You might get what you want by using @code{COMP' @i{word}
                   6972: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
                   6973: token}).
                   6974: 
1.116     anton    6975: Another way to get an XT is @code{:noname} or @code{latestxt}
1.71      anton    6976: (@pxref{Anonymous Definitions}).  For anonymous words this gives an xt
                   6977: for the only behaviour the word has (the execution semantics).  For
1.116     anton    6978: named words, @code{latestxt} produces an XT for the same behaviour it
1.71      anton    6979: would produce if the word was defined anonymously.
                   6980: 
                   6981: @example
                   6982: :noname ." hello" ;
                   6983: execute
1.47      crook    6984: @end example
                   6985: 
1.71      anton    6986: An XT occupies one cell and can be manipulated like any other cell.
                   6987: 
1.47      crook    6988: @cindex code field address
                   6989: @cindex CFA
1.71      anton    6990: In ANS Forth the XT is just an abstract data type (i.e., defined by the
                   6991: operations that produce or consume it).  For old hands: In Gforth, the
                   6992: XT is implemented as a code field address (CFA).
                   6993: 
                   6994: doc-execute
                   6995: doc-perform
                   6996: 
                   6997: @node Compilation token, Name token, Execution token, Tokens for Words
                   6998: @subsection Compilation token
1.47      crook    6999: 
                   7000: @cindex compilation token
1.71      anton    7001: @cindex CT (compilation token)
                   7002: Gforth represents the compilation semantics of a named word by a
1.47      crook    7003: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
                   7004: @i{xt} is an execution token. The compilation semantics represented by
                   7005: the compilation token can be performed with @code{execute}, which
                   7006: consumes the whole compilation token, with an additional stack effect
                   7007: determined by the represented compilation semantics.
                   7008: 
                   7009: At present, the @i{w} part of a compilation token is an execution token,
                   7010: and the @i{xt} part represents either @code{execute} or
                   7011: @code{compile,}@footnote{Depending upon the compilation semantics of the
                   7012: word. If the word has default compilation semantics, the @i{xt} will
                   7013: represent @code{compile,}. Otherwise (e.g., for immediate words), the
                   7014: @i{xt} will represent @code{execute}.}. However, don't rely on that
                   7015: knowledge, unless necessary; future versions of Gforth may introduce
                   7016: unusual compilation tokens (e.g., a compilation token that represents
                   7017: the compilation semantics of a literal).
                   7018: 
1.71      anton    7019: You can perform the compilation semantics represented by the compilation
                   7020: token with @code{execute}.  You can compile the compilation semantics
                   7021: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
                   7022: equivalent to @code{postpone @i{word}}.
                   7023: 
                   7024: doc-[comp']
                   7025: doc-comp'
                   7026: doc-postpone,
                   7027: 
                   7028: @node Name token,  , Compilation token, Tokens for Words
                   7029: @subsection Name token
1.47      crook    7030: 
                   7031: @cindex name token
1.116     anton    7032: Gforth represents named words by the @dfn{name token}, (@i{nt}).  Name
                   7033: token is an abstract data type that occurs as argument or result of the
                   7034: words below.
                   7035: 
                   7036: @c !! put this elswhere?
1.47      crook    7037: @cindex name field address
                   7038: @cindex NFA
1.116     anton    7039: The closest thing to the nt in older Forth systems is the name field
                   7040: address (NFA), but there are significant differences: in older Forth
                   7041: systems each word had a unique NFA, LFA, CFA and PFA (in this order, or
                   7042: LFA, NFA, CFA, PFA) and there were words for getting from one to the
                   7043: next.  In contrast, in Gforth 0@dots{}n nts correspond to one xt; there
                   7044: is a link field in the structure identified by the name token, but
                   7045: searching usually uses a hash table external to these structures; the
                   7046: name in Gforth has a cell-wide count-and-flags field, and the nt is not
                   7047: implemented as the address of that count field.
1.47      crook    7048: 
                   7049: doc-find-name
1.116     anton    7050: doc-latest
                   7051: doc->name
1.47      crook    7052: doc-name>int
                   7053: doc-name?int
                   7054: doc-name>comp
                   7055: doc-name>string
1.109     anton    7056: doc-id.
                   7057: doc-.name
                   7058: doc-.id
1.47      crook    7059: 
1.81      anton    7060: @c ----------------------------------------------------------
                   7061: @node Compiling words, The Text Interpreter, Tokens for Words, Words
                   7062: @section Compiling words
                   7063: @cindex compiling words
                   7064: @cindex macros
                   7065: 
                   7066: In contrast to most other languages, Forth has no strict boundary
1.82      anton    7067: between compilation and run-time.  E.g., you can run arbitrary code
                   7068: between defining words (or for computing data used by defining words
                   7069: like @code{constant}). Moreover, @code{Immediate} (@pxref{Interpretation
                   7070: and Compilation Semantics} and @code{[}...@code{]} (see below) allow
                   7071: running arbitrary code while compiling a colon definition (exception:
                   7072: you must not allot dictionary space).
                   7073: 
                   7074: @menu
                   7075: * Literals::                    Compiling data values
                   7076: * Macros::                      Compiling words
                   7077: @end menu
                   7078: 
                   7079: @node Literals, Macros, Compiling words, Compiling words
                   7080: @subsection Literals
                   7081: @cindex Literals
                   7082: 
                   7083: The simplest and most frequent example is to compute a literal during
                   7084: compilation.  E.g., the following definition prints an array of strings,
                   7085: one string per line:
                   7086: 
                   7087: @example
                   7088: : .strings ( addr u -- ) \ gforth
                   7089:     2* cells bounds U+DO
                   7090:        cr i 2@@ type
                   7091:     2 cells +LOOP ;  
                   7092: @end example
1.81      anton    7093: 
1.82      anton    7094: With a simple-minded compiler like Gforth's, this computes @code{2
                   7095: cells} on every loop iteration.  You can compute this value once and for
                   7096: all at compile time and compile it into the definition like this:
                   7097: 
                   7098: @example
                   7099: : .strings ( addr u -- ) \ gforth
                   7100:     2* cells bounds U+DO
                   7101:        cr i 2@@ type
                   7102:     [ 2 cells ] literal +LOOP ;  
                   7103: @end example
                   7104: 
                   7105: @code{[} switches the text interpreter to interpret state (you will get
                   7106: an @code{ok} prompt if you type this example interactively and insert a
                   7107: newline between @code{[} and @code{]}), so it performs the
                   7108: interpretation semantics of @code{2 cells}; this computes a number.
                   7109: @code{]} switches the text interpreter back into compile state.  It then
                   7110: performs @code{Literal}'s compilation semantics, which are to compile
                   7111: this number into the current word.  You can decompile the word with
                   7112: @code{see .strings} to see the effect on the compiled code.
1.81      anton    7113: 
1.82      anton    7114: You can also optimize the @code{2* cells} into @code{[ 2 cells ] literal
                   7115: *} in this way.
1.81      anton    7116: 
1.82      anton    7117: doc-[
                   7118: doc-]
1.81      anton    7119: doc-literal
                   7120: doc-]L
1.82      anton    7121: 
                   7122: There are also words for compiling other data types than single cells as
                   7123: literals:
                   7124: 
1.81      anton    7125: doc-2literal
                   7126: doc-fliteral
1.82      anton    7127: doc-sliteral
                   7128: 
                   7129: @cindex colon-sys, passing data across @code{:}
                   7130: @cindex @code{:}, passing data across
                   7131: You might be tempted to pass data from outside a colon definition to the
                   7132: inside on the data stack.  This does not work, because @code{:} puhes a
                   7133: colon-sys, making stuff below unaccessible.  E.g., this does not work:
                   7134: 
                   7135: @example
                   7136: 5 : foo literal ; \ error: "unstructured"
                   7137: @end example
                   7138: 
                   7139: Instead, you have to pass the value in some other way, e.g., through a
                   7140: variable:
                   7141: 
                   7142: @example
                   7143: variable temp
                   7144: 5 temp !
                   7145: : foo [ temp @@ ] literal ;
                   7146: @end example
                   7147: 
                   7148: 
                   7149: @node Macros,  , Literals, Compiling words
                   7150: @subsection Macros
                   7151: @cindex Macros
                   7152: @cindex compiling compilation semantics
                   7153: 
                   7154: @code{Literal} and friends compile data values into the current
                   7155: definition.  You can also write words that compile other words into the
                   7156: current definition.  E.g.,
                   7157: 
                   7158: @example
                   7159: : compile-+ ( -- ) \ compiled code: ( n1 n2 -- n )
                   7160:   POSTPONE + ;
                   7161: 
                   7162: : foo ( n1 n2 -- n )
                   7163:   [ compile-+ ] ;
                   7164: 1 2 foo .
                   7165: @end example
                   7166: 
                   7167: This is equivalent to @code{: foo + ;} (@code{see foo} to check this).
                   7168: What happens in this example?  @code{Postpone} compiles the compilation
                   7169: semantics of @code{+} into @code{compile-+}; later the text interpreter
                   7170: executes @code{compile-+} and thus the compilation semantics of +, which
                   7171: compile (the execution semantics of) @code{+} into
                   7172: @code{foo}.@footnote{A recent RFI answer requires that compiling words
                   7173: should only be executed in compile state, so this example is not
                   7174: guaranteed to work on all standard systems, but on any decent system it
                   7175: will work.}
                   7176: 
                   7177: doc-postpone
                   7178: doc-[compile]
                   7179: 
                   7180: Compiling words like @code{compile-+} are usually immediate (or similar)
                   7181: so you do not have to switch to interpret state to execute them;
                   7182: mopifying the last example accordingly produces:
                   7183: 
                   7184: @example
                   7185: : [compile-+] ( compilation: --; interpretation: -- )
                   7186:   \ compiled code: ( n1 n2 -- n )
                   7187:   POSTPONE + ; immediate
                   7188: 
                   7189: : foo ( n1 n2 -- n )
                   7190:   [compile-+] ;
                   7191: 1 2 foo .
                   7192: @end example
                   7193: 
                   7194: Immediate compiling words are similar to macros in other languages (in
                   7195: particular, Lisp).  The important differences to macros in, e.g., C are:
                   7196: 
                   7197: @itemize @bullet
                   7198: 
                   7199: @item
                   7200: You use the same language for defining and processing macros, not a
                   7201: separate preprocessing language and processor.
                   7202: 
                   7203: @item
                   7204: Consequently, the full power of Forth is available in macro definitions.
                   7205: E.g., you can perform arbitrarily complex computations, or generate
                   7206: different code conditionally or in a loop (e.g., @pxref{Advanced macros
                   7207: Tutorial}).  This power is very useful when writing a parser generators
                   7208: or other code-generating software.
                   7209: 
                   7210: @item
                   7211: Macros defined using @code{postpone} etc. deal with the language at a
                   7212: higher level than strings; name binding happens at macro definition
                   7213: time, so you can avoid the pitfalls of name collisions that can happen
                   7214: in C macros.  Of course, Forth is a liberal language and also allows to
                   7215: shoot yourself in the foot with text-interpreted macros like
                   7216: 
                   7217: @example
                   7218: : [compile-+] s" +" evaluate ; immediate
                   7219: @end example
                   7220: 
                   7221: Apart from binding the name at macro use time, using @code{evaluate}
                   7222: also makes your definition @code{state}-smart (@pxref{state-smartness}).
                   7223: @end itemize
                   7224: 
                   7225: You may want the macro to compile a number into a word.  The word to do
                   7226: it is @code{literal}, but you have to @code{postpone} it, so its
                   7227: compilation semantics take effect when the macro is executed, not when
                   7228: it is compiled:
                   7229: 
                   7230: @example
                   7231: : [compile-5] ( -- ) \ compiled code: ( -- n )
                   7232:   5 POSTPONE literal ; immediate
                   7233: 
                   7234: : foo [compile-5] ;
                   7235: foo .
                   7236: @end example
                   7237: 
                   7238: You may want to pass parameters to a macro, that the macro should
                   7239: compile into the current definition.  If the parameter is a number, then
                   7240: you can use @code{postpone literal} (similar for other values).
                   7241: 
                   7242: If you want to pass a word that is to be compiled, the usual way is to
                   7243: pass an execution token and @code{compile,} it:
                   7244: 
                   7245: @example
                   7246: : twice1 ( xt -- ) \ compiled code: ... -- ...
                   7247:   dup compile, compile, ;
                   7248: 
                   7249: : 2+ ( n1 -- n2 )
                   7250:   [ ' 1+ twice1 ] ;
                   7251: @end example
                   7252: 
                   7253: doc-compile,
                   7254: 
                   7255: An alternative available in Gforth, that allows you to pass compile-only
                   7256: words as parameters is to use the compilation token (@pxref{Compilation
                   7257: token}).  The same example in this technique:
                   7258: 
                   7259: @example
                   7260: : twice ( ... ct -- ... ) \ compiled code: ... -- ...
                   7261:   2dup 2>r execute 2r> execute ;
                   7262: 
                   7263: : 2+ ( n1 -- n2 )
                   7264:   [ comp' 1+ twice ] ;
                   7265: @end example
                   7266: 
                   7267: In the example above @code{2>r} and @code{2r>} ensure that @code{twice}
                   7268: works even if the executed compilation semantics has an effect on the
                   7269: data stack.
                   7270: 
                   7271: You can also define complete definitions with these words; this provides
                   7272: an alternative to using @code{does>} (@pxref{User-defined Defining
                   7273: Words}).  E.g., instead of
                   7274: 
                   7275: @example
                   7276: : curry+ ( n1 "name" -- )
                   7277:     CREATE ,
                   7278: DOES> ( n2 -- n1+n2 )
                   7279:     @@ + ;
                   7280: @end example
                   7281: 
                   7282: you could define
                   7283: 
                   7284: @example
                   7285: : curry+ ( n1 "name" -- )
                   7286:   \ name execution: ( n2 -- n1+n2 )
                   7287:   >r : r> POSTPONE literal POSTPONE + POSTPONE ; ;
1.81      anton    7288: 
1.82      anton    7289: -3 curry+ 3-
                   7290: see 3-
                   7291: @end example
1.81      anton    7292: 
1.82      anton    7293: The sequence @code{>r : r>} is necessary, because @code{:} puts a
                   7294: colon-sys on the data stack that makes everything below it unaccessible.
1.81      anton    7295: 
1.82      anton    7296: This way of writing defining words is sometimes more, sometimes less
                   7297: convenient than using @code{does>} (@pxref{Advanced does> usage
                   7298: example}).  One advantage of this method is that it can be optimized
                   7299: better, because the compiler knows that the value compiled with
                   7300: @code{literal} is fixed, whereas the data associated with a
                   7301: @code{create}d word can be changed.
1.47      crook    7302: 
1.26      crook    7303: @c ----------------------------------------------------------
1.111     anton    7304: @node The Text Interpreter, The Input Stream, Compiling words, Words
1.26      crook    7305: @section  The Text Interpreter
                   7306: @cindex interpreter - outer
                   7307: @cindex text interpreter
                   7308: @cindex outer interpreter
1.1       anton    7309: 
1.34      anton    7310: @c Should we really describe all these ugly details?  IMO the text
                   7311: @c interpreter should be much cleaner, but that may not be possible within
                   7312: @c ANS Forth. - anton
1.44      crook    7313: @c nac-> I wanted to explain how it works to show how you can exploit
                   7314: @c it in your own programs. When I was writing a cross-compiler, figuring out
                   7315: @c some of these gory details was very helpful to me. None of the textbooks
                   7316: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
                   7317: @c seems to positively avoid going into too much detail for some of
                   7318: @c the internals.
1.34      anton    7319: 
1.71      anton    7320: @c anton: ok.  I wonder, though, if this is the right place; for some stuff
                   7321: @c it is; for the ugly details, I would prefer another place.  I wonder
                   7322: @c whether we should have a chapter before "Words" that describes some
                   7323: @c basic concepts referred to in words, and a chapter after "Words" that
                   7324: @c describes implementation details.
                   7325: 
1.29      crook    7326: The text interpreter@footnote{This is an expanded version of the
                   7327: material in @ref{Introducing the Text Interpreter}.} is an endless loop
1.34      anton    7328: that processes input from the current input device. It is also called
                   7329: the outer interpreter, in contrast to the inner interpreter
                   7330: (@pxref{Engine}) which executes the compiled Forth code on interpretive
                   7331: implementations.
1.27      crook    7332: 
1.29      crook    7333: @cindex interpret state
                   7334: @cindex compile state
                   7335: The text interpreter operates in one of two states: @dfn{interpret
                   7336: state} and @dfn{compile state}. The current state is defined by the
1.71      anton    7337: aptly-named variable @code{state}.
1.29      crook    7338: 
                   7339: This section starts by describing how the text interpreter behaves when
                   7340: it is in interpret state, processing input from the user input device --
                   7341: the keyboard. This is the mode that a Forth system is in after it starts
                   7342: up.
                   7343: 
                   7344: @cindex input buffer
                   7345: @cindex terminal input buffer
                   7346: The text interpreter works from an area of memory called the @dfn{input
                   7347: buffer}@footnote{When the text interpreter is processing input from the
                   7348: keyboard, this area of memory is called the @dfn{terminal input buffer}
                   7349: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
                   7350: @code{#TIB}.}, which stores your keyboard input when you press the
1.30      anton    7351: @key{RET} key. Starting at the beginning of the input buffer, it skips
1.29      crook    7352: leading spaces (called @dfn{delimiters}) then parses a string (a
                   7353: sequence of non-space characters) until it reaches either a space
                   7354: character or the end of the buffer. Having parsed a string, it makes two
                   7355: attempts to process it:
1.27      crook    7356: 
1.29      crook    7357: @cindex dictionary
1.27      crook    7358: @itemize @bullet
                   7359: @item
1.29      crook    7360: It looks for the string in a @dfn{dictionary} of definitions. If the
                   7361: string is found, the string names a @dfn{definition} (also known as a
                   7362: @dfn{word}) and the dictionary search returns information that allows
                   7363: the text interpreter to perform the word's @dfn{interpretation
                   7364: semantics}. In most cases, this simply means that the word will be
                   7365: executed.
1.27      crook    7366: @item
                   7367: If the string is not found in the dictionary, the text interpreter
1.29      crook    7368: attempts to treat it as a number, using the rules described in
                   7369: @ref{Number Conversion}. If the string represents a legal number in the
                   7370: current radix, the number is pushed onto a parameter stack (the data
                   7371: stack for integers, the floating-point stack for floating-point
                   7372: numbers).
                   7373: @end itemize
                   7374: 
                   7375: If both attempts fail, or if the word is found in the dictionary but has
                   7376: no interpretation semantics@footnote{This happens if the word was
                   7377: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
                   7378: remainder of the input buffer, issues an error message and waits for
                   7379: more input. If one of the attempts succeeds, the text interpreter
                   7380: repeats the parsing process until the whole of the input buffer has been
                   7381: processed, at which point it prints the status message ``@code{ ok}''
                   7382: and waits for more input.
                   7383: 
1.71      anton    7384: @c anton: this should be in the input stream subsection (or below it)
                   7385: 
1.29      crook    7386: @cindex parse area
                   7387: The text interpreter keeps track of its position in the input buffer by
                   7388: updating a variable called @code{>IN} (pronounced ``to-in''). The value
                   7389: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
                   7390: of the input buffer. The region from offset @code{>IN @@} to the end of
                   7391: the input buffer is called the @dfn{parse area}@footnote{In other words,
                   7392: the text interpreter processes the contents of the input buffer by
                   7393: parsing strings from the parse area until the parse area is empty.}.
                   7394: This example shows how @code{>IN} changes as the text interpreter parses
                   7395: the input buffer:
                   7396: 
                   7397: @example
                   7398: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
                   7399:   CR ." ->" TYPE ." <-" ; IMMEDIATE 
                   7400: 
                   7401: 1 2 3 remaining + remaining . 
                   7402: 
                   7403: : foo 1 2 3 remaining SWAP remaining ;
                   7404: @end example
                   7405: 
                   7406: @noindent
                   7407: The result is:
                   7408: 
                   7409: @example
                   7410: ->+ remaining .<-
                   7411: ->.<-5  ok
                   7412: 
                   7413: ->SWAP remaining ;-<
                   7414: ->;<-  ok
                   7415: @end example
                   7416: 
                   7417: @cindex parsing words
                   7418: The value of @code{>IN} can also be modified by a word in the input
                   7419: buffer that is executed by the text interpreter.  This means that a word
                   7420: can ``trick'' the text interpreter into either skipping a section of the
                   7421: input buffer@footnote{This is how parsing words work.} or into parsing a
                   7422: section twice. For example:
1.27      crook    7423: 
1.29      crook    7424: @example
1.71      anton    7425: : lat ." <<foo>>" ;
                   7426: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
1.29      crook    7427: @end example
                   7428: 
                   7429: @noindent
                   7430: When @code{flat} is executed, this output is produced@footnote{Exercise
                   7431: for the reader: what would happen if the @code{3} were replaced with
                   7432: @code{4}?}:
                   7433: 
                   7434: @example
1.71      anton    7435: <<bar>><<foo>>
1.29      crook    7436: @end example
                   7437: 
1.71      anton    7438: This technique can be used to work around some of the interoperability
                   7439: problems of parsing words.  Of course, it's better to avoid parsing
                   7440: words where possible.
                   7441: 
1.29      crook    7442: @noindent
                   7443: Two important notes about the behaviour of the text interpreter:
1.27      crook    7444: 
                   7445: @itemize @bullet
                   7446: @item
                   7447: It processes each input string to completion before parsing additional
1.29      crook    7448: characters from the input buffer.
                   7449: @item
                   7450: It treats the input buffer as a read-only region (and so must your code).
                   7451: @end itemize
                   7452: 
                   7453: @noindent
                   7454: When the text interpreter is in compile state, its behaviour changes in
                   7455: these ways:
                   7456: 
                   7457: @itemize @bullet
                   7458: @item
                   7459: If a parsed string is found in the dictionary, the text interpreter will
                   7460: perform the word's @dfn{compilation semantics}. In most cases, this
                   7461: simply means that the execution semantics of the word will be appended
                   7462: to the current definition.
1.27      crook    7463: @item
1.29      crook    7464: When a number is encountered, it is compiled into the current definition
                   7465: (as a literal) rather than being pushed onto a parameter stack.
                   7466: @item
                   7467: If an error occurs, @code{state} is modified to put the text interpreter
                   7468: back into interpret state.
                   7469: @item
                   7470: Each time a line is entered from the keyboard, Gforth prints
                   7471: ``@code{ compiled}'' rather than `` @code{ok}''.
                   7472: @end itemize
                   7473: 
                   7474: @cindex text interpreter - input sources
                   7475: When the text interpreter is using an input device other than the
                   7476: keyboard, its behaviour changes in these ways:
                   7477: 
                   7478: @itemize @bullet
                   7479: @item
                   7480: When the parse area is empty, the text interpreter attempts to refill
                   7481: the input buffer from the input source. When the input source is
1.71      anton    7482: exhausted, the input source is set back to the previous input source.
1.29      crook    7483: @item
                   7484: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
                   7485: time the parse area is emptied.
                   7486: @item
                   7487: If an error occurs, the input source is set back to the user input
                   7488: device.
1.27      crook    7489: @end itemize
1.21      crook    7490: 
1.49      anton    7491: You can read about this in more detail in @ref{Input Sources}.
1.44      crook    7492: 
1.26      crook    7493: doc->in
1.27      crook    7494: doc-source
                   7495: 
1.26      crook    7496: doc-tib
                   7497: doc-#tib
1.1       anton    7498: 
1.44      crook    7499: 
1.26      crook    7500: @menu
1.67      anton    7501: * Input Sources::               
                   7502: * Number Conversion::           
                   7503: * Interpret/Compile states::    
                   7504: * Interpreter Directives::      
1.26      crook    7505: @end menu
1.1       anton    7506: 
1.29      crook    7507: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
                   7508: @subsection Input Sources
                   7509: @cindex input sources
                   7510: @cindex text interpreter - input sources
                   7511: 
1.44      crook    7512: By default, the text interpreter processes input from the user input
1.29      crook    7513: device (the keyboard) when Forth starts up. The text interpreter can
                   7514: process input from any of these sources:
                   7515: 
                   7516: @itemize @bullet
                   7517: @item
                   7518: The user input device -- the keyboard.
                   7519: @item
                   7520: A file, using the words described in @ref{Forth source files}.
                   7521: @item
                   7522: A block, using the words described in @ref{Blocks}.
                   7523: @item
                   7524: A text string, using @code{evaluate}.
                   7525: @end itemize
                   7526: 
                   7527: A program can identify the current input device from the values of
                   7528: @code{source-id} and @code{blk}.
                   7529: 
1.44      crook    7530: 
1.29      crook    7531: doc-source-id
                   7532: doc-blk
                   7533: 
                   7534: doc-save-input
                   7535: doc-restore-input
                   7536: 
                   7537: doc-evaluate
1.111     anton    7538: doc-query
1.1       anton    7539: 
1.29      crook    7540: 
1.44      crook    7541: 
1.29      crook    7542: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
1.26      crook    7543: @subsection Number Conversion
                   7544: @cindex number conversion
                   7545: @cindex double-cell numbers, input format
                   7546: @cindex input format for double-cell numbers
                   7547: @cindex single-cell numbers, input format
                   7548: @cindex input format for single-cell numbers
                   7549: @cindex floating-point numbers, input format
                   7550: @cindex input format for floating-point numbers
1.1       anton    7551: 
1.29      crook    7552: This section describes the rules that the text interpreter uses when it
                   7553: tries to convert a string into a number.
1.1       anton    7554: 
1.26      crook    7555: Let <digit> represent any character that is a legal digit in the current
1.29      crook    7556: number base@footnote{For example, 0-9 when the number base is decimal or
                   7557: 0-9, A-F when the number base is hexadecimal.}.
1.1       anton    7558: 
1.26      crook    7559: Let <decimal digit> represent any character in the range 0-9.
1.1       anton    7560: 
1.29      crook    7561: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
                   7562: in the braces (@i{a} or @i{b} or neither).
1.1       anton    7563: 
1.26      crook    7564: Let * represent any number of instances of the previous character
                   7565: (including none).
1.1       anton    7566: 
1.26      crook    7567: Let any other character represent itself.
1.1       anton    7568: 
1.29      crook    7569: @noindent
1.26      crook    7570: Now, the conversion rules are:
1.21      crook    7571: 
1.26      crook    7572: @itemize @bullet
                   7573: @item
                   7574: A string of the form <digit><digit>* is treated as a single-precision
1.29      crook    7575: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
1.26      crook    7576: @item
                   7577: A string of the form -<digit><digit>* is treated as a single-precision
1.29      crook    7578: (cell-sized) negative integer, and is represented using 2's-complement
1.26      crook    7579: arithmetic. Examples are -45 -5681 -0
                   7580: @item
                   7581: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
1.29      crook    7582: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
                   7583: (all three of these represent the same number).
1.26      crook    7584: @item
                   7585: A string of the form -<digit><digit>*.<digit>* is treated as a
1.29      crook    7586: double-precision (double-cell-sized) negative integer, and is
1.26      crook    7587: represented using 2's-complement arithmetic. Examples are -3465. -3.465
1.29      crook    7588: -34.65 (all three of these represent the same number).
1.26      crook    7589: @item
1.29      crook    7590: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
                   7591: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
1.35      anton    7592: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
1.29      crook    7593: number) +12.E-4
1.26      crook    7594: @end itemize
1.1       anton    7595: 
1.26      crook    7596: By default, the number base used for integer number conversion is given
1.35      anton    7597: by the contents of the variable @code{base}.  Note that a lot of
                   7598: confusion can result from unexpected values of @code{base}.  If you
                   7599: change @code{base} anywhere, make sure to save the old value and restore
                   7600: it afterwards.  In general I recommend keeping @code{base} decimal, and
                   7601: using the prefixes described below for the popular non-decimal bases.
1.1       anton    7602: 
1.29      crook    7603: doc-dpl
1.26      crook    7604: doc-base
                   7605: doc-hex
                   7606: doc-decimal
1.1       anton    7607: 
1.26      crook    7608: @cindex '-prefix for character strings
                   7609: @cindex &-prefix for decimal numbers
1.133     anton    7610: @cindex #-prefix for decimal numbers
1.26      crook    7611: @cindex %-prefix for binary numbers
                   7612: @cindex $-prefix for hexadecimal numbers
1.133     anton    7613: @cindex 0x-prefix for hexadecimal numbers
1.35      anton    7614: Gforth allows you to override the value of @code{base} by using a
1.29      crook    7615: prefix@footnote{Some Forth implementations provide a similar scheme by
                   7616: implementing @code{$} etc. as parsing words that process the subsequent
                   7617: number in the input stream and push it onto the stack. For example, see
                   7618: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
                   7619: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
                   7620: is required between the prefix and the number.} before the first digit
1.133     anton    7621: of an (integer) number. The following prefixes are supported:
1.1       anton    7622: 
1.26      crook    7623: @itemize @bullet
                   7624: @item
1.35      anton    7625: @code{&} -- decimal
1.26      crook    7626: @item
1.133     anton    7627: @code{#} -- decimal
                   7628: @item
1.35      anton    7629: @code{%} -- binary
1.26      crook    7630: @item
1.35      anton    7631: @code{$} -- hexadecimal
1.26      crook    7632: @item
1.133     anton    7633: @code{0x} -- hexadecimal, if base<33.
                   7634: @item
                   7635: @code{'} -- numeric value (e.g., ASCII code) of next character; an
                   7636: optional @code{'} may be present after the character.
1.26      crook    7637: @end itemize
1.1       anton    7638: 
1.26      crook    7639: Here are some examples, with the equivalent decimal number shown after
                   7640: in braces:
1.1       anton    7641: 
1.26      crook    7642: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
1.133     anton    7643: 'A (65),
                   7644: -'a' (-97),
1.26      crook    7645: &905 (905), $abc (2478), $ABC (2478).
1.1       anton    7646: 
1.26      crook    7647: @cindex number conversion - traps for the unwary
1.29      crook    7648: @noindent
1.26      crook    7649: Number conversion has a number of traps for the unwary:
1.1       anton    7650: 
1.26      crook    7651: @itemize @bullet
                   7652: @item
                   7653: You cannot determine the current number base using the code sequence
1.35      anton    7654: @code{base @@ .} -- the number base is always 10 in the current number
                   7655: base. Instead, use something like @code{base @@ dec.}
1.26      crook    7656: @item
                   7657: If the number base is set to a value greater than 14 (for example,
                   7658: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
                   7659: it to be intepreted as either a single-precision integer or a
                   7660: floating-point number (Gforth treats it as an integer). The ambiguity
                   7661: can be resolved by explicitly stating the sign of the mantissa and/or
                   7662: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
                   7663: ambiguity arises; either representation will be treated as a
                   7664: floating-point number.
                   7665: @item
1.29      crook    7666: There is a word @code{bin} but it does @i{not} set the number base!
1.26      crook    7667: It is used to specify file types.
                   7668: @item
1.72      anton    7669: ANS Forth requires the @code{.} of a double-precision number to be the
                   7670: final character in the string.  Gforth allows the @code{.} to be
                   7671: anywhere after the first digit.
1.26      crook    7672: @item
                   7673: The number conversion process does not check for overflow.
                   7674: @item
1.72      anton    7675: In an ANS Forth program @code{base} is required to be decimal when
                   7676: converting floating-point numbers.  In Gforth, number conversion to
                   7677: floating-point numbers always uses base &10, irrespective of the value
                   7678: of @code{base}.
1.26      crook    7679: @end itemize
1.1       anton    7680: 
1.49      anton    7681: You can read numbers into your programs with the words described in
                   7682: @ref{Input}.
1.1       anton    7683: 
1.82      anton    7684: @node Interpret/Compile states, Interpreter Directives, Number Conversion, The Text Interpreter
1.26      crook    7685: @subsection Interpret/Compile states
                   7686: @cindex Interpret/Compile states
1.1       anton    7687: 
1.29      crook    7688: A standard program is not permitted to change @code{state}
                   7689: explicitly. However, it can change @code{state} implicitly, using the
                   7690: words @code{[} and @code{]}. When @code{[} is executed it switches
                   7691: @code{state} to interpret state, and therefore the text interpreter
                   7692: starts interpreting. When @code{]} is executed it switches @code{state}
                   7693: to compile state and therefore the text interpreter starts
1.44      crook    7694: compiling. The most common usage for these words is for switching into
                   7695: interpret state and back from within a colon definition; this technique
1.49      anton    7696: can be used to compile a literal (for an example, @pxref{Literals}) or
                   7697: for conditional compilation (for an example, @pxref{Interpreter
                   7698: Directives}).
1.44      crook    7699: 
1.35      anton    7700: 
                   7701: @c This is a bad example: It's non-standard, and it's not necessary.
                   7702: @c However, I can't think of a good example for switching into compile
                   7703: @c state when there is no current word (@code{state}-smart words are not a
                   7704: @c good reason).  So maybe we should use an example for switching into
                   7705: @c interpret @code{state} in a colon def. - anton
1.44      crook    7706: @c nac-> I agree. I started out by putting in the example, then realised
                   7707: @c that it was non-ANS, so wrote more words around it. I hope this
                   7708: @c re-written version is acceptable to you. I do want to keep the example
                   7709: @c as it is helpful for showing what is and what is not portable, particularly
                   7710: @c where it outlaws a style in common use.
                   7711: 
1.72      anton    7712: @c anton: it's more important to show what's portable.  After we have done
1.83      anton    7713: @c that, we can also show what's not.  In any case, I have written a
                   7714: @c section Compiling Words which also deals with [ ].
1.35      anton    7715: 
1.95      anton    7716: @c  !! The following example does not work in Gforth 0.5.9 or later.
1.29      crook    7717: 
1.95      anton    7718: @c  @code{[} and @code{]} also give you the ability to switch into compile
                   7719: @c  state and back, but we cannot think of any useful Standard application
                   7720: @c  for this ability. Pre-ANS Forth textbooks have examples like this:
                   7721: 
                   7722: @c  @example
                   7723: @c  : AA ." this is A" ;
                   7724: @c  : BB ." this is B" ;
                   7725: @c  : CC ." this is C" ;
                   7726: 
                   7727: @c  create table ] aa bb cc [
                   7728: 
                   7729: @c  : go ( n -- ) \ n is offset into table.. 0 for 1st entry
                   7730: @c    cells table + @@ execute ;
                   7731: @c  @end example
                   7732: 
                   7733: @c  This example builds a jump table; @code{0 go} will display ``@code{this
                   7734: @c  is A}''. Using @code{[} and @code{]} in this example is equivalent to
                   7735: @c  defining @code{table} like this:
                   7736: 
                   7737: @c  @example
                   7738: @c  create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
                   7739: @c  @end example
                   7740: 
                   7741: @c  The problem with this code is that the definition of @code{table} is not
                   7742: @c  portable -- it @i{compile}s execution tokens into code space. Whilst it
                   7743: @c  @i{may} work on systems where code space and data space co-incide, the
                   7744: @c  Standard only allows data space to be assigned for a @code{CREATE}d
                   7745: @c  word. In addition, the Standard only allows @code{@@} to access data
                   7746: @c  space, whilst this example is using it to access code space. The only
                   7747: @c  portable, Standard way to build this table is to build it in data space,
                   7748: @c  like this:
                   7749: 
                   7750: @c  @example
                   7751: @c  create table ' aa , ' bb , ' cc ,
                   7752: @c  @end example
1.29      crook    7753: 
1.95      anton    7754: @c  doc-state
1.44      crook    7755: 
1.29      crook    7756: 
1.82      anton    7757: @node Interpreter Directives,  , Interpret/Compile states, The Text Interpreter
1.26      crook    7758: @subsection Interpreter Directives
                   7759: @cindex interpreter directives
1.72      anton    7760: @cindex conditional compilation
1.1       anton    7761: 
1.29      crook    7762: These words are usually used in interpret state; typically to control
                   7763: which parts of a source file are processed by the text
1.26      crook    7764: interpreter. There are only a few ANS Forth Standard words, but Gforth
                   7765: supplements these with a rich set of immediate control structure words
                   7766: to compensate for the fact that the non-immediate versions can only be
1.29      crook    7767: used in compile state (@pxref{Control Structures}). Typical usages:
                   7768: 
                   7769: @example
1.72      anton    7770: FALSE Constant HAVE-ASSEMBLER
1.29      crook    7771: .
                   7772: .
1.72      anton    7773: HAVE-ASSEMBLER [IF]
1.29      crook    7774: : ASSEMBLER-FEATURE
                   7775:   ...
                   7776: ;
                   7777: [ENDIF]
                   7778: .
                   7779: .
                   7780: : SEE
                   7781:   ... \ general-purpose SEE code
1.72      anton    7782:   [ HAVE-ASSEMBLER [IF] ]
1.29      crook    7783:   ... \ assembler-specific SEE code
                   7784:   [ [ENDIF] ]
                   7785: ;
                   7786: @end example
1.1       anton    7787: 
1.44      crook    7788: 
1.26      crook    7789: doc-[IF]
                   7790: doc-[ELSE]
                   7791: doc-[THEN]
                   7792: doc-[ENDIF]
1.1       anton    7793: 
1.26      crook    7794: doc-[IFDEF]
                   7795: doc-[IFUNDEF]
1.1       anton    7796: 
1.26      crook    7797: doc-[?DO]
                   7798: doc-[DO]
                   7799: doc-[FOR]
                   7800: doc-[LOOP]
                   7801: doc-[+LOOP]
                   7802: doc-[NEXT]
1.1       anton    7803: 
1.26      crook    7804: doc-[BEGIN]
                   7805: doc-[UNTIL]
                   7806: doc-[AGAIN]
                   7807: doc-[WHILE]
                   7808: doc-[REPEAT]
1.1       anton    7809: 
1.27      crook    7810: 
1.26      crook    7811: @c -------------------------------------------------------------
1.111     anton    7812: @node The Input Stream, Word Lists, The Text Interpreter, Words
                   7813: @section The Input Stream
                   7814: @cindex input stream
                   7815: 
                   7816: @c !! integrate this better with the "Text Interpreter" section
                   7817: The text interpreter reads from the input stream, which can come from
                   7818: several sources (@pxref{Input Sources}).  Some words, in particular
                   7819: defining words, but also words like @code{'}, read parameters from the
                   7820: input stream instead of from the stack.
                   7821: 
                   7822: Such words are called parsing words, because they parse the input
                   7823: stream.  Parsing words are hard to use in other words, because it is
                   7824: hard to pass program-generated parameters through the input stream.
                   7825: They also usually have an unintuitive combination of interpretation and
                   7826: compilation semantics when implemented naively, leading to various
                   7827: approaches that try to produce a more intuitive behaviour
                   7828: (@pxref{Combined words}).
                   7829: 
                   7830: It should be obvious by now that parsing words are a bad idea.  If you
                   7831: want to implement a parsing word for convenience, also provide a factor
                   7832: of the word that does not parse, but takes the parameters on the stack.
                   7833: To implement the parsing word on top if it, you can use the following
                   7834: words:
                   7835: 
                   7836: @c anton: these belong in the input stream section
                   7837: doc-parse
1.138     anton    7838: doc-parse-name
1.111     anton    7839: doc-parse-word
                   7840: doc-name
                   7841: doc-word
                   7842: doc-\"-parse
                   7843: doc-refill
                   7844: 
                   7845: Conversely, if you have the bad luck (or lack of foresight) to have to
                   7846: deal with parsing words without having such factors, how do you pass a
                   7847: string that is not in the input stream to it?
                   7848: 
                   7849: doc-execute-parsing
                   7850: 
                   7851: If you want to run a parsing word on a file, the following word should
                   7852: help:
                   7853: 
                   7854: doc-execute-parsing-file
                   7855: 
                   7856: @c -------------------------------------------------------------
                   7857: @node Word Lists, Environmental Queries, The Input Stream, Words
1.26      crook    7858: @section Word Lists
                   7859: @cindex word lists
1.32      anton    7860: @cindex header space
1.1       anton    7861: 
1.36      anton    7862: A wordlist is a list of named words; you can add new words and look up
                   7863: words by name (and you can remove words in a restricted way with
                   7864: markers).  Every named (and @code{reveal}ed) word is in one wordlist.
                   7865: 
                   7866: @cindex search order stack
                   7867: The text interpreter searches the wordlists present in the search order
                   7868: (a stack of wordlists), from the top to the bottom.  Within each
                   7869: wordlist, the search starts conceptually at the newest word; i.e., if
                   7870: two words in a wordlist have the same name, the newer word is found.
1.1       anton    7871: 
1.26      crook    7872: @cindex compilation word list
1.36      anton    7873: New words are added to the @dfn{compilation wordlist} (aka current
                   7874: wordlist).
1.1       anton    7875: 
1.36      anton    7876: @cindex wid
                   7877: A word list is identified by a cell-sized word list identifier (@i{wid})
                   7878: in much the same way as a file is identified by a file handle. The
                   7879: numerical value of the wid has no (portable) meaning, and might change
                   7880: from session to session.
1.1       anton    7881: 
1.29      crook    7882: The ANS Forth ``Search order'' word set is intended to provide a set of
                   7883: low-level tools that allow various different schemes to be
1.74      anton    7884: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
1.26      crook    7885: word.  @file{compat/vocabulary.fs} provides an implementation in ANS
1.45      crook    7886: Forth.
1.1       anton    7887: 
1.27      crook    7888: @comment TODO: locals section refers to here, saying that every word list (aka
                   7889: @comment vocabulary) has its own methods for searching etc. Need to document that.
1.78      anton    7890: @c anton: but better in a separate subsection on wordlist internals
1.1       anton    7891: 
1.45      crook    7892: @comment TODO: document markers, reveal, tables, mappedwordlist
                   7893: 
                   7894: @comment the gforthman- prefix is used to pick out the true definition of a
1.27      crook    7895: @comment word from the source files, rather than some alias.
1.44      crook    7896: 
1.26      crook    7897: doc-forth-wordlist
                   7898: doc-definitions
                   7899: doc-get-current
                   7900: doc-set-current
                   7901: doc-get-order
1.45      crook    7902: doc---gforthman-set-order
1.26      crook    7903: doc-wordlist
1.30      anton    7904: doc-table
1.79      anton    7905: doc->order
1.36      anton    7906: doc-previous
1.26      crook    7907: doc-also
1.45      crook    7908: doc---gforthman-forth
1.26      crook    7909: doc-only
1.45      crook    7910: doc---gforthman-order
1.15      anton    7911: 
1.26      crook    7912: doc-find
                   7913: doc-search-wordlist
1.15      anton    7914: 
1.26      crook    7915: doc-words
                   7916: doc-vlist
1.44      crook    7917: @c doc-words-deferred
1.1       anton    7918: 
1.74      anton    7919: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
1.26      crook    7920: doc-root
                   7921: doc-vocabulary
                   7922: doc-seal
                   7923: doc-vocs
                   7924: doc-current
                   7925: doc-context
1.1       anton    7926: 
1.44      crook    7927: 
1.26      crook    7928: @menu
1.75      anton    7929: * Vocabularies::                
1.67      anton    7930: * Why use word lists?::         
1.75      anton    7931: * Word list example::           
1.26      crook    7932: @end menu
                   7933: 
1.75      anton    7934: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
                   7935: @subsection Vocabularies
                   7936: @cindex Vocabularies, detailed explanation
                   7937: 
                   7938: Here is an example of creating and using a new wordlist using ANS
                   7939: Forth words:
                   7940: 
                   7941: @example
                   7942: wordlist constant my-new-words-wordlist
                   7943: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
                   7944: 
                   7945: \ add it to the search order
                   7946: also my-new-words
                   7947: 
                   7948: \ alternatively, add it to the search order and make it
                   7949: \ the compilation word list
                   7950: also my-new-words definitions
                   7951: \ type "order" to see the problem
                   7952: @end example
                   7953: 
                   7954: The problem with this example is that @code{order} has no way to
                   7955: associate the name @code{my-new-words} with the wid of the word list (in
                   7956: Gforth, @code{order} and @code{vocs} will display @code{???}  for a wid
                   7957: that has no associated name). There is no Standard way of associating a
                   7958: name with a wid.
                   7959: 
                   7960: In Gforth, this example can be re-coded using @code{vocabulary}, which
                   7961: associates a name with a wid:
                   7962: 
                   7963: @example
                   7964: vocabulary my-new-words
                   7965: 
                   7966: \ add it to the search order
                   7967: also my-new-words
                   7968: 
                   7969: \ alternatively, add it to the search order and make it
                   7970: \ the compilation word list
                   7971: my-new-words definitions
                   7972: \ type "order" to see that the problem is solved
                   7973: @end example
                   7974: 
                   7975: 
                   7976: @node Why use word lists?, Word list example, Vocabularies, Word Lists
1.26      crook    7977: @subsection Why use word lists?
                   7978: @cindex word lists - why use them?
                   7979: 
1.74      anton    7980: Here are some reasons why people use wordlists:
1.26      crook    7981: 
                   7982: @itemize @bullet
1.74      anton    7983: 
                   7984: @c anton: Gforth's hashing implementation makes the search speed
                   7985: @c independent from the number of words.  But it is linear with the number
                   7986: @c of wordlists that have to be searched, so in effect using more wordlists
                   7987: @c actually slows down compilation.
                   7988: 
                   7989: @c @item
                   7990: @c To improve compilation speed by reducing the number of header space
                   7991: @c entries that must be searched. This is achieved by creating a new
                   7992: @c word list that contains all of the definitions that are used in the
                   7993: @c definition of a Forth system but which would not usually be used by
                   7994: @c programs running on that system. That word list would be on the search
                   7995: @c list when the Forth system was compiled but would be removed from the
                   7996: @c search list for normal operation. This can be a useful technique for
                   7997: @c low-performance systems (for example, 8-bit processors in embedded
                   7998: @c systems) but is unlikely to be necessary in high-performance desktop
                   7999: @c systems.
                   8000: 
1.26      crook    8001: @item
                   8002: To prevent a set of words from being used outside the context in which
                   8003: they are valid. Two classic examples of this are an integrated editor
                   8004: (all of the edit commands are defined in a separate word list; the
                   8005: search order is set to the editor word list when the editor is invoked;
                   8006: the old search order is restored when the editor is terminated) and an
                   8007: integrated assembler (the op-codes for the machine are defined in a
                   8008: separate word list which is used when a @code{CODE} word is defined).
1.74      anton    8009: 
                   8010: @item
                   8011: To organize the words of an application or library into a user-visible
                   8012: set (in @code{forth-wordlist} or some other common wordlist) and a set
                   8013: of helper words used just for the implementation (hidden in a separate
1.75      anton    8014: wordlist).  This keeps @code{words}' output smaller, separates
                   8015: implementation and interface, and reduces the chance of name conflicts
                   8016: within the common wordlist.
1.74      anton    8017: 
1.26      crook    8018: @item
                   8019: To prevent a name-space clash between multiple definitions with the same
                   8020: name. For example, when building a cross-compiler you might have a word
                   8021: @code{IF} that generates conditional code for your target system. By
                   8022: placing this definition in a different word list you can control whether
                   8023: the host system's @code{IF} or the target system's @code{IF} get used in
                   8024: any particular context by controlling the order of the word lists on the
                   8025: search order stack.
1.74      anton    8026: 
1.26      crook    8027: @end itemize
1.1       anton    8028: 
1.74      anton    8029: The downsides of using wordlists are:
                   8030: 
                   8031: @itemize
                   8032: 
                   8033: @item
                   8034: Debugging becomes more cumbersome.
                   8035: 
                   8036: @item
                   8037: Name conflicts worked around with wordlists are still there, and you
                   8038: have to arrange the search order carefully to get the desired results;
                   8039: if you forget to do that, you get hard-to-find errors (as in any case
                   8040: where you read the code differently from the compiler; @code{see} can
1.75      anton    8041: help seeing which of several possible words the name resolves to in such
                   8042: cases).  @code{See} displays just the name of the words, not what
                   8043: wordlist they belong to, so it might be misleading.  Using unique names
                   8044: is a better approach to avoid name conflicts.
1.74      anton    8045: 
                   8046: @item
                   8047: You have to explicitly undo any changes to the search order.  In many
                   8048: cases it would be more convenient if this happened implicitly.  Gforth
                   8049: currently does not provide such a feature, but it may do so in the
                   8050: future.
                   8051: @end itemize
                   8052: 
                   8053: 
1.75      anton    8054: @node Word list example,  , Why use word lists?, Word Lists
                   8055: @subsection Word list example
                   8056: @cindex word lists - example
1.1       anton    8057: 
1.74      anton    8058: The following example is from the
                   8059: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
                   8060: garbage collector} and uses wordlists to separate public words from
                   8061: helper words:
                   8062: 
                   8063: @example
                   8064: get-current ( wid )
                   8065: vocabulary garbage-collector also garbage-collector definitions
                   8066: ... \ define helper words
                   8067: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
                   8068: ... \ define the public (i.e., API) words
                   8069:     \ they can refer to the helper words
                   8070: previous \ restore original search order (helper words become invisible)
                   8071: @end example
                   8072: 
1.26      crook    8073: @c -------------------------------------------------------------
                   8074: @node Environmental Queries, Files, Word Lists, Words
                   8075: @section Environmental Queries
                   8076: @cindex environmental queries
1.21      crook    8077: 
1.26      crook    8078: ANS Forth introduced the idea of ``environmental queries'' as a way
                   8079: for a program running on a system to determine certain characteristics of the system.
                   8080: The Standard specifies a number of strings that might be recognised by a system.
1.21      crook    8081: 
1.32      anton    8082: The Standard requires that the header space used for environmental queries
                   8083: be distinct from the header space used for definitions.
1.21      crook    8084: 
1.26      crook    8085: Typically, environmental queries are supported by creating a set of
1.29      crook    8086: definitions in a word list that is @i{only} used during environmental
1.26      crook    8087: queries; that is what Gforth does. There is no Standard way of adding
                   8088: definitions to the set of recognised environmental queries, but any
                   8089: implementation that supports the loading of optional word sets must have
                   8090: some mechanism for doing this (after loading the word set, the
                   8091: associated environmental query string must return @code{true}). In
                   8092: Gforth, the word list used to honour environmental queries can be
                   8093: manipulated just like any other word list.
1.21      crook    8094: 
1.44      crook    8095: 
1.26      crook    8096: doc-environment?
                   8097: doc-environment-wordlist
1.21      crook    8098: 
1.26      crook    8099: doc-gforth
                   8100: doc-os-class
1.21      crook    8101: 
1.44      crook    8102: 
1.26      crook    8103: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
                   8104: returning two items on the stack, querying it using @code{environment?}
                   8105: will return an additional item; the @code{true} flag that shows that the
                   8106: string was recognised.
1.21      crook    8107: 
1.26      crook    8108: @comment TODO Document the standard strings or note where they are documented herein
1.21      crook    8109: 
1.26      crook    8110: Here are some examples of using environmental queries:
1.21      crook    8111: 
1.26      crook    8112: @example
                   8113: s" address-unit-bits" environment? 0=
                   8114: [IF]
                   8115:      cr .( environmental attribute address-units-bits unknown... ) cr
1.75      anton    8116: [ELSE]
                   8117:      drop \ ensure balanced stack effect
1.26      crook    8118: [THEN]
1.21      crook    8119: 
1.75      anton    8120: \ this might occur in the prelude of a standard program that uses THROW
                   8121: s" exception" environment? [IF]
                   8122:    0= [IF]
                   8123:       : throw abort" exception thrown" ;
                   8124:    [THEN]
                   8125: [ELSE] \ we don't know, so make sure
                   8126:    : throw abort" exception thrown" ;
                   8127: [THEN]
1.21      crook    8128: 
1.26      crook    8129: s" gforth" environment? [IF] .( Gforth version ) TYPE
                   8130:                         [ELSE] .( Not Gforth..) [THEN]
1.75      anton    8131: 
                   8132: \ a program using v*
                   8133: s" gforth" environment? [IF]
                   8134:   s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
                   8135:    : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
                   8136:      >r swap 2swap swap 0e r> 0 ?DO
                   8137:        dup f@ over + 2swap dup f@ f* f+ over + 2swap
                   8138:      LOOP
                   8139:      2drop 2drop ; 
                   8140:   [THEN]
                   8141: [ELSE] \ 
                   8142:   : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
                   8143:   ...
                   8144: [THEN]
1.26      crook    8145: @end example
1.21      crook    8146: 
1.26      crook    8147: Here is an example of adding a definition to the environment word list:
1.21      crook    8148: 
1.26      crook    8149: @example
                   8150: get-current environment-wordlist set-current
                   8151: true constant block
                   8152: true constant block-ext
                   8153: set-current
                   8154: @end example
1.21      crook    8155: 
1.26      crook    8156: You can see what definitions are in the environment word list like this:
1.21      crook    8157: 
1.26      crook    8158: @example
1.79      anton    8159: environment-wordlist >order words previous
1.26      crook    8160: @end example
1.21      crook    8161: 
                   8162: 
1.26      crook    8163: @c -------------------------------------------------------------
                   8164: @node Files, Blocks, Environmental Queries, Words
                   8165: @section Files
1.28      crook    8166: @cindex files
                   8167: @cindex I/O - file-handling
1.21      crook    8168: 
1.26      crook    8169: Gforth provides facilities for accessing files that are stored in the
                   8170: host operating system's file-system. Files that are processed by Gforth
                   8171: can be divided into two categories:
1.21      crook    8172: 
1.23      crook    8173: @itemize @bullet
                   8174: @item
1.29      crook    8175: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
1.23      crook    8176: @item
1.29      crook    8177: Files that are processed by some other program (@dfn{general files}).
1.26      crook    8178: @end itemize
                   8179: 
                   8180: @menu
1.48      anton    8181: * Forth source files::          
                   8182: * General files::               
                   8183: * Search Paths::                
1.26      crook    8184: @end menu
                   8185: 
                   8186: @c -------------------------------------------------------------
                   8187: @node Forth source files, General files, Files, Files
                   8188: @subsection Forth source files
                   8189: @cindex including files
                   8190: @cindex Forth source files
1.21      crook    8191: 
1.26      crook    8192: The simplest way to interpret the contents of a file is to use one of
                   8193: these two formats:
1.21      crook    8194: 
1.26      crook    8195: @example
                   8196: include mysource.fs
                   8197: s" mysource.fs" included
                   8198: @end example
1.21      crook    8199: 
1.75      anton    8200: You usually want to include a file only if it is not included already
1.26      crook    8201: (by, say, another source file). In that case, you can use one of these
1.45      crook    8202: three formats:
1.21      crook    8203: 
1.26      crook    8204: @example
                   8205: require mysource.fs
                   8206: needs mysource.fs
                   8207: s" mysource.fs" required
                   8208: @end example
1.21      crook    8209: 
1.26      crook    8210: @cindex stack effect of included files
                   8211: @cindex including files, stack effect
1.45      crook    8212: It is good practice to write your source files such that interpreting them
                   8213: does not change the stack. Source files designed in this way can be used with
1.26      crook    8214: @code{required} and friends without complications. For example:
1.21      crook    8215: 
1.26      crook    8216: @example
1.75      anton    8217: 1024 require foo.fs drop
1.26      crook    8218: @end example
1.21      crook    8219: 
1.75      anton    8220: Here you want to pass the argument 1024 (e.g., a buffer size) to
                   8221: @file{foo.fs}.  Interpreting @file{foo.fs} has the stack effect ( n -- n
                   8222: ), which allows its use with @code{require}.  Of course with such
                   8223: parameters to required files, you have to ensure that the first
                   8224: @code{require} fits for all uses (i.e., @code{require} it early in the
                   8225: master load file).
1.44      crook    8226: 
1.26      crook    8227: doc-include-file
                   8228: doc-included
1.28      crook    8229: doc-included?
1.26      crook    8230: doc-include
                   8231: doc-required
                   8232: doc-require
                   8233: doc-needs
1.75      anton    8234: @c doc-init-included-files @c internal
                   8235: doc-sourcefilename
                   8236: doc-sourceline#
1.44      crook    8237: 
1.26      crook    8238: A definition in ANS Forth for @code{required} is provided in
                   8239: @file{compat/required.fs}.
1.21      crook    8240: 
1.26      crook    8241: @c -------------------------------------------------------------
                   8242: @node General files, Search Paths, Forth source files, Files
                   8243: @subsection General files
                   8244: @cindex general files
                   8245: @cindex file-handling
1.21      crook    8246: 
1.75      anton    8247: Files are opened/created by name and type. The following file access
                   8248: methods (FAMs) are recognised:
1.44      crook    8249: 
1.75      anton    8250: @cindex fam (file access method)
1.26      crook    8251: doc-r/o
                   8252: doc-r/w
                   8253: doc-w/o
                   8254: doc-bin
1.1       anton    8255: 
1.44      crook    8256: 
1.26      crook    8257: When a file is opened/created, it returns a file identifier,
1.29      crook    8258: @i{wfileid} that is used for all other file commands. All file
                   8259: commands also return a status value, @i{wior}, that is 0 for a
1.26      crook    8260: successful operation and an implementation-defined non-zero value in the
                   8261: case of an error.
1.21      crook    8262: 
1.44      crook    8263: 
1.26      crook    8264: doc-open-file
                   8265: doc-create-file
1.21      crook    8266: 
1.26      crook    8267: doc-close-file
                   8268: doc-delete-file
                   8269: doc-rename-file
                   8270: doc-read-file
                   8271: doc-read-line
                   8272: doc-write-file
                   8273: doc-write-line
                   8274: doc-emit-file
                   8275: doc-flush-file
1.21      crook    8276: 
1.26      crook    8277: doc-file-status
                   8278: doc-file-position
                   8279: doc-reposition-file
                   8280: doc-file-size
                   8281: doc-resize-file
1.21      crook    8282: 
1.93      anton    8283: doc-slurp-file
                   8284: doc-slurp-fid
1.112     anton    8285: doc-stdin
                   8286: doc-stdout
                   8287: doc-stderr
1.44      crook    8288: 
1.26      crook    8289: @c ---------------------------------------------------------
1.48      anton    8290: @node Search Paths,  , General files, Files
1.26      crook    8291: @subsection Search Paths
                   8292: @cindex path for @code{included}
                   8293: @cindex file search path
                   8294: @cindex @code{include} search path
                   8295: @cindex search path for files
1.21      crook    8296: 
1.26      crook    8297: If you specify an absolute filename (i.e., a filename starting with
                   8298: @file{/} or @file{~}, or with @file{:} in the second position (as in
                   8299: @samp{C:...})) for @code{included} and friends, that file is included
                   8300: just as you would expect.
1.21      crook    8301: 
1.75      anton    8302: If the filename starts with @file{./}, this refers to the directory that
                   8303: the present file was @code{included} from.  This allows files to include
                   8304: other files relative to their own position (irrespective of the current
                   8305: working directory or the absolute position).  This feature is essential
                   8306: for libraries consisting of several files, where a file may include
                   8307: other files from the library.  It corresponds to @code{#include "..."}
                   8308: in C. If the current input source is not a file, @file{.} refers to the
                   8309: directory of the innermost file being included, or, if there is no file
                   8310: being included, to the current working directory.
                   8311: 
                   8312: For relative filenames (not starting with @file{./}), Gforth uses a
                   8313: search path similar to Forth's search order (@pxref{Word Lists}). It
                   8314: tries to find the given filename in the directories present in the path,
                   8315: and includes the first one it finds. There are separate search paths for
                   8316: Forth source files and general files.  If the search path contains the
                   8317: directory @file{.}, this refers to the directory of the current file, or
                   8318: the working directory, as if the file had been specified with @file{./}.
1.21      crook    8319: 
1.26      crook    8320: Use @file{~+} to refer to the current working directory (as in the
                   8321: @code{bash}).
1.1       anton    8322: 
1.75      anton    8323: @c anton: fold the following subsubsections into this subsection?
1.1       anton    8324: 
1.48      anton    8325: @menu
1.75      anton    8326: * Source Search Paths::         
1.48      anton    8327: * General Search Paths::        
                   8328: @end menu
                   8329: 
1.26      crook    8330: @c ---------------------------------------------------------
1.75      anton    8331: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
                   8332: @subsubsection Source Search Paths
                   8333: @cindex search path control, source files
1.5       anton    8334: 
1.26      crook    8335: The search path is initialized when you start Gforth (@pxref{Invoking
1.75      anton    8336: Gforth}). You can display it and change it using @code{fpath} in
                   8337: combination with the general path handling words.
1.5       anton    8338: 
1.75      anton    8339: doc-fpath
                   8340: @c the functionality of the following words is easily available through
                   8341: @c   fpath and the general path words.  The may go away.
                   8342: @c doc-.fpath
                   8343: @c doc-fpath+
                   8344: @c doc-fpath=
                   8345: @c doc-open-fpath-file
1.44      crook    8346: 
                   8347: @noindent
1.26      crook    8348: Here is an example of using @code{fpath} and @code{require}:
1.5       anton    8349: 
1.26      crook    8350: @example
1.75      anton    8351: fpath path= /usr/lib/forth/|./
1.26      crook    8352: require timer.fs
                   8353: @end example
1.5       anton    8354: 
1.75      anton    8355: 
1.26      crook    8356: @c ---------------------------------------------------------
1.75      anton    8357: @node General Search Paths,  , Source Search Paths, Search Paths
1.26      crook    8358: @subsubsection General Search Paths
1.75      anton    8359: @cindex search path control, source files
1.5       anton    8360: 
1.26      crook    8361: Your application may need to search files in several directories, like
                   8362: @code{included} does. To facilitate this, Gforth allows you to define
                   8363: and use your own search paths, by providing generic equivalents of the
                   8364: Forth search path words:
1.5       anton    8365: 
1.75      anton    8366: doc-open-path-file
                   8367: doc-path-allot
                   8368: doc-clear-path
                   8369: doc-also-path
1.26      crook    8370: doc-.path
                   8371: doc-path+
                   8372: doc-path=
1.5       anton    8373: 
1.75      anton    8374: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
1.44      crook    8375: 
1.75      anton    8376: Here's an example of creating an empty search path:
                   8377: @c
1.26      crook    8378: @example
1.75      anton    8379: create mypath 500 path-allot \ maximum length 500 chars (is checked)
1.26      crook    8380: @end example
1.5       anton    8381: 
1.26      crook    8382: @c -------------------------------------------------------------
                   8383: @node Blocks, Other I/O, Files, Words
                   8384: @section Blocks
1.28      crook    8385: @cindex I/O - blocks
                   8386: @cindex blocks
                   8387: 
                   8388: When you run Gforth on a modern desk-top computer, it runs under the
                   8389: control of an operating system which provides certain services.  One of
                   8390: these services is @var{file services}, which allows Forth source code
                   8391: and data to be stored in files and read into Gforth (@pxref{Files}).
                   8392: 
                   8393: Traditionally, Forth has been an important programming language on
                   8394: systems where it has interfaced directly to the underlying hardware with
                   8395: no intervening operating system. Forth provides a mechanism, called
1.29      crook    8396: @dfn{blocks}, for accessing mass storage on such systems.
1.28      crook    8397: 
                   8398: A block is a 1024-byte data area, which can be used to hold data or
                   8399: Forth source code. No structure is imposed on the contents of the
                   8400: block. A block is identified by its number; blocks are numbered
                   8401: contiguously from 1 to an implementation-defined maximum.
                   8402: 
                   8403: A typical system that used blocks but no operating system might use a
                   8404: single floppy-disk drive for mass storage, with the disks formatted to
                   8405: provide 256-byte sectors. Blocks would be implemented by assigning the
                   8406: first four sectors of the disk to block 1, the second four sectors to
                   8407: block 2 and so on, up to the limit of the capacity of the disk. The disk
                   8408: would not contain any file system information, just the set of blocks.
                   8409: 
1.29      crook    8410: @cindex blocks file
1.28      crook    8411: On systems that do provide file services, blocks are typically
1.29      crook    8412: implemented by storing a sequence of blocks within a single @dfn{blocks
1.28      crook    8413: file}.  The size of the blocks file will be an exact multiple of 1024
                   8414: bytes, corresponding to the number of blocks it contains. This is the
                   8415: mechanism that Gforth uses.
                   8416: 
1.29      crook    8417: @cindex @file{blocks.fb}
1.75      anton    8418: Only one blocks file can be open at a time. If you use block words without
1.28      crook    8419: having specified a blocks file, Gforth defaults to the blocks file
                   8420: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
1.75      anton    8421: locate a blocks file (@pxref{Source Search Paths}).
1.28      crook    8422: 
1.29      crook    8423: @cindex block buffers
1.28      crook    8424: When you read and write blocks under program control, Gforth uses a
1.29      crook    8425: number of @dfn{block buffers} as intermediate storage. These buffers are
1.28      crook    8426: not used when you use @code{load} to interpret the contents of a block.
                   8427: 
1.75      anton    8428: The behaviour of the block buffers is analagous to that of a cache.
                   8429: Each block buffer has three states:
1.28      crook    8430: 
                   8431: @itemize @bullet
                   8432: @item
                   8433: Unassigned
                   8434: @item
                   8435: Assigned-clean
                   8436: @item
                   8437: Assigned-dirty
                   8438: @end itemize
                   8439: 
1.29      crook    8440: Initially, all block buffers are @i{unassigned}. In order to access a
1.28      crook    8441: block, the block (specified by its block number) must be assigned to a
                   8442: block buffer.
                   8443: 
                   8444: The assignment of a block to a block buffer is performed by @code{block}
                   8445: or @code{buffer}. Use @code{block} when you wish to modify the existing
                   8446: contents of a block. Use @code{buffer} when you don't care about the
                   8447: existing contents of the block@footnote{The ANS Forth definition of
1.35      anton    8448: @code{buffer} is intended not to cause disk I/O; if the data associated
1.28      crook    8449: with the particular block is already stored in a block buffer due to an
                   8450: earlier @code{block} command, @code{buffer} will return that block
                   8451: buffer and the existing contents of the block will be
                   8452: available. Otherwise, @code{buffer} will simply assign a new, empty
1.29      crook    8453: block buffer for the block.}.
1.28      crook    8454: 
1.47      crook    8455: Once a block has been assigned to a block buffer using @code{block} or
1.75      anton    8456: @code{buffer}, that block buffer becomes the @i{current block
                   8457: buffer}. Data may only be manipulated (read or written) within the
                   8458: current block buffer.
1.47      crook    8459: 
                   8460: When the contents of the current block buffer has been modified it is
1.48      anton    8461: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
1.75      anton    8462: either abandon the changes (by doing nothing) or mark the block as
                   8463: changed (assigned-dirty), using @code{update}. Using @code{update} does
                   8464: not change the blocks file; it simply changes a block buffer's state to
                   8465: @i{assigned-dirty}.  The block will be written implicitly when it's
                   8466: buffer is needed for another block, or explicitly by @code{flush} or
                   8467: @code{save-buffers}.
                   8468: 
                   8469: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
                   8470: blocks file on disk. Leaving Gforth with @code{bye} also performs a
                   8471: @code{flush}.
1.28      crook    8472: 
1.29      crook    8473: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
1.28      crook    8474: algorithm to assign a block buffer to a block. That means that any
                   8475: particular block can only be assigned to one specific block buffer,
1.29      crook    8476: called (for the particular operation) the @i{victim buffer}. If the
1.47      crook    8477: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
                   8478: the new block immediately. If it is @i{assigned-dirty} its current
                   8479: contents are written back to the blocks file on disk before it is
1.28      crook    8480: allocated to the new block.
                   8481: 
                   8482: Although no structure is imposed on the contents of a block, it is
                   8483: traditional to display the contents as 16 lines each of 64 characters.  A
                   8484: block provides a single, continuous stream of input (for example, it
                   8485: acts as a single parse area) -- there are no end-of-line characters
                   8486: within a block, and no end-of-file character at the end of a
                   8487: block. There are two consequences of this:
1.26      crook    8488: 
1.28      crook    8489: @itemize @bullet
                   8490: @item
                   8491: The last character of one line wraps straight into the first character
                   8492: of the following line
                   8493: @item
                   8494: The word @code{\} -- comment to end of line -- requires special
                   8495: treatment; in the context of a block it causes all characters until the
                   8496: end of the current 64-character ``line'' to be ignored.
                   8497: @end itemize
                   8498: 
                   8499: In Gforth, when you use @code{block} with a non-existent block number,
1.45      crook    8500: the current blocks file will be extended to the appropriate size and the
1.28      crook    8501: block buffer will be initialised with spaces.
                   8502: 
1.47      crook    8503: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
                   8504: for details) but doesn't encourage the use of blocks; the mechanism is
                   8505: only provided for backward compatibility -- ANS Forth requires blocks to
                   8506: be available when files are.
1.28      crook    8507: 
                   8508: Common techniques that are used when working with blocks include:
                   8509: 
                   8510: @itemize @bullet
                   8511: @item
                   8512: A screen editor that allows you to edit blocks without leaving the Forth
                   8513: environment.
                   8514: @item
                   8515: Shadow screens; where every code block has an associated block
                   8516: containing comments (for example: code in odd block numbers, comments in
                   8517: even block numbers). Typically, the block editor provides a convenient
                   8518: mechanism to toggle between code and comments.
                   8519: @item
                   8520: Load blocks; a single block (typically block 1) contains a number of
                   8521: @code{thru} commands which @code{load} the whole of the application.
                   8522: @end itemize
1.26      crook    8523: 
1.29      crook    8524: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
                   8525: integrated into a Forth programming environment.
1.26      crook    8526: 
                   8527: @comment TODO what about errors on open-blocks?
1.44      crook    8528: 
1.26      crook    8529: doc-open-blocks
                   8530: doc-use
1.75      anton    8531: doc-block-offset
1.26      crook    8532: doc-get-block-fid
                   8533: doc-block-position
1.28      crook    8534: 
1.75      anton    8535: doc-list
1.28      crook    8536: doc-scr
                   8537: 
1.45      crook    8538: doc---gforthman-block
1.28      crook    8539: doc-buffer
                   8540: 
1.75      anton    8541: doc-empty-buffers
                   8542: doc-empty-buffer
1.26      crook    8543: doc-update
1.28      crook    8544: doc-updated?
1.26      crook    8545: doc-save-buffers
1.75      anton    8546: doc-save-buffer
1.26      crook    8547: doc-flush
1.28      crook    8548: 
1.26      crook    8549: doc-load
                   8550: doc-thru
                   8551: doc-+load
                   8552: doc-+thru
1.45      crook    8553: doc---gforthman--->
1.26      crook    8554: doc-block-included
                   8555: 
1.44      crook    8556: 
1.26      crook    8557: @c -------------------------------------------------------------
1.126     pazsan   8558: @node Other I/O, OS command line arguments, Blocks, Words
1.26      crook    8559: @section Other I/O
1.28      crook    8560: @cindex I/O - keyboard and display
1.26      crook    8561: 
                   8562: @menu
                   8563: * Simple numeric output::       Predefined formats
                   8564: * Formatted numeric output::    Formatted (pictured) output
                   8565: * String Formats::              How Forth stores strings in memory
1.67      anton    8566: * Displaying characters and strings::  Other stuff
1.26      crook    8567: * Input::                       Input
1.112     anton    8568: * Pipes::                       How to create your own pipes
1.26      crook    8569: @end menu
                   8570: 
                   8571: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
                   8572: @subsection Simple numeric output
1.28      crook    8573: @cindex numeric output - simple/free-format
1.5       anton    8574: 
1.26      crook    8575: The simplest output functions are those that display numbers from the
                   8576: data or floating-point stacks. Floating-point output is always displayed
                   8577: using base 10. Numbers displayed from the data stack use the value stored
                   8578: in @code{base}.
1.5       anton    8579: 
1.44      crook    8580: 
1.26      crook    8581: doc-.
                   8582: doc-dec.
                   8583: doc-hex.
                   8584: doc-u.
                   8585: doc-.r
                   8586: doc-u.r
                   8587: doc-d.
                   8588: doc-ud.
                   8589: doc-d.r
                   8590: doc-ud.r
                   8591: doc-f.
                   8592: doc-fe.
                   8593: doc-fs.
1.111     anton    8594: doc-f.rdp
1.44      crook    8595: 
1.26      crook    8596: Examples of printing the number 1234.5678E23 in the different floating-point output
                   8597: formats are shown below:
1.5       anton    8598: 
                   8599: @example
1.26      crook    8600: f. 123456779999999000000000000.
                   8601: fe. 123.456779999999E24
                   8602: fs. 1.23456779999999E26
1.5       anton    8603: @end example
                   8604: 
                   8605: 
1.26      crook    8606: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
                   8607: @subsection Formatted numeric output
1.28      crook    8608: @cindex formatted numeric output
1.26      crook    8609: @cindex pictured numeric output
1.28      crook    8610: @cindex numeric output - formatted
1.26      crook    8611: 
1.29      crook    8612: Forth traditionally uses a technique called @dfn{pictured numeric
1.26      crook    8613: output} for formatted printing of integers.  In this technique, digits
                   8614: are extracted from the number (using the current output radix defined by
                   8615: @code{base}), converted to ASCII codes and appended to a string that is
                   8616: built in a scratch-pad area of memory (@pxref{core-idef,
                   8617: Implementation-defined options, Implementation-defined
                   8618: options}). Arbitrary characters can be appended to the string during the
                   8619: extraction process. The completed string is specified by an address
                   8620: and length and can be manipulated (@code{TYPE}ed, copied, modified)
                   8621: under program control.
1.5       anton    8622: 
1.75      anton    8623: All of the integer output words described in the previous section
                   8624: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
                   8625: numeric output.
1.5       anton    8626: 
1.47      crook    8627: Three important things to remember about pictured numeric output:
1.5       anton    8628: 
1.26      crook    8629: @itemize @bullet
                   8630: @item
1.28      crook    8631: It always operates on double-precision numbers; to display a
1.49      anton    8632: single-precision number, convert it first (for ways of doing this
                   8633: @pxref{Double precision}).
1.26      crook    8634: @item
1.28      crook    8635: It always treats the double-precision number as though it were
                   8636: unsigned. The examples below show ways of printing signed numbers.
1.26      crook    8637: @item
                   8638: The string is built up from right to left; least significant digit first.
                   8639: @end itemize
1.5       anton    8640: 
1.44      crook    8641: 
1.26      crook    8642: doc-<#
1.47      crook    8643: doc-<<#
1.26      crook    8644: doc-#
                   8645: doc-#s
                   8646: doc-hold
                   8647: doc-sign
                   8648: doc-#>
1.47      crook    8649: doc-#>>
1.5       anton    8650: 
1.26      crook    8651: doc-represent
1.111     anton    8652: doc-f>str-rdp
                   8653: doc-f>buf-rdp
1.5       anton    8654: 
1.44      crook    8655: 
                   8656: @noindent
1.26      crook    8657: Here are some examples of using pictured numeric output:
1.5       anton    8658: 
                   8659: @example
1.26      crook    8660: : my-u. ( u -- )
                   8661:   \ Simplest use of pns.. behaves like Standard u. 
                   8662:   0              \ convert to unsigned double
1.75      anton    8663:   <<#            \ start conversion
1.26      crook    8664:   #s             \ convert all digits
                   8665:   #>             \ complete conversion
1.75      anton    8666:   TYPE SPACE     \ display, with trailing space
                   8667:   #>> ;          \ release hold area
1.5       anton    8668: 
1.26      crook    8669: : cents-only ( u -- )
                   8670:   0              \ convert to unsigned double
1.75      anton    8671:   <<#            \ start conversion
1.26      crook    8672:   # #            \ convert two least-significant digits
                   8673:   #>             \ complete conversion, discard other digits
1.75      anton    8674:   TYPE SPACE     \ display, with trailing space
                   8675:   #>> ;          \ release hold area
1.5       anton    8676: 
1.26      crook    8677: : dollars-and-cents ( u -- )
                   8678:   0              \ convert to unsigned double
1.75      anton    8679:   <<#            \ start conversion
1.26      crook    8680:   # #            \ convert two least-significant digits
                   8681:   [char] . hold  \ insert decimal point
                   8682:   #s             \ convert remaining digits
                   8683:   [char] $ hold  \ append currency symbol
                   8684:   #>             \ complete conversion
1.75      anton    8685:   TYPE SPACE     \ display, with trailing space
                   8686:   #>> ;          \ release hold area
1.5       anton    8687: 
1.26      crook    8688: : my-. ( n -- )
                   8689:   \ handling negatives.. behaves like Standard .
                   8690:   s>d            \ convert to signed double
                   8691:   swap over dabs \ leave sign byte followed by unsigned double
1.75      anton    8692:   <<#            \ start conversion
1.26      crook    8693:   #s             \ convert all digits
                   8694:   rot sign       \ get at sign byte, append "-" if needed
                   8695:   #>             \ complete conversion
1.75      anton    8696:   TYPE SPACE     \ display, with trailing space
                   8697:   #>> ;          \ release hold area
1.5       anton    8698: 
1.26      crook    8699: : account. ( n -- )
1.75      anton    8700:   \ accountants don't like minus signs, they use parentheses
1.26      crook    8701:   \ for negative numbers
                   8702:   s>d            \ convert to signed double
                   8703:   swap over dabs \ leave sign byte followed by unsigned double
1.75      anton    8704:   <<#            \ start conversion
1.26      crook    8705:   2 pick         \ get copy of sign byte
                   8706:   0< IF [char] ) hold THEN \ right-most character of output
                   8707:   #s             \ convert all digits
                   8708:   rot            \ get at sign byte
                   8709:   0< IF [char] ( hold THEN
                   8710:   #>             \ complete conversion
1.75      anton    8711:   TYPE SPACE     \ display, with trailing space
                   8712:   #>> ;          \ release hold area
                   8713: 
1.5       anton    8714: @end example
                   8715: 
1.26      crook    8716: Here are some examples of using these words:
1.5       anton    8717: 
                   8718: @example
1.26      crook    8719: 1 my-u. 1
                   8720: hex -1 my-u. decimal FFFFFFFF
                   8721: 1 cents-only 01
                   8722: 1234 cents-only 34
                   8723: 2 dollars-and-cents $0.02
                   8724: 1234 dollars-and-cents $12.34
                   8725: 123 my-. 123
                   8726: -123 my. -123
                   8727: 123 account. 123
                   8728: -456 account. (456)
1.5       anton    8729: @end example
                   8730: 
                   8731: 
1.26      crook    8732: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
                   8733: @subsection String Formats
1.27      crook    8734: @cindex strings - see character strings
                   8735: @cindex character strings - formats
1.28      crook    8736: @cindex I/O - see character strings
1.75      anton    8737: @cindex counted strings
                   8738: 
                   8739: @c anton: this does not really belong here; maybe the memory section,
                   8740: @c  or the principles chapter
1.26      crook    8741: 
1.27      crook    8742: Forth commonly uses two different methods for representing character
                   8743: strings:
1.26      crook    8744: 
                   8745: @itemize @bullet
                   8746: @item
                   8747: @cindex address of counted string
1.45      crook    8748: @cindex counted string
1.29      crook    8749: As a @dfn{counted string}, represented by a @i{c-addr}. The char
                   8750: addressed by @i{c-addr} contains a character-count, @i{n}, of the
                   8751: string and the string occupies the subsequent @i{n} char addresses in
1.26      crook    8752: memory.
                   8753: @item
1.29      crook    8754: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
                   8755: of the string in characters, and @i{c-addr} is the address of the
1.26      crook    8756: first byte of the string.
                   8757: @end itemize
                   8758: 
                   8759: ANS Forth encourages the use of the second format when representing
1.75      anton    8760: strings.
1.26      crook    8761: 
1.44      crook    8762: 
1.26      crook    8763: doc-count
                   8764: 
1.44      crook    8765: 
1.49      anton    8766: For words that move, copy and search for strings see @ref{Memory
                   8767: Blocks}. For words that display characters and strings see
                   8768: @ref{Displaying characters and strings}.
1.26      crook    8769: 
                   8770: @node Displaying characters and strings, Input, String Formats, Other I/O
                   8771: @subsection Displaying characters and strings
1.27      crook    8772: @cindex characters - compiling and displaying
                   8773: @cindex character strings - compiling and displaying
1.26      crook    8774: 
                   8775: This section starts with a glossary of Forth words and ends with a set
                   8776: of examples.
                   8777: 
1.44      crook    8778: 
1.26      crook    8779: doc-bl
                   8780: doc-space
                   8781: doc-spaces
                   8782: doc-emit
                   8783: doc-toupper
                   8784: doc-."
                   8785: doc-.(
1.98      anton    8786: doc-.\"
1.26      crook    8787: doc-type
1.44      crook    8788: doc-typewhite
1.26      crook    8789: doc-cr
1.27      crook    8790: @cindex cursor control
1.26      crook    8791: doc-at-xy
                   8792: doc-page
                   8793: doc-s"
1.98      anton    8794: doc-s\"
1.26      crook    8795: doc-c"
                   8796: doc-char
                   8797: doc-[char]
                   8798: 
1.44      crook    8799: 
                   8800: @noindent
1.26      crook    8801: As an example, consider the following text, stored in a file @file{test.fs}:
1.5       anton    8802: 
                   8803: @example
1.26      crook    8804: .( text-1)
                   8805: : my-word
                   8806:   ." text-2" cr
                   8807:   .( text-3)
                   8808: ;
                   8809: 
                   8810: ." text-4"
                   8811: 
                   8812: : my-char
                   8813:   [char] ALPHABET emit
                   8814:   char emit
                   8815: ;
1.5       anton    8816: @end example
                   8817: 
1.26      crook    8818: When you load this code into Gforth, the following output is generated:
1.5       anton    8819: 
1.26      crook    8820: @example
1.30      anton    8821: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
1.26      crook    8822: @end example
1.5       anton    8823: 
1.26      crook    8824: @itemize @bullet
                   8825: @item
                   8826: Messages @code{text-1} and @code{text-3} are displayed because @code{.(} 
                   8827: is an immediate word; it behaves in the same way whether it is used inside
                   8828: or outside a colon definition.
                   8829: @item
                   8830: Message @code{text-4} is displayed because of Gforth's added interpretation
                   8831: semantics for @code{."}.
                   8832: @item
1.29      crook    8833: Message @code{text-2} is @i{not} displayed, because the text interpreter
1.26      crook    8834: performs the compilation semantics for @code{."} within the definition of
                   8835: @code{my-word}.
                   8836: @end itemize
1.5       anton    8837: 
1.26      crook    8838: Here are some examples of executing @code{my-word} and @code{my-char}:
1.5       anton    8839: 
1.26      crook    8840: @example
1.30      anton    8841: @kbd{my-word @key{RET}} text-2
1.26      crook    8842:  ok
1.30      anton    8843: @kbd{my-char fred @key{RET}} Af ok
                   8844: @kbd{my-char jim @key{RET}} Aj ok
1.26      crook    8845: @end example
1.5       anton    8846: 
                   8847: @itemize @bullet
                   8848: @item
1.26      crook    8849: Message @code{text-2} is displayed because of the run-time behaviour of
                   8850: @code{."}.
                   8851: @item
                   8852: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
                   8853: on the stack at run-time. @code{emit} always displays the character
                   8854: when @code{my-char} is executed.
                   8855: @item
                   8856: @code{char} parses a string at run-time and the second @code{emit} displays
                   8857: the first character of the string.
1.5       anton    8858: @item
1.26      crook    8859: If you type @code{see my-char} you can see that @code{[char]} discarded
                   8860: the text ``LPHABET'' and only compiled the display code for ``A'' into the
                   8861: definition of @code{my-char}.
1.5       anton    8862: @end itemize
                   8863: 
                   8864: 
                   8865: 
1.112     anton    8866: @node Input, Pipes, Displaying characters and strings, Other I/O
1.26      crook    8867: @subsection Input
                   8868: @cindex input
1.28      crook    8869: @cindex I/O - see input
                   8870: @cindex parsing a string
1.5       anton    8871: 
1.49      anton    8872: For ways of storing character strings in memory see @ref{String Formats}.
1.5       anton    8873: 
1.27      crook    8874: @comment TODO examples for >number >float accept key key? pad parse word refill
1.29      crook    8875: @comment then index them
1.27      crook    8876: 
1.44      crook    8877: 
1.27      crook    8878: doc-key
                   8879: doc-key?
1.45      crook    8880: doc-ekey
1.141     anton    8881: doc-ekey>char
1.45      crook    8882: doc-ekey?
1.141     anton    8883: 
                   8884: Gforth recognizes various keys available on ANSI terminals (in MS-DOS
                   8885: you need the ANSI.SYS driver to get that behaviour).  These are the
                   8886: keyboard events produced by various common keys:
                   8887: 
                   8888: doc-k-left
                   8889: doc-k-right
                   8890: doc-k-up       
                   8891: doc-k-down     
                   8892: doc-k-home     
                   8893: doc-k-end      
                   8894: doc-k-prior
                   8895: doc-k-next
                   8896: doc-k-insert
                   8897: doc-k-delete
                   8898: 
                   8899: The function keys (aka keypad keys) are:
                   8900: 
                   8901: doc-k1
                   8902: doc-k2
                   8903: doc-k3
                   8904: doc-k4
                   8905: doc-k5
                   8906: doc-k6
                   8907: doc-k7
                   8908: doc-k8
                   8909: doc-k9
                   8910: doc-k10
                   8911: doc-k11
                   8912: doc-k12
                   8913: 
                   8914: Note that K11 and K12 are not as widely available.  The shifted
                   8915: function keys are also not very widely available:
                   8916: 
                   8917: doc-s-k8
                   8918: doc-s-k1
                   8919: doc-s-k2
                   8920: doc-s-k3
                   8921: doc-s-k4
                   8922: doc-s-k5
                   8923: doc-s-k6
                   8924: doc-s-k7
                   8925: doc-s-k8
                   8926: doc-s-k9
                   8927: doc-s-k10
                   8928: doc-s-k11
                   8929: doc-s-k12
                   8930: 
                   8931: Words for inputting one line from the keyboard:
                   8932: 
                   8933: doc-accept
                   8934: doc-edit-line
                   8935: 
                   8936: Conversion words:
                   8937: 
1.143     anton    8938: doc-s>number?
                   8939: doc-s>unumber?
1.26      crook    8940: doc->number
                   8941: doc->float
1.143     anton    8942: 
1.141     anton    8943: 
1.27      crook    8944: @comment obsolescent words..
1.141     anton    8945: Obsolescent input and conversion words:
                   8946: 
1.27      crook    8947: doc-convert
1.26      crook    8948: doc-expect
1.27      crook    8949: doc-span
1.5       anton    8950: 
                   8951: 
1.112     anton    8952: @node Pipes,  , Input, Other I/O
                   8953: @subsection Pipes
                   8954: @cindex pipes, creating your own
                   8955: 
                   8956: In addition to using Gforth in pipes created by other processes
                   8957: (@pxref{Gforth in pipes}), you can create your own pipe with
                   8958: @code{open-pipe}, and read from or write to it.
                   8959: 
                   8960: doc-open-pipe
                   8961: doc-close-pipe
                   8962: 
                   8963: If you write to a pipe, Gforth can throw a @code{broken-pipe-error}; if
                   8964: you don't catch this exception, Gforth will catch it and exit, usually
                   8965: silently (@pxref{Gforth in pipes}).  Since you probably do not want
                   8966: this, you should wrap a @code{catch} or @code{try} block around the code
                   8967: from @code{open-pipe} to @code{close-pipe}, so you can deal with the
                   8968: problem yourself, and then return to regular processing.
                   8969: 
                   8970: doc-broken-pipe-error
                   8971: 
                   8972: 
1.121     anton    8973: @node OS command line arguments, Locals, Other I/O, Words
                   8974: @section OS command line arguments
                   8975: @cindex OS command line arguments
                   8976: @cindex command line arguments, OS
                   8977: @cindex arguments, OS command line
                   8978: 
                   8979: The usual way to pass arguments to Gforth programs on the command line
                   8980: is via the @option{-e} option, e.g.
                   8981: 
                   8982: @example
                   8983: gforth -e "123 456" foo.fs -e bye
                   8984: @end example
                   8985: 
                   8986: However, you may want to interpret the command-line arguments directly.
                   8987: In that case, you can access the (image-specific) command-line arguments
1.123     anton    8988: through @code{next-arg}:
1.121     anton    8989: 
1.123     anton    8990: doc-next-arg
1.121     anton    8991: 
1.123     anton    8992: Here's an example program @file{echo.fs} for @code{next-arg}:
1.121     anton    8993: 
                   8994: @example
                   8995: : echo ( -- )
1.122     anton    8996:     begin
1.123     anton    8997:        next-arg 2dup 0 0 d<> while
                   8998:            type space
                   8999:     repeat
                   9000:     2drop ;
1.121     anton    9001: 
                   9002: echo cr bye
                   9003: @end example
                   9004: 
                   9005: This can be invoked with
                   9006: 
                   9007: @example
                   9008: gforth echo.fs hello world
                   9009: @end example
1.123     anton    9010: 
                   9011: and it will print
                   9012: 
                   9013: @example
                   9014: hello world
                   9015: @end example
                   9016: 
                   9017: The next lower level of dealing with the OS command line are the
                   9018: following words:
                   9019: 
                   9020: doc-arg
                   9021: doc-shift-args
                   9022: 
                   9023: Finally, at the lowest level Gforth provides the following words:
                   9024: 
                   9025: doc-argc
                   9026: doc-argv
1.121     anton    9027: 
1.78      anton    9028: @c -------------------------------------------------------------
1.126     pazsan   9029: @node Locals, Structures, OS command line arguments, Words
1.78      anton    9030: @section Locals
                   9031: @cindex locals
                   9032: 
                   9033: Local variables can make Forth programming more enjoyable and Forth
                   9034: programs easier to read. Unfortunately, the locals of ANS Forth are
                   9035: laden with restrictions. Therefore, we provide not only the ANS Forth
                   9036: locals wordset, but also our own, more powerful locals wordset (we
                   9037: implemented the ANS Forth locals wordset through our locals wordset).
1.44      crook    9038: 
1.78      anton    9039: The ideas in this section have also been published in M. Anton Ertl,
                   9040: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
                   9041: Automatic Scoping of Local Variables}}, EuroForth '94.
1.12      anton    9042: 
                   9043: @menu
1.78      anton    9044: * Gforth locals::               
                   9045: * ANS Forth locals::            
1.5       anton    9046: @end menu
                   9047: 
1.78      anton    9048: @node Gforth locals, ANS Forth locals, Locals, Locals
                   9049: @subsection Gforth locals
                   9050: @cindex Gforth locals
                   9051: @cindex locals, Gforth style
1.5       anton    9052: 
1.78      anton    9053: Locals can be defined with
1.44      crook    9054: 
1.78      anton    9055: @example
                   9056: @{ local1 local2 ... -- comment @}
                   9057: @end example
                   9058: or
                   9059: @example
                   9060: @{ local1 local2 ... @}
                   9061: @end example
1.5       anton    9062: 
1.78      anton    9063: E.g.,
                   9064: @example
                   9065: : max @{ n1 n2 -- n3 @}
                   9066:  n1 n2 > if
                   9067:    n1
                   9068:  else
                   9069:    n2
                   9070:  endif ;
                   9071: @end example
1.44      crook    9072: 
1.78      anton    9073: The similarity of locals definitions with stack comments is intended. A
                   9074: locals definition often replaces the stack comment of a word. The order
                   9075: of the locals corresponds to the order in a stack comment and everything
                   9076: after the @code{--} is really a comment.
1.77      anton    9077: 
1.78      anton    9078: This similarity has one disadvantage: It is too easy to confuse locals
                   9079: declarations with stack comments, causing bugs and making them hard to
                   9080: find. However, this problem can be avoided by appropriate coding
                   9081: conventions: Do not use both notations in the same program. If you do,
                   9082: they should be distinguished using additional means, e.g. by position.
1.77      anton    9083: 
1.78      anton    9084: @cindex types of locals
                   9085: @cindex locals types
                   9086: The name of the local may be preceded by a type specifier, e.g.,
                   9087: @code{F:} for a floating point value:
1.5       anton    9088: 
1.78      anton    9089: @example
                   9090: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
                   9091: \ complex multiplication
                   9092:  Ar Br f* Ai Bi f* f-
                   9093:  Ar Bi f* Ai Br f* f+ ;
                   9094: @end example
1.44      crook    9095: 
1.78      anton    9096: @cindex flavours of locals
                   9097: @cindex locals flavours
                   9098: @cindex value-flavoured locals
                   9099: @cindex variable-flavoured locals
                   9100: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
                   9101: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
                   9102: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
                   9103: with @code{W:}, @code{D:} etc.) produces its value and can be changed
                   9104: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
                   9105: produces its address (which becomes invalid when the variable's scope is
                   9106: left). E.g., the standard word @code{emit} can be defined in terms of
                   9107: @code{type} like this:
1.5       anton    9108: 
1.78      anton    9109: @example
                   9110: : emit @{ C^ char* -- @}
                   9111:     char* 1 type ;
                   9112: @end example
1.5       anton    9113: 
1.78      anton    9114: @cindex default type of locals
                   9115: @cindex locals, default type
                   9116: A local without type specifier is a @code{W:} local. Both flavours of
                   9117: locals are initialized with values from the data or FP stack.
1.44      crook    9118: 
1.78      anton    9119: Currently there is no way to define locals with user-defined data
                   9120: structures, but we are working on it.
1.5       anton    9121: 
1.78      anton    9122: Gforth allows defining locals everywhere in a colon definition. This
                   9123: poses the following questions:
1.5       anton    9124: 
1.78      anton    9125: @menu
                   9126: * Where are locals visible by name?::  
                   9127: * How long do locals live?::    
                   9128: * Locals programming style::    
                   9129: * Locals implementation::       
                   9130: @end menu
1.44      crook    9131: 
1.78      anton    9132: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
                   9133: @subsubsection Where are locals visible by name?
                   9134: @cindex locals visibility
                   9135: @cindex visibility of locals
                   9136: @cindex scope of locals
1.5       anton    9137: 
1.78      anton    9138: Basically, the answer is that locals are visible where you would expect
                   9139: it in block-structured languages, and sometimes a little longer. If you
                   9140: want to restrict the scope of a local, enclose its definition in
                   9141: @code{SCOPE}...@code{ENDSCOPE}.
1.5       anton    9142: 
                   9143: 
1.78      anton    9144: doc-scope
                   9145: doc-endscope
1.5       anton    9146: 
                   9147: 
1.78      anton    9148: These words behave like control structure words, so you can use them
                   9149: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
                   9150: arbitrary ways.
1.77      anton    9151: 
1.78      anton    9152: If you want a more exact answer to the visibility question, here's the
                   9153: basic principle: A local is visible in all places that can only be
                   9154: reached through the definition of the local@footnote{In compiler
                   9155: construction terminology, all places dominated by the definition of the
                   9156: local.}. In other words, it is not visible in places that can be reached
                   9157: without going through the definition of the local. E.g., locals defined
                   9158: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
                   9159: defined in @code{BEGIN}...@code{UNTIL} are visible after the
                   9160: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1.77      anton    9161: 
1.78      anton    9162: The reasoning behind this solution is: We want to have the locals
                   9163: visible as long as it is meaningful. The user can always make the
                   9164: visibility shorter by using explicit scoping. In a place that can
                   9165: only be reached through the definition of a local, the meaning of a
                   9166: local name is clear. In other places it is not: How is the local
                   9167: initialized at the control flow path that does not contain the
                   9168: definition? Which local is meant, if the same name is defined twice in
                   9169: two independent control flow paths?
1.77      anton    9170: 
1.78      anton    9171: This should be enough detail for nearly all users, so you can skip the
                   9172: rest of this section. If you really must know all the gory details and
                   9173: options, read on.
1.77      anton    9174: 
1.78      anton    9175: In order to implement this rule, the compiler has to know which places
                   9176: are unreachable. It knows this automatically after @code{AHEAD},
                   9177: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
                   9178: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
                   9179: compiler that the control flow never reaches that place. If
                   9180: @code{UNREACHABLE} is not used where it could, the only consequence is
                   9181: that the visibility of some locals is more limited than the rule above
                   9182: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
                   9183: lie to the compiler), buggy code will be produced.
1.77      anton    9184: 
1.5       anton    9185: 
1.78      anton    9186: doc-unreachable
1.5       anton    9187: 
1.23      crook    9188: 
1.78      anton    9189: Another problem with this rule is that at @code{BEGIN}, the compiler
                   9190: does not know which locals will be visible on the incoming
                   9191: back-edge. All problems discussed in the following are due to this
                   9192: ignorance of the compiler (we discuss the problems using @code{BEGIN}
                   9193: loops as examples; the discussion also applies to @code{?DO} and other
                   9194: loops). Perhaps the most insidious example is:
1.26      crook    9195: @example
1.78      anton    9196: AHEAD
                   9197: BEGIN
                   9198:   x
                   9199: [ 1 CS-ROLL ] THEN
                   9200:   @{ x @}
                   9201:   ...
                   9202: UNTIL
1.26      crook    9203: @end example
1.23      crook    9204: 
1.78      anton    9205: This should be legal according to the visibility rule. The use of
                   9206: @code{x} can only be reached through the definition; but that appears
                   9207: textually below the use.
                   9208: 
                   9209: From this example it is clear that the visibility rules cannot be fully
                   9210: implemented without major headaches. Our implementation treats common
                   9211: cases as advertised and the exceptions are treated in a safe way: The
                   9212: compiler makes a reasonable guess about the locals visible after a
                   9213: @code{BEGIN}; if it is too pessimistic, the
                   9214: user will get a spurious error about the local not being defined; if the
                   9215: compiler is too optimistic, it will notice this later and issue a
                   9216: warning. In the case above the compiler would complain about @code{x}
                   9217: being undefined at its use. You can see from the obscure examples in
                   9218: this section that it takes quite unusual control structures to get the
                   9219: compiler into trouble, and even then it will often do fine.
1.23      crook    9220: 
1.78      anton    9221: If the @code{BEGIN} is reachable from above, the most optimistic guess
                   9222: is that all locals visible before the @code{BEGIN} will also be
                   9223: visible after the @code{BEGIN}. This guess is valid for all loops that
                   9224: are entered only through the @code{BEGIN}, in particular, for normal
                   9225: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
                   9226: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
                   9227: compiler. When the branch to the @code{BEGIN} is finally generated by
                   9228: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
                   9229: warns the user if it was too optimistic:
1.26      crook    9230: @example
1.78      anton    9231: IF
                   9232:   @{ x @}
                   9233: BEGIN
                   9234:   \ x ? 
                   9235: [ 1 cs-roll ] THEN
                   9236:   ...
                   9237: UNTIL
1.26      crook    9238: @end example
1.23      crook    9239: 
1.78      anton    9240: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
                   9241: optimistically assumes that it lives until the @code{THEN}. It notices
                   9242: this difference when it compiles the @code{UNTIL} and issues a
                   9243: warning. The user can avoid the warning, and make sure that @code{x}
                   9244: is not used in the wrong area by using explicit scoping:
                   9245: @example
                   9246: IF
                   9247:   SCOPE
                   9248:   @{ x @}
                   9249:   ENDSCOPE
                   9250: BEGIN
                   9251: [ 1 cs-roll ] THEN
                   9252:   ...
                   9253: UNTIL
                   9254: @end example
1.23      crook    9255: 
1.78      anton    9256: Since the guess is optimistic, there will be no spurious error messages
                   9257: about undefined locals.
1.44      crook    9258: 
1.78      anton    9259: If the @code{BEGIN} is not reachable from above (e.g., after
                   9260: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
                   9261: optimistic guess, as the locals visible after the @code{BEGIN} may be
                   9262: defined later. Therefore, the compiler assumes that no locals are
                   9263: visible after the @code{BEGIN}. However, the user can use
                   9264: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
                   9265: visible at the BEGIN as at the point where the top control-flow stack
                   9266: item was created.
1.23      crook    9267: 
1.44      crook    9268: 
1.78      anton    9269: doc-assume-live
1.26      crook    9270: 
1.23      crook    9271: 
1.78      anton    9272: @noindent
                   9273: E.g.,
                   9274: @example
                   9275: @{ x @}
                   9276: AHEAD
                   9277: ASSUME-LIVE
                   9278: BEGIN
                   9279:   x
                   9280: [ 1 CS-ROLL ] THEN
                   9281:   ...
                   9282: UNTIL
                   9283: @end example
1.44      crook    9284: 
1.78      anton    9285: Other cases where the locals are defined before the @code{BEGIN} can be
                   9286: handled by inserting an appropriate @code{CS-ROLL} before the
                   9287: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
                   9288: behind the @code{ASSUME-LIVE}).
1.23      crook    9289: 
1.78      anton    9290: Cases where locals are defined after the @code{BEGIN} (but should be
                   9291: visible immediately after the @code{BEGIN}) can only be handled by
                   9292: rearranging the loop. E.g., the ``most insidious'' example above can be
                   9293: arranged into:
                   9294: @example
                   9295: BEGIN
                   9296:   @{ x @}
                   9297:   ... 0=
                   9298: WHILE
                   9299:   x
                   9300: REPEAT
                   9301: @end example
1.44      crook    9302: 
1.78      anton    9303: @node How long do locals live?, Locals programming style, Where are locals visible by name?, Gforth locals
                   9304: @subsubsection How long do locals live?
                   9305: @cindex locals lifetime
                   9306: @cindex lifetime of locals
1.23      crook    9307: 
1.78      anton    9308: The right answer for the lifetime question would be: A local lives at
                   9309: least as long as it can be accessed. For a value-flavoured local this
                   9310: means: until the end of its visibility. However, a variable-flavoured
                   9311: local could be accessed through its address far beyond its visibility
                   9312: scope. Ultimately, this would mean that such locals would have to be
                   9313: garbage collected. Since this entails un-Forth-like implementation
                   9314: complexities, I adopted the same cowardly solution as some other
                   9315: languages (e.g., C): The local lives only as long as it is visible;
                   9316: afterwards its address is invalid (and programs that access it
                   9317: afterwards are erroneous).
1.23      crook    9318: 
1.78      anton    9319: @node Locals programming style, Locals implementation, How long do locals live?, Gforth locals
                   9320: @subsubsection Locals programming style
                   9321: @cindex locals programming style
                   9322: @cindex programming style, locals
1.23      crook    9323: 
1.78      anton    9324: The freedom to define locals anywhere has the potential to change
                   9325: programming styles dramatically. In particular, the need to use the
                   9326: return stack for intermediate storage vanishes. Moreover, all stack
                   9327: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
                   9328: determined arguments) can be eliminated: If the stack items are in the
                   9329: wrong order, just write a locals definition for all of them; then
                   9330: write the items in the order you want.
1.23      crook    9331: 
1.78      anton    9332: This seems a little far-fetched and eliminating stack manipulations is
                   9333: unlikely to become a conscious programming objective. Still, the number
                   9334: of stack manipulations will be reduced dramatically if local variables
                   9335: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
                   9336: a traditional implementation of @code{max}).
1.23      crook    9337: 
1.78      anton    9338: This shows one potential benefit of locals: making Forth programs more
                   9339: readable. Of course, this benefit will only be realized if the
                   9340: programmers continue to honour the principle of factoring instead of
                   9341: using the added latitude to make the words longer.
1.23      crook    9342: 
1.78      anton    9343: @cindex single-assignment style for locals
                   9344: Using @code{TO} can and should be avoided.  Without @code{TO},
                   9345: every value-flavoured local has only a single assignment and many
                   9346: advantages of functional languages apply to Forth. I.e., programs are
                   9347: easier to analyse, to optimize and to read: It is clear from the
                   9348: definition what the local stands for, it does not turn into something
                   9349: different later.
1.23      crook    9350: 
1.78      anton    9351: E.g., a definition using @code{TO} might look like this:
                   9352: @example
                   9353: : strcmp @{ addr1 u1 addr2 u2 -- n @}
                   9354:  u1 u2 min 0
                   9355:  ?do
                   9356:    addr1 c@@ addr2 c@@ -
                   9357:    ?dup-if
                   9358:      unloop exit
                   9359:    then
                   9360:    addr1 char+ TO addr1
                   9361:    addr2 char+ TO addr2
                   9362:  loop
                   9363:  u1 u2 - ;
1.26      crook    9364: @end example
1.78      anton    9365: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
                   9366: every loop iteration. @code{strcmp} is a typical example of the
                   9367: readability problems of using @code{TO}. When you start reading
                   9368: @code{strcmp}, you think that @code{addr1} refers to the start of the
                   9369: string. Only near the end of the loop you realize that it is something
                   9370: else.
1.23      crook    9371: 
1.78      anton    9372: This can be avoided by defining two locals at the start of the loop that
                   9373: are initialized with the right value for the current iteration.
                   9374: @example
                   9375: : strcmp @{ addr1 u1 addr2 u2 -- n @}
                   9376:  addr1 addr2
                   9377:  u1 u2 min 0 
                   9378:  ?do @{ s1 s2 @}
                   9379:    s1 c@@ s2 c@@ -
                   9380:    ?dup-if
                   9381:      unloop exit
                   9382:    then
                   9383:    s1 char+ s2 char+
                   9384:  loop
                   9385:  2drop
                   9386:  u1 u2 - ;
                   9387: @end example
                   9388: Here it is clear from the start that @code{s1} has a different value
                   9389: in every loop iteration.
1.23      crook    9390: 
1.78      anton    9391: @node Locals implementation,  , Locals programming style, Gforth locals
                   9392: @subsubsection Locals implementation
                   9393: @cindex locals implementation
                   9394: @cindex implementation of locals
1.23      crook    9395: 
1.78      anton    9396: @cindex locals stack
                   9397: Gforth uses an extra locals stack. The most compelling reason for
                   9398: this is that the return stack is not float-aligned; using an extra stack
                   9399: also eliminates the problems and restrictions of using the return stack
                   9400: as locals stack. Like the other stacks, the locals stack grows toward
                   9401: lower addresses. A few primitives allow an efficient implementation:
                   9402: 
                   9403: 
                   9404: doc-@local#
                   9405: doc-f@local#
                   9406: doc-laddr#
                   9407: doc-lp+!#
                   9408: doc-lp!
                   9409: doc->l
                   9410: doc-f>l
                   9411: 
                   9412: 
                   9413: In addition to these primitives, some specializations of these
                   9414: primitives for commonly occurring inline arguments are provided for
                   9415: efficiency reasons, e.g., @code{@@local0} as specialization of
                   9416: @code{@@local#} for the inline argument 0. The following compiling words
                   9417: compile the right specialized version, or the general version, as
                   9418: appropriate:
1.23      crook    9419: 
1.5       anton    9420: 
1.107     dvdkhlng 9421: @c doc-compile-@local
                   9422: @c doc-compile-f@local
1.78      anton    9423: doc-compile-lp+!
1.5       anton    9424: 
                   9425: 
1.78      anton    9426: Combinations of conditional branches and @code{lp+!#} like
                   9427: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
                   9428: is taken) are provided for efficiency and correctness in loops.
1.5       anton    9429: 
1.78      anton    9430: A special area in the dictionary space is reserved for keeping the
                   9431: local variable names. @code{@{} switches the dictionary pointer to this
                   9432: area and @code{@}} switches it back and generates the locals
                   9433: initializing code. @code{W:} etc.@ are normal defining words. This
                   9434: special area is cleared at the start of every colon definition.
1.5       anton    9435: 
1.78      anton    9436: @cindex word list for defining locals
                   9437: A special feature of Gforth's dictionary is used to implement the
                   9438: definition of locals without type specifiers: every word list (aka
                   9439: vocabulary) has its own methods for searching
                   9440: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
                   9441: with a special search method: When it is searched for a word, it
                   9442: actually creates that word using @code{W:}. @code{@{} changes the search
                   9443: order to first search the word list containing @code{@}}, @code{W:} etc.,
                   9444: and then the word list for defining locals without type specifiers.
1.5       anton    9445: 
1.78      anton    9446: The lifetime rules support a stack discipline within a colon
                   9447: definition: The lifetime of a local is either nested with other locals
                   9448: lifetimes or it does not overlap them.
1.23      crook    9449: 
1.78      anton    9450: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
                   9451: pointer manipulation is generated. Between control structure words
                   9452: locals definitions can push locals onto the locals stack. @code{AGAIN}
                   9453: is the simplest of the other three control flow words. It has to
                   9454: restore the locals stack depth of the corresponding @code{BEGIN}
                   9455: before branching. The code looks like this:
                   9456: @format
                   9457: @code{lp+!#} current-locals-size @minus{} dest-locals-size
                   9458: @code{branch} <begin>
                   9459: @end format
1.26      crook    9460: 
1.78      anton    9461: @code{UNTIL} is a little more complicated: If it branches back, it
                   9462: must adjust the stack just like @code{AGAIN}. But if it falls through,
                   9463: the locals stack must not be changed. The compiler generates the
                   9464: following code:
                   9465: @format
                   9466: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
                   9467: @end format
                   9468: The locals stack pointer is only adjusted if the branch is taken.
1.44      crook    9469: 
1.78      anton    9470: @code{THEN} can produce somewhat inefficient code:
                   9471: @format
                   9472: @code{lp+!#} current-locals-size @minus{} orig-locals-size
                   9473: <orig target>:
                   9474: @code{lp+!#} orig-locals-size @minus{} new-locals-size
                   9475: @end format
                   9476: The second @code{lp+!#} adjusts the locals stack pointer from the
                   9477: level at the @i{orig} point to the level after the @code{THEN}. The
                   9478: first @code{lp+!#} adjusts the locals stack pointer from the current
                   9479: level to the level at the orig point, so the complete effect is an
                   9480: adjustment from the current level to the right level after the
                   9481: @code{THEN}.
1.26      crook    9482: 
1.78      anton    9483: @cindex locals information on the control-flow stack
                   9484: @cindex control-flow stack items, locals information
                   9485: In a conventional Forth implementation a dest control-flow stack entry
                   9486: is just the target address and an orig entry is just the address to be
                   9487: patched. Our locals implementation adds a word list to every orig or dest
                   9488: item. It is the list of locals visible (or assumed visible) at the point
                   9489: described by the entry. Our implementation also adds a tag to identify
                   9490: the kind of entry, in particular to differentiate between live and dead
                   9491: (reachable and unreachable) orig entries.
1.26      crook    9492: 
1.78      anton    9493: A few unusual operations have to be performed on locals word lists:
1.44      crook    9494: 
1.5       anton    9495: 
1.78      anton    9496: doc-common-list
                   9497: doc-sub-list?
                   9498: doc-list-size
1.52      anton    9499: 
                   9500: 
1.78      anton    9501: Several features of our locals word list implementation make these
                   9502: operations easy to implement: The locals word lists are organised as
                   9503: linked lists; the tails of these lists are shared, if the lists
                   9504: contain some of the same locals; and the address of a name is greater
                   9505: than the address of the names behind it in the list.
1.5       anton    9506: 
1.78      anton    9507: Another important implementation detail is the variable
                   9508: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
                   9509: determine if they can be reached directly or only through the branch
                   9510: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
                   9511: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
                   9512: definition, by @code{BEGIN} and usually by @code{THEN}.
1.5       anton    9513: 
1.78      anton    9514: Counted loops are similar to other loops in most respects, but
                   9515: @code{LEAVE} requires special attention: It performs basically the same
                   9516: service as @code{AHEAD}, but it does not create a control-flow stack
                   9517: entry. Therefore the information has to be stored elsewhere;
                   9518: traditionally, the information was stored in the target fields of the
                   9519: branches created by the @code{LEAVE}s, by organizing these fields into a
                   9520: linked list. Unfortunately, this clever trick does not provide enough
                   9521: space for storing our extended control flow information. Therefore, we
                   9522: introduce another stack, the leave stack. It contains the control-flow
                   9523: stack entries for all unresolved @code{LEAVE}s.
1.44      crook    9524: 
1.78      anton    9525: Local names are kept until the end of the colon definition, even if
                   9526: they are no longer visible in any control-flow path. In a few cases
                   9527: this may lead to increased space needs for the locals name area, but
                   9528: usually less than reclaiming this space would cost in code size.
1.5       anton    9529: 
1.44      crook    9530: 
1.78      anton    9531: @node ANS Forth locals,  , Gforth locals, Locals
                   9532: @subsection ANS Forth locals
                   9533: @cindex locals, ANS Forth style
1.5       anton    9534: 
1.78      anton    9535: The ANS Forth locals wordset does not define a syntax for locals, but
                   9536: words that make it possible to define various syntaxes. One of the
                   9537: possible syntaxes is a subset of the syntax we used in the Gforth locals
                   9538: wordset, i.e.:
1.29      crook    9539: 
                   9540: @example
1.78      anton    9541: @{ local1 local2 ... -- comment @}
                   9542: @end example
                   9543: @noindent
                   9544: or
                   9545: @example
                   9546: @{ local1 local2 ... @}
1.29      crook    9547: @end example
                   9548: 
1.78      anton    9549: The order of the locals corresponds to the order in a stack comment. The
                   9550: restrictions are:
1.5       anton    9551: 
1.78      anton    9552: @itemize @bullet
                   9553: @item
                   9554: Locals can only be cell-sized values (no type specifiers are allowed).
                   9555: @item
                   9556: Locals can be defined only outside control structures.
                   9557: @item
                   9558: Locals can interfere with explicit usage of the return stack. For the
                   9559: exact (and long) rules, see the standard. If you don't use return stack
                   9560: accessing words in a definition using locals, you will be all right. The
                   9561: purpose of this rule is to make locals implementation on the return
                   9562: stack easier.
                   9563: @item
                   9564: The whole definition must be in one line.
                   9565: @end itemize
1.5       anton    9566: 
1.78      anton    9567: Locals defined in ANS Forth behave like @code{VALUE}s
                   9568: (@pxref{Values}). I.e., they are initialized from the stack. Using their
                   9569: name produces their value. Their value can be changed using @code{TO}.
1.77      anton    9570: 
1.78      anton    9571: Since the syntax above is supported by Gforth directly, you need not do
                   9572: anything to use it. If you want to port a program using this syntax to
                   9573: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
                   9574: syntax on the other system.
1.5       anton    9575: 
1.78      anton    9576: Note that a syntax shown in the standard, section A.13 looks
                   9577: similar, but is quite different in having the order of locals
                   9578: reversed. Beware!
1.5       anton    9579: 
1.78      anton    9580: The ANS Forth locals wordset itself consists of one word:
1.5       anton    9581: 
1.78      anton    9582: doc-(local)
1.5       anton    9583: 
1.78      anton    9584: The ANS Forth locals extension wordset defines a syntax using
                   9585: @code{locals|}, but it is so awful that we strongly recommend not to use
                   9586: it. We have implemented this syntax to make porting to Gforth easy, but
                   9587: do not document it here. The problem with this syntax is that the locals
                   9588: are defined in an order reversed with respect to the standard stack
                   9589: comment notation, making programs harder to read, and easier to misread
                   9590: and miswrite. The only merit of this syntax is that it is easy to
                   9591: implement using the ANS Forth locals wordset.
1.53      anton    9592: 
                   9593: 
1.78      anton    9594: @c ----------------------------------------------------------
                   9595: @node Structures, Object-oriented Forth, Locals, Words
                   9596: @section  Structures
                   9597: @cindex structures
                   9598: @cindex records
1.53      anton    9599: 
1.78      anton    9600: This section presents the structure package that comes with Gforth. A
                   9601: version of the package implemented in ANS Forth is available in
                   9602: @file{compat/struct.fs}. This package was inspired by a posting on
                   9603: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
                   9604: possibly John Hayes). A version of this section has been published in
                   9605: M. Anton Ertl,
                   9606: @uref{http://www.complang.tuwien.ac.at/forth/objects/structs.html, Yet
                   9607: Another Forth Structures Package}, Forth Dimensions 19(3), pages
                   9608: 13--16. Marcel Hendrix provided helpful comments.
1.53      anton    9609: 
1.78      anton    9610: @menu
                   9611: * Why explicit structure support?::  
                   9612: * Structure Usage::             
                   9613: * Structure Naming Convention::  
                   9614: * Structure Implementation::    
                   9615: * Structure Glossary::          
                   9616: @end menu
1.55      anton    9617: 
1.78      anton    9618: @node Why explicit structure support?, Structure Usage, Structures, Structures
                   9619: @subsection Why explicit structure support?
1.53      anton    9620: 
1.78      anton    9621: @cindex address arithmetic for structures
                   9622: @cindex structures using address arithmetic
                   9623: If we want to use a structure containing several fields, we could simply
                   9624: reserve memory for it, and access the fields using address arithmetic
                   9625: (@pxref{Address arithmetic}). As an example, consider a structure with
                   9626: the following fields
1.57      anton    9627: 
1.78      anton    9628: @table @code
                   9629: @item a
                   9630: is a float
                   9631: @item b
                   9632: is a cell
                   9633: @item c
                   9634: is a float
                   9635: @end table
1.57      anton    9636: 
1.78      anton    9637: Given the (float-aligned) base address of the structure we get the
                   9638: address of the field
1.52      anton    9639: 
1.78      anton    9640: @table @code
                   9641: @item a
                   9642: without doing anything further.
                   9643: @item b
                   9644: with @code{float+}
                   9645: @item c
                   9646: with @code{float+ cell+ faligned}
                   9647: @end table
1.52      anton    9648: 
1.78      anton    9649: It is easy to see that this can become quite tiring. 
1.52      anton    9650: 
1.78      anton    9651: Moreover, it is not very readable, because seeing a
                   9652: @code{cell+} tells us neither which kind of structure is
                   9653: accessed nor what field is accessed; we have to somehow infer the kind
                   9654: of structure, and then look up in the documentation, which field of
                   9655: that structure corresponds to that offset.
1.53      anton    9656: 
1.78      anton    9657: Finally, this kind of address arithmetic also causes maintenance
                   9658: troubles: If you add or delete a field somewhere in the middle of the
                   9659: structure, you have to find and change all computations for the fields
                   9660: afterwards.
1.52      anton    9661: 
1.78      anton    9662: So, instead of using @code{cell+} and friends directly, how
                   9663: about storing the offsets in constants:
1.52      anton    9664: 
1.78      anton    9665: @example
                   9666: 0 constant a-offset
                   9667: 0 float+ constant b-offset
                   9668: 0 float+ cell+ faligned c-offset
                   9669: @end example
1.64      pazsan   9670: 
1.78      anton    9671: Now we can get the address of field @code{x} with @code{x-offset
                   9672: +}. This is much better in all respects. Of course, you still
                   9673: have to change all later offset definitions if you add a field. You can
                   9674: fix this by declaring the offsets in the following way:
1.57      anton    9675: 
1.78      anton    9676: @example
                   9677: 0 constant a-offset
                   9678: a-offset float+ constant b-offset
                   9679: b-offset cell+ faligned constant c-offset
                   9680: @end example
1.57      anton    9681: 
1.78      anton    9682: Since we always use the offsets with @code{+}, we could use a defining
                   9683: word @code{cfield} that includes the @code{+} in the action of the
                   9684: defined word:
1.64      pazsan   9685: 
1.78      anton    9686: @example
                   9687: : cfield ( n "name" -- )
                   9688:     create ,
                   9689: does> ( name execution: addr1 -- addr2 )
                   9690:     @@ + ;
1.64      pazsan   9691: 
1.78      anton    9692: 0 cfield a
                   9693: 0 a float+ cfield b
                   9694: 0 b cell+ faligned cfield c
                   9695: @end example
1.64      pazsan   9696: 
1.78      anton    9697: Instead of @code{x-offset +}, we now simply write @code{x}.
1.64      pazsan   9698: 
1.78      anton    9699: The structure field words now can be used quite nicely. However,
                   9700: their definition is still a bit cumbersome: We have to repeat the
                   9701: name, the information about size and alignment is distributed before
                   9702: and after the field definitions etc.  The structure package presented
                   9703: here addresses these problems.
1.64      pazsan   9704: 
1.78      anton    9705: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
                   9706: @subsection Structure Usage
                   9707: @cindex structure usage
1.57      anton    9708: 
1.78      anton    9709: @cindex @code{field} usage
                   9710: @cindex @code{struct} usage
                   9711: @cindex @code{end-struct} usage
                   9712: You can define a structure for a (data-less) linked list with:
1.57      anton    9713: @example
1.78      anton    9714: struct
                   9715:     cell% field list-next
                   9716: end-struct list%
1.57      anton    9717: @end example
                   9718: 
1.78      anton    9719: With the address of the list node on the stack, you can compute the
                   9720: address of the field that contains the address of the next node with
                   9721: @code{list-next}. E.g., you can determine the length of a list
                   9722: with:
1.57      anton    9723: 
                   9724: @example
1.78      anton    9725: : list-length ( list -- n )
                   9726: \ "list" is a pointer to the first element of a linked list
                   9727: \ "n" is the length of the list
                   9728:     0 BEGIN ( list1 n1 )
                   9729:         over
                   9730:     WHILE ( list1 n1 )
                   9731:         1+ swap list-next @@ swap
                   9732:     REPEAT
                   9733:     nip ;
1.57      anton    9734: @end example
                   9735: 
1.78      anton    9736: You can reserve memory for a list node in the dictionary with
                   9737: @code{list% %allot}, which leaves the address of the list node on the
                   9738: stack. For the equivalent allocation on the heap you can use @code{list%
                   9739: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
                   9740: use @code{list% %allocate}). You can get the the size of a list
                   9741: node with @code{list% %size} and its alignment with @code{list%
                   9742: %alignment}.
                   9743: 
                   9744: Note that in ANS Forth the body of a @code{create}d word is
                   9745: @code{aligned} but not necessarily @code{faligned};
                   9746: therefore, if you do a:
1.57      anton    9747: 
                   9748: @example
1.78      anton    9749: create @emph{name} foo% %allot drop
1.57      anton    9750: @end example
                   9751: 
1.78      anton    9752: @noindent
                   9753: then the memory alloted for @code{foo%} is guaranteed to start at the
                   9754: body of @code{@emph{name}} only if @code{foo%} contains only character,
                   9755: cell and double fields.  Therefore, if your structure contains floats,
                   9756: better use
1.57      anton    9757: 
                   9758: @example
1.78      anton    9759: foo% %allot constant @emph{name}
1.57      anton    9760: @end example
                   9761: 
1.78      anton    9762: @cindex structures containing structures
                   9763: You can include a structure @code{foo%} as a field of
                   9764: another structure, like this:
1.65      anton    9765: @example
1.78      anton    9766: struct
                   9767: ...
                   9768:     foo% field ...
                   9769: ...
                   9770: end-struct ...
1.65      anton    9771: @end example
1.52      anton    9772: 
1.78      anton    9773: @cindex structure extension
                   9774: @cindex extended records
                   9775: Instead of starting with an empty structure, you can extend an
                   9776: existing structure. E.g., a plain linked list without data, as defined
                   9777: above, is hardly useful; You can extend it to a linked list of integers,
                   9778: like this:@footnote{This feature is also known as @emph{extended
                   9779: records}. It is the main innovation in the Oberon language; in other
                   9780: words, adding this feature to Modula-2 led Wirth to create a new
                   9781: language, write a new compiler etc.  Adding this feature to Forth just
                   9782: required a few lines of code.}
1.52      anton    9783: 
1.78      anton    9784: @example
                   9785: list%
                   9786:     cell% field intlist-int
                   9787: end-struct intlist%
                   9788: @end example
1.55      anton    9789: 
1.78      anton    9790: @code{intlist%} is a structure with two fields:
                   9791: @code{list-next} and @code{intlist-int}.
1.55      anton    9792: 
1.78      anton    9793: @cindex structures containing arrays
                   9794: You can specify an array type containing @emph{n} elements of
                   9795: type @code{foo%} like this:
1.55      anton    9796: 
                   9797: @example
1.78      anton    9798: foo% @emph{n} *
1.56      anton    9799: @end example
1.55      anton    9800: 
1.78      anton    9801: You can use this array type in any place where you can use a normal
                   9802: type, e.g., when defining a @code{field}, or with
                   9803: @code{%allot}.
                   9804: 
                   9805: @cindex first field optimization
                   9806: The first field is at the base address of a structure and the word for
                   9807: this field (e.g., @code{list-next}) actually does not change the address
                   9808: on the stack. You may be tempted to leave it away in the interest of
                   9809: run-time and space efficiency. This is not necessary, because the
                   9810: structure package optimizes this case: If you compile a first-field
                   9811: words, no code is generated. So, in the interest of readability and
                   9812: maintainability you should include the word for the field when accessing
                   9813: the field.
1.52      anton    9814: 
                   9815: 
1.78      anton    9816: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
                   9817: @subsection Structure Naming Convention
                   9818: @cindex structure naming convention
1.52      anton    9819: 
1.78      anton    9820: The field names that come to (my) mind are often quite generic, and,
                   9821: if used, would cause frequent name clashes. E.g., many structures
                   9822: probably contain a @code{counter} field. The structure names
                   9823: that come to (my) mind are often also the logical choice for the names
                   9824: of words that create such a structure.
1.52      anton    9825: 
1.78      anton    9826: Therefore, I have adopted the following naming conventions: 
1.52      anton    9827: 
1.78      anton    9828: @itemize @bullet
                   9829: @cindex field naming convention
                   9830: @item
                   9831: The names of fields are of the form
                   9832: @code{@emph{struct}-@emph{field}}, where
                   9833: @code{@emph{struct}} is the basic name of the structure, and
                   9834: @code{@emph{field}} is the basic name of the field. You can
                   9835: think of field words as converting the (address of the)
                   9836: structure into the (address of the) field.
1.52      anton    9837: 
1.78      anton    9838: @cindex structure naming convention
                   9839: @item
                   9840: The names of structures are of the form
                   9841: @code{@emph{struct}%}, where
                   9842: @code{@emph{struct}} is the basic name of the structure.
                   9843: @end itemize
1.52      anton    9844: 
1.78      anton    9845: This naming convention does not work that well for fields of extended
                   9846: structures; e.g., the integer list structure has a field
                   9847: @code{intlist-int}, but has @code{list-next}, not
                   9848: @code{intlist-next}.
1.53      anton    9849: 
1.78      anton    9850: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
                   9851: @subsection Structure Implementation
                   9852: @cindex structure implementation
                   9853: @cindex implementation of structures
1.52      anton    9854: 
1.78      anton    9855: The central idea in the implementation is to pass the data about the
                   9856: structure being built on the stack, not in some global
                   9857: variable. Everything else falls into place naturally once this design
                   9858: decision is made.
1.53      anton    9859: 
1.78      anton    9860: The type description on the stack is of the form @emph{align
                   9861: size}. Keeping the size on the top-of-stack makes dealing with arrays
                   9862: very simple.
1.53      anton    9863: 
1.78      anton    9864: @code{field} is a defining word that uses @code{Create}
                   9865: and @code{DOES>}. The body of the field contains the offset
                   9866: of the field, and the normal @code{DOES>} action is simply:
1.53      anton    9867: 
                   9868: @example
1.78      anton    9869: @@ +
1.53      anton    9870: @end example
                   9871: 
1.78      anton    9872: @noindent
                   9873: i.e., add the offset to the address, giving the stack effect
                   9874: @i{addr1 -- addr2} for a field.
                   9875: 
                   9876: @cindex first field optimization, implementation
                   9877: This simple structure is slightly complicated by the optimization
                   9878: for fields with offset 0, which requires a different
                   9879: @code{DOES>}-part (because we cannot rely on there being
                   9880: something on the stack if such a field is invoked during
                   9881: compilation). Therefore, we put the different @code{DOES>}-parts
                   9882: in separate words, and decide which one to invoke based on the
                   9883: offset. For a zero offset, the field is basically a noop; it is
                   9884: immediate, and therefore no code is generated when it is compiled.
1.53      anton    9885: 
1.78      anton    9886: @node Structure Glossary,  , Structure Implementation, Structures
                   9887: @subsection Structure Glossary
                   9888: @cindex structure glossary
1.53      anton    9889: 
1.5       anton    9890: 
1.78      anton    9891: doc-%align
                   9892: doc-%alignment
                   9893: doc-%alloc
                   9894: doc-%allocate
                   9895: doc-%allot
                   9896: doc-cell%
                   9897: doc-char%
                   9898: doc-dfloat%
                   9899: doc-double%
                   9900: doc-end-struct
                   9901: doc-field
                   9902: doc-float%
                   9903: doc-naligned
                   9904: doc-sfloat%
                   9905: doc-%size
                   9906: doc-struct
1.54      anton    9907: 
                   9908: 
1.26      crook    9909: @c -------------------------------------------------------------
1.78      anton    9910: @node Object-oriented Forth, Programming Tools, Structures, Words
                   9911: @section Object-oriented Forth
                   9912: 
                   9913: Gforth comes with three packages for object-oriented programming:
                   9914: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
                   9915: is preloaded, so you have to @code{include} them before use. The most
                   9916: important differences between these packages (and others) are discussed
                   9917: in @ref{Comparison with other object models}. All packages are written
                   9918: in ANS Forth and can be used with any other ANS Forth.
1.5       anton    9919: 
1.78      anton    9920: @menu
                   9921: * Why object-oriented programming?::  
                   9922: * Object-Oriented Terminology::  
                   9923: * Objects::                     
                   9924: * OOF::                         
                   9925: * Mini-OOF::                    
                   9926: * Comparison with other object models::  
                   9927: @end menu
1.5       anton    9928: 
1.78      anton    9929: @c ----------------------------------------------------------------
                   9930: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
                   9931: @subsection Why object-oriented programming?
                   9932: @cindex object-oriented programming motivation
                   9933: @cindex motivation for object-oriented programming
1.44      crook    9934: 
1.78      anton    9935: Often we have to deal with several data structures (@emph{objects}),
                   9936: that have to be treated similarly in some respects, but differently in
                   9937: others. Graphical objects are the textbook example: circles, triangles,
                   9938: dinosaurs, icons, and others, and we may want to add more during program
                   9939: development. We want to apply some operations to any graphical object,
                   9940: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
                   9941: has to do something different for every kind of object.
                   9942: @comment TODO add some other operations eg perimeter, area
                   9943: @comment and tie in to concrete examples later..
1.5       anton    9944: 
1.78      anton    9945: We could implement @code{draw} as a big @code{CASE}
                   9946: control structure that executes the appropriate code depending on the
                   9947: kind of object to be drawn. This would be not be very elegant, and,
                   9948: moreover, we would have to change @code{draw} every time we add
                   9949: a new kind of graphical object (say, a spaceship).
1.44      crook    9950: 
1.78      anton    9951: What we would rather do is: When defining spaceships, we would tell
                   9952: the system: ``Here's how you @code{draw} a spaceship; you figure
                   9953: out the rest''.
1.5       anton    9954: 
1.78      anton    9955: This is the problem that all systems solve that (rightfully) call
                   9956: themselves object-oriented; the object-oriented packages presented here
                   9957: solve this problem (and not much else).
                   9958: @comment TODO ?list properties of oo systems.. oo vs o-based?
1.44      crook    9959: 
1.78      anton    9960: @c ------------------------------------------------------------------------
                   9961: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
                   9962: @subsection Object-Oriented Terminology
                   9963: @cindex object-oriented terminology
                   9964: @cindex terminology for object-oriented programming
1.5       anton    9965: 
1.78      anton    9966: This section is mainly for reference, so you don't have to understand
                   9967: all of it right away.  The terminology is mainly Smalltalk-inspired.  In
                   9968: short:
1.44      crook    9969: 
1.78      anton    9970: @table @emph
                   9971: @cindex class
                   9972: @item class
                   9973: a data structure definition with some extras.
1.5       anton    9974: 
1.78      anton    9975: @cindex object
                   9976: @item object
                   9977: an instance of the data structure described by the class definition.
1.5       anton    9978: 
1.78      anton    9979: @cindex instance variables
                   9980: @item instance variables
                   9981: fields of the data structure.
1.5       anton    9982: 
1.78      anton    9983: @cindex selector
                   9984: @cindex method selector
                   9985: @cindex virtual function
                   9986: @item selector
                   9987: (or @emph{method selector}) a word (e.g.,
                   9988: @code{draw}) that performs an operation on a variety of data
                   9989: structures (classes). A selector describes @emph{what} operation to
                   9990: perform. In C++ terminology: a (pure) virtual function.
1.5       anton    9991: 
1.78      anton    9992: @cindex method
                   9993: @item method
                   9994: the concrete definition that performs the operation
                   9995: described by the selector for a specific class. A method specifies
                   9996: @emph{how} the operation is performed for a specific class.
1.5       anton    9997: 
1.78      anton    9998: @cindex selector invocation
                   9999: @cindex message send
                   10000: @cindex invoking a selector
                   10001: @item selector invocation
                   10002: a call of a selector. One argument of the call (the TOS (top-of-stack))
                   10003: is used for determining which method is used. In Smalltalk terminology:
                   10004: a message (consisting of the selector and the other arguments) is sent
                   10005: to the object.
1.5       anton    10006: 
1.78      anton    10007: @cindex receiving object
                   10008: @item receiving object
                   10009: the object used for determining the method executed by a selector
                   10010: invocation. In the @file{objects.fs} model, it is the object that is on
                   10011: the TOS when the selector is invoked. (@emph{Receiving} comes from
                   10012: the Smalltalk @emph{message} terminology.)
1.5       anton    10013: 
1.78      anton    10014: @cindex child class
                   10015: @cindex parent class
                   10016: @cindex inheritance
                   10017: @item child class
                   10018: a class that has (@emph{inherits}) all properties (instance variables,
                   10019: selectors, methods) from a @emph{parent class}. In Smalltalk
                   10020: terminology: The subclass inherits from the superclass. In C++
                   10021: terminology: The derived class inherits from the base class.
1.5       anton    10022: 
1.78      anton    10023: @end table
1.5       anton    10024: 
1.78      anton    10025: @c If you wonder about the message sending terminology, it comes from
                   10026: @c a time when each object had it's own task and objects communicated via
                   10027: @c message passing; eventually the Smalltalk developers realized that
                   10028: @c they can do most things through simple (indirect) calls. They kept the
                   10029: @c terminology.
1.5       anton    10030: 
1.78      anton    10031: @c --------------------------------------------------------------
                   10032: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
                   10033: @subsection The @file{objects.fs} model
                   10034: @cindex objects
                   10035: @cindex object-oriented programming
1.26      crook    10036: 
1.78      anton    10037: @cindex @file{objects.fs}
                   10038: @cindex @file{oof.fs}
1.26      crook    10039: 
1.78      anton    10040: This section describes the @file{objects.fs} package. This material also
                   10041: has been published in M. Anton Ertl,
                   10042: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
                   10043: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
                   10044: 37--43.
                   10045: @c McKewan's and Zsoter's packages
1.26      crook    10046: 
1.78      anton    10047: This section assumes that you have read @ref{Structures}.
1.5       anton    10048: 
1.78      anton    10049: The techniques on which this model is based have been used to implement
                   10050: the parser generator, Gray, and have also been used in Gforth for
                   10051: implementing the various flavours of word lists (hashed or not,
                   10052: case-sensitive or not, special-purpose word lists for locals etc.).
1.5       anton    10053: 
                   10054: 
1.26      crook    10055: @menu
1.78      anton    10056: * Properties of the Objects model::  
                   10057: * Basic Objects Usage::         
                   10058: * The Objects base class::      
                   10059: * Creating objects::            
                   10060: * Object-Oriented Programming Style::  
                   10061: * Class Binding::               
                   10062: * Method conveniences::         
                   10063: * Classes and Scoping::         
                   10064: * Dividing classes::            
                   10065: * Object Interfaces::           
                   10066: * Objects Implementation::      
                   10067: * Objects Glossary::            
1.26      crook    10068: @end menu
1.5       anton    10069: 
1.78      anton    10070: Marcel Hendrix provided helpful comments on this section.
1.5       anton    10071: 
1.78      anton    10072: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
                   10073: @subsubsection Properties of the @file{objects.fs} model
                   10074: @cindex @file{objects.fs} properties
1.5       anton    10075: 
1.78      anton    10076: @itemize @bullet
                   10077: @item
                   10078: It is straightforward to pass objects on the stack. Passing
                   10079: selectors on the stack is a little less convenient, but possible.
1.44      crook    10080: 
1.78      anton    10081: @item
                   10082: Objects are just data structures in memory, and are referenced by their
                   10083: address. You can create words for objects with normal defining words
                   10084: like @code{constant}. Likewise, there is no difference between instance
                   10085: variables that contain objects and those that contain other data.
1.5       anton    10086: 
1.78      anton    10087: @item
                   10088: Late binding is efficient and easy to use.
1.44      crook    10089: 
1.78      anton    10090: @item
                   10091: It avoids parsing, and thus avoids problems with state-smartness
                   10092: and reduced extensibility; for convenience there are a few parsing
                   10093: words, but they have non-parsing counterparts. There are also a few
                   10094: defining words that parse. This is hard to avoid, because all standard
                   10095: defining words parse (except @code{:noname}); however, such
                   10096: words are not as bad as many other parsing words, because they are not
                   10097: state-smart.
1.5       anton    10098: 
1.78      anton    10099: @item
                   10100: It does not try to incorporate everything. It does a few things and does
                   10101: them well (IMO). In particular, this model was not designed to support
                   10102: information hiding (although it has features that may help); you can use
                   10103: a separate package for achieving this.
1.5       anton    10104: 
1.78      anton    10105: @item
                   10106: It is layered; you don't have to learn and use all features to use this
                   10107: model. Only a few features are necessary (@pxref{Basic Objects Usage},
                   10108: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
                   10109: are optional and independent of each other.
1.5       anton    10110: 
1.78      anton    10111: @item
                   10112: An implementation in ANS Forth is available.
1.5       anton    10113: 
1.78      anton    10114: @end itemize
1.5       anton    10115: 
1.44      crook    10116: 
1.78      anton    10117: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
                   10118: @subsubsection Basic @file{objects.fs} Usage
                   10119: @cindex basic objects usage
                   10120: @cindex objects, basic usage
1.5       anton    10121: 
1.78      anton    10122: You can define a class for graphical objects like this:
1.44      crook    10123: 
1.78      anton    10124: @cindex @code{class} usage
                   10125: @cindex @code{end-class} usage
                   10126: @cindex @code{selector} usage
1.5       anton    10127: @example
1.78      anton    10128: object class \ "object" is the parent class
                   10129:   selector draw ( x y graphical -- )
                   10130: end-class graphical
                   10131: @end example
                   10132: 
                   10133: This code defines a class @code{graphical} with an
                   10134: operation @code{draw}.  We can perform the operation
                   10135: @code{draw} on any @code{graphical} object, e.g.:
                   10136: 
                   10137: @example
                   10138: 100 100 t-rex draw
1.26      crook    10139: @end example
1.5       anton    10140: 
1.78      anton    10141: @noindent
                   10142: where @code{t-rex} is a word (say, a constant) that produces a
                   10143: graphical object.
                   10144: 
                   10145: @comment TODO add a 2nd operation eg perimeter.. and use for
                   10146: @comment a concrete example
1.5       anton    10147: 
1.78      anton    10148: @cindex abstract class
                   10149: How do we create a graphical object? With the present definitions,
                   10150: we cannot create a useful graphical object. The class
                   10151: @code{graphical} describes graphical objects in general, but not
                   10152: any concrete graphical object type (C++ users would call it an
                   10153: @emph{abstract class}); e.g., there is no method for the selector
                   10154: @code{draw} in the class @code{graphical}.
1.5       anton    10155: 
1.78      anton    10156: For concrete graphical objects, we define child classes of the
                   10157: class @code{graphical}, e.g.:
1.5       anton    10158: 
1.78      anton    10159: @cindex @code{overrides} usage
                   10160: @cindex @code{field} usage in class definition
1.26      crook    10161: @example
1.78      anton    10162: graphical class \ "graphical" is the parent class
                   10163:   cell% field circle-radius
1.5       anton    10164: 
1.78      anton    10165: :noname ( x y circle -- )
                   10166:   circle-radius @@ draw-circle ;
                   10167: overrides draw
1.5       anton    10168: 
1.78      anton    10169: :noname ( n-radius circle -- )
                   10170:   circle-radius ! ;
                   10171: overrides construct
1.5       anton    10172: 
1.78      anton    10173: end-class circle
                   10174: @end example
1.44      crook    10175: 
1.78      anton    10176: Here we define a class @code{circle} as a child of @code{graphical},
                   10177: with field @code{circle-radius} (which behaves just like a field
                   10178: (@pxref{Structures}); it defines (using @code{overrides}) new methods
                   10179: for the selectors @code{draw} and @code{construct} (@code{construct} is
                   10180: defined in @code{object}, the parent class of @code{graphical}).
1.5       anton    10181: 
1.78      anton    10182: Now we can create a circle on the heap (i.e.,
                   10183: @code{allocate}d memory) with:
1.44      crook    10184: 
1.78      anton    10185: @cindex @code{heap-new} usage
1.5       anton    10186: @example
1.78      anton    10187: 50 circle heap-new constant my-circle
1.5       anton    10188: @end example
                   10189: 
1.78      anton    10190: @noindent
                   10191: @code{heap-new} invokes @code{construct}, thus
                   10192: initializing the field @code{circle-radius} with 50. We can draw
                   10193: this new circle at (100,100) with:
1.5       anton    10194: 
                   10195: @example
1.78      anton    10196: 100 100 my-circle draw
1.5       anton    10197: @end example
                   10198: 
1.78      anton    10199: @cindex selector invocation, restrictions
                   10200: @cindex class definition, restrictions
                   10201: Note: You can only invoke a selector if the object on the TOS
                   10202: (the receiving object) belongs to the class where the selector was
                   10203: defined or one of its descendents; e.g., you can invoke
                   10204: @code{draw} only for objects belonging to @code{graphical}
                   10205: or its descendents (e.g., @code{circle}).  Immediately before
                   10206: @code{end-class}, the search order has to be the same as
                   10207: immediately after @code{class}.
                   10208: 
                   10209: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
                   10210: @subsubsection The @file{object.fs} base class
                   10211: @cindex @code{object} class
                   10212: 
                   10213: When you define a class, you have to specify a parent class.  So how do
                   10214: you start defining classes? There is one class available from the start:
                   10215: @code{object}. It is ancestor for all classes and so is the
                   10216: only class that has no parent. It has two selectors: @code{construct}
                   10217: and @code{print}.
                   10218: 
                   10219: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
                   10220: @subsubsection Creating objects
                   10221: @cindex creating objects
                   10222: @cindex object creation
                   10223: @cindex object allocation options
                   10224: 
                   10225: @cindex @code{heap-new} discussion
                   10226: @cindex @code{dict-new} discussion
                   10227: @cindex @code{construct} discussion
                   10228: You can create and initialize an object of a class on the heap with
                   10229: @code{heap-new} ( ... class -- object ) and in the dictionary
                   10230: (allocation with @code{allot}) with @code{dict-new} (
                   10231: ... class -- object ). Both words invoke @code{construct}, which
                   10232: consumes the stack items indicated by "..." above.
                   10233: 
                   10234: @cindex @code{init-object} discussion
                   10235: @cindex @code{class-inst-size} discussion
                   10236: If you want to allocate memory for an object yourself, you can get its
                   10237: alignment and size with @code{class-inst-size 2@@} ( class --
                   10238: align size ). Once you have memory for an object, you can initialize
                   10239: it with @code{init-object} ( ... class object -- );
                   10240: @code{construct} does only a part of the necessary work.
                   10241: 
                   10242: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
                   10243: @subsubsection Object-Oriented Programming Style
                   10244: @cindex object-oriented programming style
                   10245: @cindex programming style, object-oriented
1.5       anton    10246: 
1.78      anton    10247: This section is not exhaustive.
1.5       anton    10248: 
1.78      anton    10249: @cindex stack effects of selectors
                   10250: @cindex selectors and stack effects
                   10251: In general, it is a good idea to ensure that all methods for the
                   10252: same selector have the same stack effect: when you invoke a selector,
                   10253: you often have no idea which method will be invoked, so, unless all
                   10254: methods have the same stack effect, you will not know the stack effect
                   10255: of the selector invocation.
1.5       anton    10256: 
1.78      anton    10257: One exception to this rule is methods for the selector
                   10258: @code{construct}. We know which method is invoked, because we
                   10259: specify the class to be constructed at the same place. Actually, I
                   10260: defined @code{construct} as a selector only to give the users a
                   10261: convenient way to specify initialization. The way it is used, a
                   10262: mechanism different from selector invocation would be more natural
                   10263: (but probably would take more code and more space to explain).
1.5       anton    10264: 
1.78      anton    10265: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
                   10266: @subsubsection Class Binding
                   10267: @cindex class binding
                   10268: @cindex early binding
1.5       anton    10269: 
1.78      anton    10270: @cindex late binding
                   10271: Normal selector invocations determine the method at run-time depending
                   10272: on the class of the receiving object. This run-time selection is called
                   10273: @i{late binding}.
1.5       anton    10274: 
1.78      anton    10275: Sometimes it's preferable to invoke a different method. For example,
                   10276: you might want to use the simple method for @code{print}ing
                   10277: @code{object}s instead of the possibly long-winded @code{print} method
                   10278: of the receiver class. You can achieve this by replacing the invocation
                   10279: of @code{print} with:
1.5       anton    10280: 
1.78      anton    10281: @cindex @code{[bind]} usage
1.5       anton    10282: @example
1.78      anton    10283: [bind] object print
1.5       anton    10284: @end example
                   10285: 
1.78      anton    10286: @noindent
                   10287: in compiled code or:
                   10288: 
                   10289: @cindex @code{bind} usage
1.5       anton    10290: @example
1.78      anton    10291: bind object print
1.5       anton    10292: @end example
                   10293: 
1.78      anton    10294: @cindex class binding, alternative to
                   10295: @noindent
                   10296: in interpreted code. Alternatively, you can define the method with a
                   10297: name (e.g., @code{print-object}), and then invoke it through the
                   10298: name. Class binding is just a (often more convenient) way to achieve
                   10299: the same effect; it avoids name clutter and allows you to invoke
                   10300: methods directly without naming them first.
1.5       anton    10301: 
1.78      anton    10302: @cindex superclass binding
                   10303: @cindex parent class binding
                   10304: A frequent use of class binding is this: When we define a method
                   10305: for a selector, we often want the method to do what the selector does
                   10306: in the parent class, and a little more. There is a special word for
                   10307: this purpose: @code{[parent]}; @code{[parent]
                   10308: @emph{selector}} is equivalent to @code{[bind] @emph{parent
                   10309: selector}}, where @code{@emph{parent}} is the parent
                   10310: class of the current class. E.g., a method definition might look like:
1.44      crook    10311: 
1.78      anton    10312: @cindex @code{[parent]} usage
                   10313: @example
                   10314: :noname
                   10315:   dup [parent] foo \ do parent's foo on the receiving object
                   10316:   ... \ do some more
                   10317: ; overrides foo
                   10318: @end example
1.6       pazsan   10319: 
1.78      anton    10320: @cindex class binding as optimization
                   10321: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
                   10322: March 1997), Andrew McKewan presents class binding as an optimization
                   10323: technique. I recommend not using it for this purpose unless you are in
                   10324: an emergency. Late binding is pretty fast with this model anyway, so the
                   10325: benefit of using class binding is small; the cost of using class binding
                   10326: where it is not appropriate is reduced maintainability.
1.44      crook    10327: 
1.78      anton    10328: While we are at programming style questions: You should bind
                   10329: selectors only to ancestor classes of the receiving object. E.g., say,
                   10330: you know that the receiving object is of class @code{foo} or its
                   10331: descendents; then you should bind only to @code{foo} and its
                   10332: ancestors.
1.12      anton    10333: 
1.78      anton    10334: @node Method conveniences, Classes and Scoping, Class Binding, Objects
                   10335: @subsubsection Method conveniences
                   10336: @cindex method conveniences
1.44      crook    10337: 
1.78      anton    10338: In a method you usually access the receiving object pretty often.  If
                   10339: you define the method as a plain colon definition (e.g., with
                   10340: @code{:noname}), you may have to do a lot of stack
                   10341: gymnastics. To avoid this, you can define the method with @code{m:
                   10342: ... ;m}. E.g., you could define the method for
                   10343: @code{draw}ing a @code{circle} with
1.6       pazsan   10344: 
1.78      anton    10345: @cindex @code{this} usage
                   10346: @cindex @code{m:} usage
                   10347: @cindex @code{;m} usage
                   10348: @example
                   10349: m: ( x y circle -- )
                   10350:   ( x y ) this circle-radius @@ draw-circle ;m
                   10351: @end example
1.6       pazsan   10352: 
1.78      anton    10353: @cindex @code{exit} in @code{m: ... ;m}
                   10354: @cindex @code{exitm} discussion
                   10355: @cindex @code{catch} in @code{m: ... ;m}
                   10356: When this method is executed, the receiver object is removed from the
                   10357: stack; you can access it with @code{this} (admittedly, in this
                   10358: example the use of @code{m: ... ;m} offers no advantage). Note
                   10359: that I specify the stack effect for the whole method (i.e. including
                   10360: the receiver object), not just for the code between @code{m:}
                   10361: and @code{;m}. You cannot use @code{exit} in
                   10362: @code{m:...;m}; instead, use
                   10363: @code{exitm}.@footnote{Moreover, for any word that calls
                   10364: @code{catch} and was defined before loading
                   10365: @code{objects.fs}, you have to redefine it like I redefined
                   10366: @code{catch}: @code{: catch this >r catch r> to-this ;}}
1.12      anton    10367: 
1.78      anton    10368: @cindex @code{inst-var} usage
                   10369: You will frequently use sequences of the form @code{this
                   10370: @emph{field}} (in the example above: @code{this
                   10371: circle-radius}). If you use the field only in this way, you can
                   10372: define it with @code{inst-var} and eliminate the
                   10373: @code{this} before the field name. E.g., the @code{circle}
                   10374: class above could also be defined with:
1.6       pazsan   10375: 
1.78      anton    10376: @example
                   10377: graphical class
                   10378:   cell% inst-var radius
1.6       pazsan   10379: 
1.78      anton    10380: m: ( x y circle -- )
                   10381:   radius @@ draw-circle ;m
                   10382: overrides draw
1.6       pazsan   10383: 
1.78      anton    10384: m: ( n-radius circle -- )
                   10385:   radius ! ;m
                   10386: overrides construct
1.6       pazsan   10387: 
1.78      anton    10388: end-class circle
                   10389: @end example
1.6       pazsan   10390: 
1.78      anton    10391: @code{radius} can only be used in @code{circle} and its
                   10392: descendent classes and inside @code{m:...;m}.
1.6       pazsan   10393: 
1.78      anton    10394: @cindex @code{inst-value} usage
                   10395: You can also define fields with @code{inst-value}, which is
                   10396: to @code{inst-var} what @code{value} is to
                   10397: @code{variable}.  You can change the value of such a field with
                   10398: @code{[to-inst]}.  E.g., we could also define the class
                   10399: @code{circle} like this:
1.44      crook    10400: 
1.78      anton    10401: @example
                   10402: graphical class
                   10403:   inst-value radius
1.6       pazsan   10404: 
1.78      anton    10405: m: ( x y circle -- )
                   10406:   radius draw-circle ;m
                   10407: overrides draw
1.44      crook    10408: 
1.78      anton    10409: m: ( n-radius circle -- )
                   10410:   [to-inst] radius ;m
                   10411: overrides construct
1.6       pazsan   10412: 
1.78      anton    10413: end-class circle
                   10414: @end example
1.6       pazsan   10415: 
1.78      anton    10416: @c !! :m is easy to confuse with m:.  Another name would be better.
1.6       pazsan   10417: 
1.78      anton    10418: @c Finally, you can define named methods with @code{:m}.  One use of this
                   10419: @c feature is the definition of words that occur only in one class and are
                   10420: @c not intended to be overridden, but which still need method context
                   10421: @c (e.g., for accessing @code{inst-var}s).  Another use is for methods that
                   10422: @c would be bound frequently, if defined anonymously.
1.6       pazsan   10423: 
                   10424: 
1.78      anton    10425: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
                   10426: @subsubsection Classes and Scoping
                   10427: @cindex classes and scoping
                   10428: @cindex scoping and classes
1.6       pazsan   10429: 
1.78      anton    10430: Inheritance is frequent, unlike structure extension. This exacerbates
                   10431: the problem with the field name convention (@pxref{Structure Naming
                   10432: Convention}): One always has to remember in which class the field was
                   10433: originally defined; changing a part of the class structure would require
                   10434: changes for renaming in otherwise unaffected code.
1.6       pazsan   10435: 
1.78      anton    10436: @cindex @code{inst-var} visibility
                   10437: @cindex @code{inst-value} visibility
                   10438: To solve this problem, I added a scoping mechanism (which was not in my
                   10439: original charter): A field defined with @code{inst-var} (or
                   10440: @code{inst-value}) is visible only in the class where it is defined and in
                   10441: the descendent classes of this class.  Using such fields only makes
                   10442: sense in @code{m:}-defined methods in these classes anyway.
1.6       pazsan   10443: 
1.78      anton    10444: This scoping mechanism allows us to use the unadorned field name,
                   10445: because name clashes with unrelated words become much less likely.
1.6       pazsan   10446: 
1.78      anton    10447: @cindex @code{protected} discussion
                   10448: @cindex @code{private} discussion
                   10449: Once we have this mechanism, we can also use it for controlling the
                   10450: visibility of other words: All words defined after
                   10451: @code{protected} are visible only in the current class and its
                   10452: descendents. @code{public} restores the compilation
                   10453: (i.e. @code{current}) word list that was in effect before. If you
                   10454: have several @code{protected}s without an intervening
                   10455: @code{public} or @code{set-current}, @code{public}
                   10456: will restore the compilation word list in effect before the first of
                   10457: these @code{protected}s.
1.6       pazsan   10458: 
1.78      anton    10459: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
                   10460: @subsubsection Dividing classes
                   10461: @cindex Dividing classes
                   10462: @cindex @code{methods}...@code{end-methods}
1.6       pazsan   10463: 
1.78      anton    10464: You may want to do the definition of methods separate from the
                   10465: definition of the class, its selectors, fields, and instance variables,
                   10466: i.e., separate the implementation from the definition.  You can do this
                   10467: in the following way:
1.6       pazsan   10468: 
1.78      anton    10469: @example
                   10470: graphical class
                   10471:   inst-value radius
                   10472: end-class circle
1.6       pazsan   10473: 
1.78      anton    10474: ... \ do some other stuff
1.6       pazsan   10475: 
1.78      anton    10476: circle methods \ now we are ready
1.44      crook    10477: 
1.78      anton    10478: m: ( x y circle -- )
                   10479:   radius draw-circle ;m
                   10480: overrides draw
1.6       pazsan   10481: 
1.78      anton    10482: m: ( n-radius circle -- )
                   10483:   [to-inst] radius ;m
                   10484: overrides construct
1.44      crook    10485: 
1.78      anton    10486: end-methods
                   10487: @end example
1.7       pazsan   10488: 
1.78      anton    10489: You can use several @code{methods}...@code{end-methods} sections.  The
                   10490: only things you can do to the class in these sections are: defining
                   10491: methods, and overriding the class's selectors.  You must not define new
                   10492: selectors or fields.
1.7       pazsan   10493: 
1.78      anton    10494: Note that you often have to override a selector before using it.  In
                   10495: particular, you usually have to override @code{construct} with a new
                   10496: method before you can invoke @code{heap-new} and friends.  E.g., you
                   10497: must not create a circle before the @code{overrides construct} sequence
                   10498: in the example above.
1.7       pazsan   10499: 
1.78      anton    10500: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
                   10501: @subsubsection Object Interfaces
                   10502: @cindex object interfaces
                   10503: @cindex interfaces for objects
1.7       pazsan   10504: 
1.78      anton    10505: In this model you can only call selectors defined in the class of the
                   10506: receiving objects or in one of its ancestors. If you call a selector
                   10507: with a receiving object that is not in one of these classes, the
                   10508: result is undefined; if you are lucky, the program crashes
                   10509: immediately.
1.7       pazsan   10510: 
1.78      anton    10511: @cindex selectors common to hardly-related classes
                   10512: Now consider the case when you want to have a selector (or several)
                   10513: available in two classes: You would have to add the selector to a
                   10514: common ancestor class, in the worst case to @code{object}. You
                   10515: may not want to do this, e.g., because someone else is responsible for
                   10516: this ancestor class.
1.7       pazsan   10517: 
1.78      anton    10518: The solution for this problem is interfaces. An interface is a
                   10519: collection of selectors. If a class implements an interface, the
                   10520: selectors become available to the class and its descendents. A class
                   10521: can implement an unlimited number of interfaces. For the problem
                   10522: discussed above, we would define an interface for the selector(s), and
                   10523: both classes would implement the interface.
1.7       pazsan   10524: 
1.78      anton    10525: As an example, consider an interface @code{storage} for
                   10526: writing objects to disk and getting them back, and a class
                   10527: @code{foo} that implements it. The code would look like this:
1.7       pazsan   10528: 
1.78      anton    10529: @cindex @code{interface} usage
                   10530: @cindex @code{end-interface} usage
                   10531: @cindex @code{implementation} usage
                   10532: @example
                   10533: interface
                   10534:   selector write ( file object -- )
                   10535:   selector read1 ( file object -- )
                   10536: end-interface storage
1.13      pazsan   10537: 
1.78      anton    10538: bar class
                   10539:   storage implementation
1.13      pazsan   10540: 
1.78      anton    10541: ... overrides write
                   10542: ... overrides read1
                   10543: ...
                   10544: end-class foo
                   10545: @end example
1.13      pazsan   10546: 
1.78      anton    10547: @noindent
                   10548: (I would add a word @code{read} @i{( file -- object )} that uses
                   10549: @code{read1} internally, but that's beyond the point illustrated
                   10550: here.)
1.13      pazsan   10551: 
1.78      anton    10552: Note that you cannot use @code{protected} in an interface; and
                   10553: of course you cannot define fields.
1.13      pazsan   10554: 
1.78      anton    10555: In the Neon model, all selectors are available for all classes;
                   10556: therefore it does not need interfaces. The price you pay in this model
                   10557: is slower late binding, and therefore, added complexity to avoid late
                   10558: binding.
1.13      pazsan   10559: 
1.78      anton    10560: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
                   10561: @subsubsection @file{objects.fs} Implementation
                   10562: @cindex @file{objects.fs} implementation
1.13      pazsan   10563: 
1.78      anton    10564: @cindex @code{object-map} discussion
                   10565: An object is a piece of memory, like one of the data structures
                   10566: described with @code{struct...end-struct}. It has a field
                   10567: @code{object-map} that points to the method map for the object's
                   10568: class.
1.13      pazsan   10569: 
1.78      anton    10570: @cindex method map
                   10571: @cindex virtual function table
                   10572: The @emph{method map}@footnote{This is Self terminology; in C++
                   10573: terminology: virtual function table.} is an array that contains the
                   10574: execution tokens (@i{xt}s) of the methods for the object's class. Each
                   10575: selector contains an offset into a method map.
1.13      pazsan   10576: 
1.78      anton    10577: @cindex @code{selector} implementation, class
                   10578: @code{selector} is a defining word that uses
                   10579: @code{CREATE} and @code{DOES>}. The body of the
                   10580: selector contains the offset; the @code{DOES>} action for a
                   10581: class selector is, basically:
1.8       pazsan   10582: 
                   10583: @example
1.78      anton    10584: ( object addr ) @@ over object-map @@ + @@ execute
1.13      pazsan   10585: @end example
                   10586: 
1.78      anton    10587: Since @code{object-map} is the first field of the object, it
                   10588: does not generate any code. As you can see, calling a selector has a
                   10589: small, constant cost.
1.26      crook    10590: 
1.78      anton    10591: @cindex @code{current-interface} discussion
                   10592: @cindex class implementation and representation
                   10593: A class is basically a @code{struct} combined with a method
                   10594: map. During the class definition the alignment and size of the class
                   10595: are passed on the stack, just as with @code{struct}s, so
                   10596: @code{field} can also be used for defining class
                   10597: fields. However, passing more items on the stack would be
                   10598: inconvenient, so @code{class} builds a data structure in memory,
                   10599: which is accessed through the variable
                   10600: @code{current-interface}. After its definition is complete, the
                   10601: class is represented on the stack by a pointer (e.g., as parameter for
                   10602: a child class definition).
1.26      crook    10603: 
1.78      anton    10604: A new class starts off with the alignment and size of its parent,
                   10605: and a copy of the parent's method map. Defining new fields extends the
                   10606: size and alignment; likewise, defining new selectors extends the
                   10607: method map. @code{overrides} just stores a new @i{xt} in the method
                   10608: map at the offset given by the selector.
1.13      pazsan   10609: 
1.78      anton    10610: @cindex class binding, implementation
                   10611: Class binding just gets the @i{xt} at the offset given by the selector
                   10612: from the class's method map and @code{compile,}s (in the case of
                   10613: @code{[bind]}) it.
1.13      pazsan   10614: 
1.78      anton    10615: @cindex @code{this} implementation
                   10616: @cindex @code{catch} and @code{this}
                   10617: @cindex @code{this} and @code{catch}
                   10618: I implemented @code{this} as a @code{value}. At the
                   10619: start of an @code{m:...;m} method the old @code{this} is
                   10620: stored to the return stack and restored at the end; and the object on
                   10621: the TOS is stored @code{TO this}. This technique has one
                   10622: disadvantage: If the user does not leave the method via
                   10623: @code{;m}, but via @code{throw} or @code{exit},
                   10624: @code{this} is not restored (and @code{exit} may
                   10625: crash). To deal with the @code{throw} problem, I have redefined
                   10626: @code{catch} to save and restore @code{this}; the same
                   10627: should be done with any word that can catch an exception. As for
                   10628: @code{exit}, I simply forbid it (as a replacement, there is
                   10629: @code{exitm}).
1.13      pazsan   10630: 
1.78      anton    10631: @cindex @code{inst-var} implementation
                   10632: @code{inst-var} is just the same as @code{field}, with
                   10633: a different @code{DOES>} action:
1.13      pazsan   10634: @example
1.78      anton    10635: @@ this +
1.8       pazsan   10636: @end example
1.78      anton    10637: Similar for @code{inst-value}.
1.8       pazsan   10638: 
1.78      anton    10639: @cindex class scoping implementation
                   10640: Each class also has a word list that contains the words defined with
                   10641: @code{inst-var} and @code{inst-value}, and its protected
                   10642: words. It also has a pointer to its parent. @code{class} pushes
                   10643: the word lists of the class and all its ancestors onto the search order stack,
                   10644: and @code{end-class} drops them.
1.20      pazsan   10645: 
1.78      anton    10646: @cindex interface implementation
                   10647: An interface is like a class without fields, parent and protected
                   10648: words; i.e., it just has a method map. If a class implements an
                   10649: interface, its method map contains a pointer to the method map of the
                   10650: interface. The positive offsets in the map are reserved for class
                   10651: methods, therefore interface map pointers have negative
                   10652: offsets. Interfaces have offsets that are unique throughout the
                   10653: system, unlike class selectors, whose offsets are only unique for the
                   10654: classes where the selector is available (invokable).
1.20      pazsan   10655: 
1.78      anton    10656: This structure means that interface selectors have to perform one
                   10657: indirection more than class selectors to find their method. Their body
                   10658: contains the interface map pointer offset in the class method map, and
                   10659: the method offset in the interface method map. The
                   10660: @code{does>} action for an interface selector is, basically:
1.20      pazsan   10661: 
                   10662: @example
1.78      anton    10663: ( object selector-body )
                   10664: 2dup selector-interface @@ ( object selector-body object interface-offset )
                   10665: swap object-map @@ + @@ ( object selector-body map )
                   10666: swap selector-offset @@ + @@ execute
1.20      pazsan   10667: @end example
                   10668: 
1.78      anton    10669: where @code{object-map} and @code{selector-offset} are
                   10670: first fields and generate no code.
1.20      pazsan   10671: 
1.78      anton    10672: As a concrete example, consider the following code:
1.20      pazsan   10673: 
                   10674: @example
1.78      anton    10675: interface
                   10676:   selector if1sel1
                   10677:   selector if1sel2
                   10678: end-interface if1
1.20      pazsan   10679: 
1.78      anton    10680: object class
                   10681:   if1 implementation
                   10682:   selector cl1sel1
                   10683:   cell% inst-var cl1iv1
1.20      pazsan   10684: 
1.78      anton    10685: ' m1 overrides construct
                   10686: ' m2 overrides if1sel1
                   10687: ' m3 overrides if1sel2
                   10688: ' m4 overrides cl1sel2
                   10689: end-class cl1
1.20      pazsan   10690: 
1.78      anton    10691: create obj1 object dict-new drop
                   10692: create obj2 cl1    dict-new drop
                   10693: @end example
1.20      pazsan   10694: 
1.78      anton    10695: The data structure created by this code (including the data structure
                   10696: for @code{object}) is shown in the
                   10697: @uref{objects-implementation.eps,figure}, assuming a cell size of 4.
                   10698: @comment TODO add this diagram..
1.20      pazsan   10699: 
1.78      anton    10700: @node Objects Glossary,  , Objects Implementation, Objects
                   10701: @subsubsection @file{objects.fs} Glossary
                   10702: @cindex @file{objects.fs} Glossary
1.20      pazsan   10703: 
                   10704: 
1.78      anton    10705: doc---objects-bind
                   10706: doc---objects-<bind>
                   10707: doc---objects-bind'
                   10708: doc---objects-[bind]
                   10709: doc---objects-class
                   10710: doc---objects-class->map
                   10711: doc---objects-class-inst-size
                   10712: doc---objects-class-override!
1.79      anton    10713: doc---objects-class-previous
                   10714: doc---objects-class>order
1.78      anton    10715: doc---objects-construct
                   10716: doc---objects-current'
                   10717: doc---objects-[current]
                   10718: doc---objects-current-interface
                   10719: doc---objects-dict-new
                   10720: doc---objects-end-class
                   10721: doc---objects-end-class-noname
                   10722: doc---objects-end-interface
                   10723: doc---objects-end-interface-noname
                   10724: doc---objects-end-methods
                   10725: doc---objects-exitm
                   10726: doc---objects-heap-new
                   10727: doc---objects-implementation
                   10728: doc---objects-init-object
                   10729: doc---objects-inst-value
                   10730: doc---objects-inst-var
                   10731: doc---objects-interface
                   10732: doc---objects-m:
                   10733: doc---objects-:m
                   10734: doc---objects-;m
                   10735: doc---objects-method
                   10736: doc---objects-methods
                   10737: doc---objects-object
                   10738: doc---objects-overrides
                   10739: doc---objects-[parent]
                   10740: doc---objects-print
                   10741: doc---objects-protected
                   10742: doc---objects-public
                   10743: doc---objects-selector
                   10744: doc---objects-this
                   10745: doc---objects-<to-inst>
                   10746: doc---objects-[to-inst]
                   10747: doc---objects-to-this
                   10748: doc---objects-xt-new
1.20      pazsan   10749: 
                   10750: 
1.78      anton    10751: @c -------------------------------------------------------------
                   10752: @node OOF, Mini-OOF, Objects, Object-oriented Forth
                   10753: @subsection The @file{oof.fs} model
                   10754: @cindex oof
                   10755: @cindex object-oriented programming
1.20      pazsan   10756: 
1.78      anton    10757: @cindex @file{objects.fs}
                   10758: @cindex @file{oof.fs}
1.20      pazsan   10759: 
1.78      anton    10760: This section describes the @file{oof.fs} package.
1.20      pazsan   10761: 
1.78      anton    10762: The package described in this section has been used in bigFORTH since 1991, and
                   10763: used for two large applications: a chromatographic system used to
                   10764: create new medicaments, and a graphic user interface library (MINOS).
1.20      pazsan   10765: 
1.78      anton    10766: You can find a description (in German) of @file{oof.fs} in @cite{Object
                   10767: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
                   10768: 10(2), 1994.
1.20      pazsan   10769: 
1.78      anton    10770: @menu
                   10771: * Properties of the OOF model::  
                   10772: * Basic OOF Usage::             
                   10773: * The OOF base class::          
                   10774: * Class Declaration::           
                   10775: * Class Implementation::        
                   10776: @end menu
1.20      pazsan   10777: 
1.78      anton    10778: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
                   10779: @subsubsection Properties of the @file{oof.fs} model
                   10780: @cindex @file{oof.fs} properties
1.20      pazsan   10781: 
1.78      anton    10782: @itemize @bullet
                   10783: @item
                   10784: This model combines object oriented programming with information
                   10785: hiding. It helps you writing large application, where scoping is
                   10786: necessary, because it provides class-oriented scoping.
1.20      pazsan   10787: 
1.78      anton    10788: @item
                   10789: Named objects, object pointers, and object arrays can be created,
                   10790: selector invocation uses the ``object selector'' syntax. Selector invocation
                   10791: to objects and/or selectors on the stack is a bit less convenient, but
                   10792: possible.
1.44      crook    10793: 
1.78      anton    10794: @item
                   10795: Selector invocation and instance variable usage of the active object is
                   10796: straightforward, since both make use of the active object.
1.44      crook    10797: 
1.78      anton    10798: @item
                   10799: Late binding is efficient and easy to use.
1.20      pazsan   10800: 
1.78      anton    10801: @item
                   10802: State-smart objects parse selectors. However, extensibility is provided
                   10803: using a (parsing) selector @code{postpone} and a selector @code{'}.
1.20      pazsan   10804: 
1.78      anton    10805: @item
                   10806: An implementation in ANS Forth is available.
1.20      pazsan   10807: 
1.78      anton    10808: @end itemize
1.23      crook    10809: 
                   10810: 
1.78      anton    10811: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
                   10812: @subsubsection Basic @file{oof.fs} Usage
                   10813: @cindex @file{oof.fs} usage
1.23      crook    10814: 
1.78      anton    10815: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
1.23      crook    10816: 
1.78      anton    10817: You can define a class for graphical objects like this:
1.23      crook    10818: 
1.78      anton    10819: @cindex @code{class} usage
                   10820: @cindex @code{class;} usage
                   10821: @cindex @code{method} usage
                   10822: @example
                   10823: object class graphical \ "object" is the parent class
1.139     pazsan   10824:   method draw ( x y -- )
1.78      anton    10825: class;
                   10826: @end example
1.23      crook    10827: 
1.78      anton    10828: This code defines a class @code{graphical} with an
                   10829: operation @code{draw}.  We can perform the operation
                   10830: @code{draw} on any @code{graphical} object, e.g.:
1.23      crook    10831: 
1.78      anton    10832: @example
                   10833: 100 100 t-rex draw
                   10834: @end example
1.23      crook    10835: 
1.78      anton    10836: @noindent
                   10837: where @code{t-rex} is an object or object pointer, created with e.g.
                   10838: @code{graphical : t-rex}.
1.23      crook    10839: 
1.78      anton    10840: @cindex abstract class
                   10841: How do we create a graphical object? With the present definitions,
                   10842: we cannot create a useful graphical object. The class
                   10843: @code{graphical} describes graphical objects in general, but not
                   10844: any concrete graphical object type (C++ users would call it an
                   10845: @emph{abstract class}); e.g., there is no method for the selector
                   10846: @code{draw} in the class @code{graphical}.
1.23      crook    10847: 
1.78      anton    10848: For concrete graphical objects, we define child classes of the
                   10849: class @code{graphical}, e.g.:
1.23      crook    10850: 
1.78      anton    10851: @example
                   10852: graphical class circle \ "graphical" is the parent class
                   10853:   cell var circle-radius
                   10854: how:
                   10855:   : draw ( x y -- )
                   10856:     circle-radius @@ draw-circle ;
1.23      crook    10857: 
1.139     pazsan   10858:   : init ( n-radius -- )
1.78      anton    10859:     circle-radius ! ;
                   10860: class;
                   10861: @end example
1.1       anton    10862: 
1.78      anton    10863: Here we define a class @code{circle} as a child of @code{graphical},
                   10864: with a field @code{circle-radius}; it defines new methods for the
                   10865: selectors @code{draw} and @code{init} (@code{init} is defined in
                   10866: @code{object}, the parent class of @code{graphical}).
1.1       anton    10867: 
1.78      anton    10868: Now we can create a circle in the dictionary with:
1.1       anton    10869: 
1.78      anton    10870: @example
                   10871: 50 circle : my-circle
                   10872: @end example
1.21      crook    10873: 
1.78      anton    10874: @noindent
                   10875: @code{:} invokes @code{init}, thus initializing the field
                   10876: @code{circle-radius} with 50. We can draw this new circle at (100,100)
                   10877: with:
1.1       anton    10878: 
1.78      anton    10879: @example
                   10880: 100 100 my-circle draw
                   10881: @end example
1.1       anton    10882: 
1.78      anton    10883: @cindex selector invocation, restrictions
                   10884: @cindex class definition, restrictions
                   10885: Note: You can only invoke a selector if the receiving object belongs to
                   10886: the class where the selector was defined or one of its descendents;
                   10887: e.g., you can invoke @code{draw} only for objects belonging to
                   10888: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
                   10889: mechanism will check if you try to invoke a selector that is not
                   10890: defined in this class hierarchy, so you'll get an error at compilation
                   10891: time.
1.1       anton    10892: 
                   10893: 
1.78      anton    10894: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
                   10895: @subsubsection The @file{oof.fs} base class
                   10896: @cindex @file{oof.fs} base class
1.1       anton    10897: 
1.78      anton    10898: When you define a class, you have to specify a parent class.  So how do
                   10899: you start defining classes? There is one class available from the start:
                   10900: @code{object}. You have to use it as ancestor for all classes. It is the
                   10901: only class that has no parent. Classes are also objects, except that
                   10902: they don't have instance variables; class manipulation such as
                   10903: inheritance or changing definitions of a class is handled through
                   10904: selectors of the class @code{object}.
1.1       anton    10905: 
1.78      anton    10906: @code{object} provides a number of selectors:
1.1       anton    10907: 
1.78      anton    10908: @itemize @bullet
                   10909: @item
                   10910: @code{class} for subclassing, @code{definitions} to add definitions
                   10911: later on, and @code{class?} to get type informations (is the class a
                   10912: subclass of the class passed on the stack?).
1.1       anton    10913: 
1.78      anton    10914: doc---object-class
                   10915: doc---object-definitions
                   10916: doc---object-class?
1.1       anton    10917: 
                   10918: 
1.26      crook    10919: @item
1.78      anton    10920: @code{init} and @code{dispose} as constructor and destructor of the
                   10921: object. @code{init} is invocated after the object's memory is allocated,
                   10922: while @code{dispose} also handles deallocation. Thus if you redefine
                   10923: @code{dispose}, you have to call the parent's dispose with @code{super
                   10924: dispose}, too.
                   10925: 
                   10926: doc---object-init
                   10927: doc---object-dispose
                   10928: 
1.1       anton    10929: 
1.26      crook    10930: @item
1.78      anton    10931: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
                   10932: @code{[]} to create named and unnamed objects and object arrays or
                   10933: object pointers.
                   10934: 
                   10935: doc---object-new
                   10936: doc---object-new[]
                   10937: doc---object-:
                   10938: doc---object-ptr
                   10939: doc---object-asptr
                   10940: doc---object-[]
                   10941: 
1.1       anton    10942: 
1.26      crook    10943: @item
1.78      anton    10944: @code{::} and @code{super} for explicit scoping. You should use explicit
                   10945: scoping only for super classes or classes with the same set of instance
                   10946: variables. Explicitly-scoped selectors use early binding.
1.21      crook    10947: 
1.78      anton    10948: doc---object-::
                   10949: doc---object-super
1.21      crook    10950: 
                   10951: 
1.26      crook    10952: @item
1.78      anton    10953: @code{self} to get the address of the object
1.21      crook    10954: 
1.78      anton    10955: doc---object-self
1.21      crook    10956: 
                   10957: 
1.78      anton    10958: @item
                   10959: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
                   10960: pointers and instance defers.
1.21      crook    10961: 
1.78      anton    10962: doc---object-bind
                   10963: doc---object-bound
                   10964: doc---object-link
                   10965: doc---object-is
1.21      crook    10966: 
                   10967: 
1.78      anton    10968: @item
                   10969: @code{'} to obtain selector tokens, @code{send} to invocate selectors
                   10970: form the stack, and @code{postpone} to generate selector invocation code.
1.21      crook    10971: 
1.78      anton    10972: doc---object-'
                   10973: doc---object-postpone
1.21      crook    10974: 
                   10975: 
1.78      anton    10976: @item
                   10977: @code{with} and @code{endwith} to select the active object from the
                   10978: stack, and enable its scope. Using @code{with} and @code{endwith}
                   10979: also allows you to create code using selector @code{postpone} without being
                   10980: trapped by the state-smart objects.
1.21      crook    10981: 
1.78      anton    10982: doc---object-with
                   10983: doc---object-endwith
1.21      crook    10984: 
                   10985: 
1.78      anton    10986: @end itemize
1.21      crook    10987: 
1.78      anton    10988: @node Class Declaration, Class Implementation, The OOF base class, OOF
                   10989: @subsubsection Class Declaration
                   10990: @cindex class declaration
1.21      crook    10991: 
1.78      anton    10992: @itemize @bullet
                   10993: @item
                   10994: Instance variables
1.21      crook    10995: 
1.78      anton    10996: doc---oof-var
1.21      crook    10997: 
                   10998: 
1.78      anton    10999: @item
                   11000: Object pointers
1.21      crook    11001: 
1.78      anton    11002: doc---oof-ptr
                   11003: doc---oof-asptr
1.21      crook    11004: 
                   11005: 
1.78      anton    11006: @item
                   11007: Instance defers
1.21      crook    11008: 
1.78      anton    11009: doc---oof-defer
1.21      crook    11010: 
                   11011: 
1.78      anton    11012: @item
                   11013: Method selectors
1.21      crook    11014: 
1.78      anton    11015: doc---oof-early
                   11016: doc---oof-method
1.21      crook    11017: 
                   11018: 
1.78      anton    11019: @item
                   11020: Class-wide variables
1.21      crook    11021: 
1.78      anton    11022: doc---oof-static
1.21      crook    11023: 
                   11024: 
1.78      anton    11025: @item
                   11026: End declaration
1.1       anton    11027: 
1.78      anton    11028: doc---oof-how:
                   11029: doc---oof-class;
1.21      crook    11030: 
                   11031: 
1.78      anton    11032: @end itemize
1.21      crook    11033: 
1.78      anton    11034: @c -------------------------------------------------------------
                   11035: @node Class Implementation,  , Class Declaration, OOF
                   11036: @subsubsection Class Implementation
                   11037: @cindex class implementation
1.21      crook    11038: 
1.78      anton    11039: @c -------------------------------------------------------------
                   11040: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
                   11041: @subsection The @file{mini-oof.fs} model
                   11042: @cindex mini-oof
1.21      crook    11043: 
1.78      anton    11044: Gforth's third object oriented Forth package is a 12-liner. It uses a
1.79      anton    11045: mixture of the @file{objects.fs} and the @file{oof.fs} syntax,
1.78      anton    11046: and reduces to the bare minimum of features. This is based on a posting
                   11047: of Bernd Paysan in comp.lang.forth.
1.21      crook    11048: 
1.78      anton    11049: @menu
                   11050: * Basic Mini-OOF Usage::        
                   11051: * Mini-OOF Example::            
                   11052: * Mini-OOF Implementation::     
                   11053: @end menu
1.21      crook    11054: 
1.78      anton    11055: @c -------------------------------------------------------------
                   11056: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
                   11057: @subsubsection Basic @file{mini-oof.fs} Usage
                   11058: @cindex mini-oof usage
1.21      crook    11059: 
1.78      anton    11060: There is a base class (@code{class}, which allocates one cell for the
                   11061: object pointer) plus seven other words: to define a method, a variable,
                   11062: a class; to end a class, to resolve binding, to allocate an object and
                   11063: to compile a class method.
                   11064: @comment TODO better description of the last one
1.26      crook    11065: 
1.21      crook    11066: 
1.78      anton    11067: doc-object
                   11068: doc-method
                   11069: doc-var
                   11070: doc-class
                   11071: doc-end-class
                   11072: doc-defines
                   11073: doc-new
                   11074: doc-::
1.21      crook    11075: 
                   11076: 
                   11077: 
1.78      anton    11078: @c -------------------------------------------------------------
                   11079: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
                   11080: @subsubsection Mini-OOF Example
                   11081: @cindex mini-oof example
1.1       anton    11082: 
1.78      anton    11083: A short example shows how to use this package. This example, in slightly
                   11084: extended form, is supplied as @file{moof-exm.fs}
                   11085: @comment TODO could flesh this out with some comments from the Forthwrite article
1.20      pazsan   11086: 
1.26      crook    11087: @example
1.78      anton    11088: object class
                   11089:   method init
                   11090:   method draw
                   11091: end-class graphical
1.26      crook    11092: @end example
1.20      pazsan   11093: 
1.78      anton    11094: This code defines a class @code{graphical} with an
                   11095: operation @code{draw}.  We can perform the operation
                   11096: @code{draw} on any @code{graphical} object, e.g.:
1.20      pazsan   11097: 
1.26      crook    11098: @example
1.78      anton    11099: 100 100 t-rex draw
1.26      crook    11100: @end example
1.12      anton    11101: 
1.78      anton    11102: where @code{t-rex} is an object or object pointer, created with e.g.
                   11103: @code{graphical new Constant t-rex}.
1.12      anton    11104: 
1.78      anton    11105: For concrete graphical objects, we define child classes of the
                   11106: class @code{graphical}, e.g.:
1.12      anton    11107: 
1.26      crook    11108: @example
                   11109: graphical class
1.78      anton    11110:   cell var circle-radius
                   11111: end-class circle \ "graphical" is the parent class
1.12      anton    11112: 
1.78      anton    11113: :noname ( x y -- )
                   11114:   circle-radius @@ draw-circle ; circle defines draw
                   11115: :noname ( r -- )
                   11116:   circle-radius ! ; circle defines init
                   11117: @end example
1.12      anton    11118: 
1.78      anton    11119: There is no implicit init method, so we have to define one. The creation
                   11120: code of the object now has to call init explicitely.
1.21      crook    11121: 
1.78      anton    11122: @example
                   11123: circle new Constant my-circle
                   11124: 50 my-circle init
1.12      anton    11125: @end example
                   11126: 
1.78      anton    11127: It is also possible to add a function to create named objects with
                   11128: automatic call of @code{init}, given that all objects have @code{init}
                   11129: on the same place:
1.38      anton    11130: 
1.78      anton    11131: @example
                   11132: : new: ( .. o "name" -- )
                   11133:     new dup Constant init ;
                   11134: 80 circle new: large-circle
                   11135: @end example
1.12      anton    11136: 
1.78      anton    11137: We can draw this new circle at (100,100) with:
1.12      anton    11138: 
1.78      anton    11139: @example
                   11140: 100 100 my-circle draw
                   11141: @end example
1.12      anton    11142: 
1.78      anton    11143: @node Mini-OOF Implementation,  , Mini-OOF Example, Mini-OOF
                   11144: @subsubsection @file{mini-oof.fs} Implementation
1.12      anton    11145: 
1.78      anton    11146: Object-oriented systems with late binding typically use a
                   11147: ``vtable''-approach: the first variable in each object is a pointer to a
                   11148: table, which contains the methods as function pointers. The vtable
                   11149: may also contain other information.
1.12      anton    11150: 
1.79      anton    11151: So first, let's declare selectors:
1.37      anton    11152: 
                   11153: @example
1.79      anton    11154: : method ( m v "name" -- m' v ) Create  over , swap cell+ swap
1.78      anton    11155:   DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
                   11156: @end example
1.37      anton    11157: 
1.79      anton    11158: During selector declaration, the number of selectors and instance
                   11159: variables is on the stack (in address units). @code{method} creates one
                   11160: selector and increments the selector number. To execute a selector, it
1.78      anton    11161: takes the object, fetches the vtable pointer, adds the offset, and
1.79      anton    11162: executes the method @i{xt} stored there. Each selector takes the object
                   11163: it is invoked with as top of stack parameter; it passes the parameters
                   11164: (including the object) unchanged to the appropriate method which should
1.78      anton    11165: consume that object.
1.37      anton    11166: 
1.78      anton    11167: Now, we also have to declare instance variables
1.37      anton    11168: 
1.78      anton    11169: @example
1.79      anton    11170: : var ( m v size "name" -- m v' ) Create  over , +
1.78      anton    11171:   DOES> ( o -- addr ) @@ + ;
1.37      anton    11172: @end example
                   11173: 
1.78      anton    11174: As before, a word is created with the current offset. Instance
                   11175: variables can have different sizes (cells, floats, doubles, chars), so
                   11176: all we do is take the size and add it to the offset. If your machine
                   11177: has alignment restrictions, put the proper @code{aligned} or
                   11178: @code{faligned} before the variable, to adjust the variable
                   11179: offset. That's why it is on the top of stack.
1.37      anton    11180: 
1.78      anton    11181: We need a starting point (the base object) and some syntactic sugar:
1.37      anton    11182: 
1.78      anton    11183: @example
                   11184: Create object  1 cells , 2 cells ,
1.79      anton    11185: : class ( class -- class selectors vars ) dup 2@@ ;
1.78      anton    11186: @end example
1.12      anton    11187: 
1.78      anton    11188: For inheritance, the vtable of the parent object has to be
                   11189: copied when a new, derived class is declared. This gives all the
                   11190: methods of the parent class, which can be overridden, though.
1.12      anton    11191: 
1.78      anton    11192: @example
1.79      anton    11193: : end-class  ( class selectors vars "name" -- )
1.78      anton    11194:   Create  here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
                   11195:   cell+ dup cell+ r> rot @@ 2 cells /string move ;
                   11196: @end example
1.12      anton    11197: 
1.78      anton    11198: The first line creates the vtable, initialized with
                   11199: @code{noop}s. The second line is the inheritance mechanism, it
                   11200: copies the xts from the parent vtable.
1.12      anton    11201: 
1.78      anton    11202: We still have no way to define new methods, let's do that now:
1.12      anton    11203: 
1.26      crook    11204: @example
1.79      anton    11205: : defines ( xt class "name" -- ) ' >body @@ + ! ;
1.78      anton    11206: @end example
1.12      anton    11207: 
1.78      anton    11208: To allocate a new object, we need a word, too:
1.12      anton    11209: 
1.78      anton    11210: @example
                   11211: : new ( class -- o )  here over @@ allot swap over ! ;
1.12      anton    11212: @end example
                   11213: 
1.78      anton    11214: Sometimes derived classes want to access the method of the
                   11215: parent object. There are two ways to achieve this with Mini-OOF:
                   11216: first, you could use named words, and second, you could look up the
                   11217: vtable of the parent object.
1.12      anton    11218: 
1.78      anton    11219: @example
                   11220: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
                   11221: @end example
1.12      anton    11222: 
                   11223: 
1.78      anton    11224: Nothing can be more confusing than a good example, so here is
                   11225: one. First let's declare a text object (called
                   11226: @code{button}), that stores text and position:
1.12      anton    11227: 
1.78      anton    11228: @example
                   11229: object class
                   11230:   cell var text
                   11231:   cell var len
                   11232:   cell var x
                   11233:   cell var y
                   11234:   method init
                   11235:   method draw
                   11236: end-class button
                   11237: @end example
1.12      anton    11238: 
1.78      anton    11239: @noindent
                   11240: Now, implement the two methods, @code{draw} and @code{init}:
1.21      crook    11241: 
1.26      crook    11242: @example
1.78      anton    11243: :noname ( o -- )
                   11244:  >r r@@ x @@ r@@ y @@ at-xy  r@@ text @@ r> len @@ type ;
                   11245:  button defines draw
                   11246: :noname ( addr u o -- )
                   11247:  >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
                   11248:  button defines init
1.26      crook    11249: @end example
1.12      anton    11250: 
1.78      anton    11251: @noindent
                   11252: To demonstrate inheritance, we define a class @code{bold-button}, with no
1.79      anton    11253: new data and no new selectors:
1.78      anton    11254: 
                   11255: @example
                   11256: button class
                   11257: end-class bold-button
1.12      anton    11258: 
1.78      anton    11259: : bold   27 emit ." [1m" ;
                   11260: : normal 27 emit ." [0m" ;
                   11261: @end example
1.1       anton    11262: 
1.78      anton    11263: @noindent
                   11264: The class @code{bold-button} has a different draw method to
                   11265: @code{button}, but the new method is defined in terms of the draw method
                   11266: for @code{button}:
1.20      pazsan   11267: 
1.78      anton    11268: @example
                   11269: :noname bold [ button :: draw ] normal ; bold-button defines draw
                   11270: @end example
1.21      crook    11271: 
1.78      anton    11272: @noindent
1.79      anton    11273: Finally, create two objects and apply selectors:
1.21      crook    11274: 
1.26      crook    11275: @example
1.78      anton    11276: button new Constant foo
                   11277: s" thin foo" foo init
                   11278: page
                   11279: foo draw
                   11280: bold-button new Constant bar
                   11281: s" fat bar" bar init
                   11282: 1 bar y !
                   11283: bar draw
1.26      crook    11284: @end example
1.21      crook    11285: 
                   11286: 
1.78      anton    11287: @node Comparison with other object models,  , Mini-OOF, Object-oriented Forth
                   11288: @subsection Comparison with other object models
                   11289: @cindex comparison of object models
                   11290: @cindex object models, comparison
                   11291: 
                   11292: Many object-oriented Forth extensions have been proposed (@cite{A survey
                   11293: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
                   11294: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
                   11295: relation of the object models described here to two well-known and two
                   11296: closely-related (by the use of method maps) models.  Andras Zsoter
                   11297: helped us with this section.
                   11298: 
                   11299: @cindex Neon model
                   11300: The most popular model currently seems to be the Neon model (see
                   11301: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
                   11302: 1997) by Andrew McKewan) but this model has a number of limitations
                   11303: @footnote{A longer version of this critique can be
                   11304: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
                   11305: Dimensions, May 1997) by Anton Ertl.}:
                   11306: 
                   11307: @itemize @bullet
                   11308: @item
                   11309: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
                   11310: to pass objects on the stack.
1.21      crook    11311: 
1.78      anton    11312: @item
                   11313: It requires that the selector parses the input stream (at
1.79      anton    11314: compile time); this leads to reduced extensibility and to bugs that are
1.78      anton    11315: hard to find.
1.21      crook    11316: 
1.78      anton    11317: @item
1.79      anton    11318: It allows using every selector on every object; this eliminates the
                   11319: need for interfaces, but makes it harder to create efficient
                   11320: implementations.
1.78      anton    11321: @end itemize
1.21      crook    11322: 
1.78      anton    11323: @cindex Pountain's object-oriented model
                   11324: Another well-known publication is @cite{Object-Oriented Forth} (Academic
                   11325: Press, London, 1987) by Dick Pountain. However, it is not really about
                   11326: object-oriented programming, because it hardly deals with late
                   11327: binding. Instead, it focuses on features like information hiding and
                   11328: overloading that are characteristic of modular languages like Ada (83).
1.26      crook    11329: 
1.78      anton    11330: @cindex Zsoter's object-oriented model
1.79      anton    11331: In @uref{http://www.forth.org/oopf.html, Does late binding have to be
                   11332: slow?} (Forth Dimensions 18(1) 1996, pages 31-35) Andras Zsoter
                   11333: describes a model that makes heavy use of an active object (like
                   11334: @code{this} in @file{objects.fs}): The active object is not only used
                   11335: for accessing all fields, but also specifies the receiving object of
                   11336: every selector invocation; you have to change the active object
                   11337: explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
                   11338: changes more or less implicitly at @code{m: ... ;m}. Such a change at
                   11339: the method entry point is unnecessary with Zsoter's model, because the
                   11340: receiving object is the active object already. On the other hand, the
                   11341: explicit change is absolutely necessary in that model, because otherwise
                   11342: no one could ever change the active object. An ANS Forth implementation
                   11343: of this model is available through
                   11344: @uref{http://www.forth.org/oopf.html}.
1.21      crook    11345: 
1.78      anton    11346: @cindex @file{oof.fs}, differences to other models
                   11347: The @file{oof.fs} model combines information hiding and overloading
                   11348: resolution (by keeping names in various word lists) with object-oriented
                   11349: programming. It sets the active object implicitly on method entry, but
                   11350: also allows explicit changing (with @code{>o...o>} or with
                   11351: @code{with...endwith}). It uses parsing and state-smart objects and
                   11352: classes for resolving overloading and for early binding: the object or
                   11353: class parses the selector and determines the method from this. If the
                   11354: selector is not parsed by an object or class, it performs a call to the
                   11355: selector for the active object (late binding), like Zsoter's model.
                   11356: Fields are always accessed through the active object. The big
                   11357: disadvantage of this model is the parsing and the state-smartness, which
                   11358: reduces extensibility and increases the opportunities for subtle bugs;
                   11359: essentially, you are only safe if you never tick or @code{postpone} an
                   11360: object or class (Bernd disagrees, but I (Anton) am not convinced).
1.21      crook    11361: 
1.78      anton    11362: @cindex @file{mini-oof.fs}, differences to other models
                   11363: The @file{mini-oof.fs} model is quite similar to a very stripped-down
                   11364: version of the @file{objects.fs} model, but syntactically it is a
                   11365: mixture of the @file{objects.fs} and @file{oof.fs} models.
1.21      crook    11366: 
                   11367: 
1.78      anton    11368: @c -------------------------------------------------------------
                   11369: @node Programming Tools, Assembler and Code Words, Object-oriented Forth, Words
                   11370: @section Programming Tools
                   11371: @cindex programming tools
1.21      crook    11372: 
1.78      anton    11373: @c !! move this and assembler down below OO stuff.
1.21      crook    11374: 
1.78      anton    11375: @menu
                   11376: * Examining::                   
                   11377: * Forgetting words::            
                   11378: * Debugging::                   Simple and quick.
                   11379: * Assertions::                  Making your programs self-checking.
                   11380: * Singlestep Debugger::         Executing your program word by word.
                   11381: @end menu
1.21      crook    11382: 
1.78      anton    11383: @node Examining, Forgetting words, Programming Tools, Programming Tools
                   11384: @subsection Examining data and code
                   11385: @cindex examining data and code
                   11386: @cindex data examination
                   11387: @cindex code examination
1.44      crook    11388: 
1.78      anton    11389: The following words inspect the stack non-destructively:
1.21      crook    11390: 
1.78      anton    11391: doc-.s
                   11392: doc-f.s
1.44      crook    11393: 
1.78      anton    11394: There is a word @code{.r} but it does @i{not} display the return stack!
                   11395: It is used for formatted numeric output (@pxref{Simple numeric output}).
1.21      crook    11396: 
1.78      anton    11397: doc-depth
                   11398: doc-fdepth
                   11399: doc-clearstack
1.124     anton    11400: doc-clearstacks
1.21      crook    11401: 
1.78      anton    11402: The following words inspect memory.
1.21      crook    11403: 
1.78      anton    11404: doc-?
                   11405: doc-dump
1.21      crook    11406: 
1.78      anton    11407: And finally, @code{see} allows to inspect code:
1.21      crook    11408: 
1.78      anton    11409: doc-see
                   11410: doc-xt-see
1.111     anton    11411: doc-simple-see
                   11412: doc-simple-see-range
1.21      crook    11413: 
1.78      anton    11414: @node Forgetting words, Debugging, Examining, Programming Tools
                   11415: @subsection Forgetting words
                   11416: @cindex words, forgetting
                   11417: @cindex forgeting words
1.21      crook    11418: 
1.78      anton    11419: @c  anton: other, maybe better places for this subsection: Defining Words;
                   11420: @c  Dictionary allocation.  At least a reference should be there.
1.21      crook    11421: 
1.78      anton    11422: Forth allows you to forget words (and everything that was alloted in the
                   11423: dictonary after them) in a LIFO manner.
1.21      crook    11424: 
1.78      anton    11425: doc-marker
1.21      crook    11426: 
1.78      anton    11427: The most common use of this feature is during progam development: when
                   11428: you change a source file, forget all the words it defined and load it
                   11429: again (since you also forget everything defined after the source file
                   11430: was loaded, you have to reload that, too).  Note that effects like
                   11431: storing to variables and destroyed system words are not undone when you
                   11432: forget words.  With a system like Gforth, that is fast enough at
                   11433: starting up and compiling, I find it more convenient to exit and restart
                   11434: Gforth, as this gives me a clean slate.
1.21      crook    11435: 
1.78      anton    11436: Here's an example of using @code{marker} at the start of a source file
                   11437: that you are debugging; it ensures that you only ever have one copy of
                   11438: the file's definitions compiled at any time:
1.21      crook    11439: 
1.78      anton    11440: @example
                   11441: [IFDEF] my-code
                   11442:     my-code
                   11443: [ENDIF]
1.26      crook    11444: 
1.78      anton    11445: marker my-code
                   11446: init-included-files
1.21      crook    11447: 
1.78      anton    11448: \ .. definitions start here
                   11449: \ .
                   11450: \ .
                   11451: \ end
                   11452: @end example
1.21      crook    11453: 
1.26      crook    11454: 
1.78      anton    11455: @node Debugging, Assertions, Forgetting words, Programming Tools
                   11456: @subsection Debugging
                   11457: @cindex debugging
1.21      crook    11458: 
1.78      anton    11459: Languages with a slow edit/compile/link/test development loop tend to
                   11460: require sophisticated tracing/stepping debuggers to facilate debugging.
1.21      crook    11461: 
1.78      anton    11462: A much better (faster) way in fast-compiling languages is to add
                   11463: printing code at well-selected places, let the program run, look at
                   11464: the output, see where things went wrong, add more printing code, etc.,
                   11465: until the bug is found.
1.21      crook    11466: 
1.78      anton    11467: The simple debugging aids provided in @file{debugs.fs}
                   11468: are meant to support this style of debugging.
1.21      crook    11469: 
1.78      anton    11470: The word @code{~~} prints debugging information (by default the source
                   11471: location and the stack contents). It is easy to insert. If you use Emacs
                   11472: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
                   11473: query-replace them with nothing). The deferred words
1.101     anton    11474: @code{printdebugdata} and @code{.debugline} control the output of
1.78      anton    11475: @code{~~}. The default source location output format works well with
                   11476: Emacs' compilation mode, so you can step through the program at the
                   11477: source level using @kbd{C-x `} (the advantage over a stepping debugger
                   11478: is that you can step in any direction and you know where the crash has
                   11479: happened or where the strange data has occurred).
1.21      crook    11480: 
1.78      anton    11481: doc-~~
                   11482: doc-printdebugdata
1.101     anton    11483: doc-.debugline
1.21      crook    11484: 
1.106     anton    11485: @cindex filenames in @code{~~} output
                   11486: @code{~~} (and assertions) will usually print the wrong file name if a
                   11487: marker is executed in the same file after their occurance.  They will
                   11488: print @samp{*somewhere*} as file name if a marker is executed in the
                   11489: same file before their occurance.
                   11490: 
                   11491: 
1.78      anton    11492: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
                   11493: @subsection Assertions
                   11494: @cindex assertions
1.21      crook    11495: 
1.78      anton    11496: It is a good idea to make your programs self-checking, especially if you
                   11497: make an assumption that may become invalid during maintenance (for
                   11498: example, that a certain field of a data structure is never zero). Gforth
                   11499: supports @dfn{assertions} for this purpose. They are used like this:
1.21      crook    11500: 
                   11501: @example
1.78      anton    11502: assert( @i{flag} )
1.26      crook    11503: @end example
                   11504: 
1.78      anton    11505: The code between @code{assert(} and @code{)} should compute a flag, that
                   11506: should be true if everything is alright and false otherwise. It should
                   11507: not change anything else on the stack. The overall stack effect of the
                   11508: assertion is @code{( -- )}. E.g.
1.21      crook    11509: 
1.26      crook    11510: @example
1.78      anton    11511: assert( 1 1 + 2 = ) \ what we learn in school
                   11512: assert( dup 0<> ) \ assert that the top of stack is not zero
                   11513: assert( false ) \ this code should not be reached
1.21      crook    11514: @end example
                   11515: 
1.78      anton    11516: The need for assertions is different at different times. During
                   11517: debugging, we want more checking, in production we sometimes care more
                   11518: for speed. Therefore, assertions can be turned off, i.e., the assertion
                   11519: becomes a comment. Depending on the importance of an assertion and the
                   11520: time it takes to check it, you may want to turn off some assertions and
                   11521: keep others turned on. Gforth provides several levels of assertions for
                   11522: this purpose:
                   11523: 
                   11524: 
                   11525: doc-assert0(
                   11526: doc-assert1(
                   11527: doc-assert2(
                   11528: doc-assert3(
                   11529: doc-assert(
                   11530: doc-)
1.21      crook    11531: 
                   11532: 
1.78      anton    11533: The variable @code{assert-level} specifies the highest assertions that
                   11534: are turned on. I.e., at the default @code{assert-level} of one,
                   11535: @code{assert0(} and @code{assert1(} assertions perform checking, while
                   11536: @code{assert2(} and @code{assert3(} assertions are treated as comments.
1.26      crook    11537: 
1.78      anton    11538: The value of @code{assert-level} is evaluated at compile-time, not at
                   11539: run-time. Therefore you cannot turn assertions on or off at run-time;
                   11540: you have to set the @code{assert-level} appropriately before compiling a
                   11541: piece of code. You can compile different pieces of code at different
                   11542: @code{assert-level}s (e.g., a trusted library at level 1 and
                   11543: newly-written code at level 3).
1.26      crook    11544: 
                   11545: 
1.78      anton    11546: doc-assert-level
1.26      crook    11547: 
                   11548: 
1.78      anton    11549: If an assertion fails, a message compatible with Emacs' compilation mode
                   11550: is produced and the execution is aborted (currently with @code{ABORT"}.
                   11551: If there is interest, we will introduce a special throw code. But if you
                   11552: intend to @code{catch} a specific condition, using @code{throw} is
                   11553: probably more appropriate than an assertion).
1.106     anton    11554: 
                   11555: @cindex filenames in assertion output
                   11556: Assertions (and @code{~~}) will usually print the wrong file name if a
                   11557: marker is executed in the same file after their occurance.  They will
                   11558: print @samp{*somewhere*} as file name if a marker is executed in the
                   11559: same file before their occurance.
1.44      crook    11560: 
1.78      anton    11561: Definitions in ANS Forth for these assertion words are provided
                   11562: in @file{compat/assert.fs}.
1.26      crook    11563: 
1.44      crook    11564: 
1.78      anton    11565: @node Singlestep Debugger,  , Assertions, Programming Tools
                   11566: @subsection Singlestep Debugger
                   11567: @cindex singlestep Debugger
                   11568: @cindex debugging Singlestep
1.44      crook    11569: 
1.112     anton    11570: The singlestep debugger does not work in this release.
                   11571: 
1.78      anton    11572: When you create a new word there's often the need to check whether it
                   11573: behaves correctly or not. You can do this by typing @code{dbg
                   11574: badword}. A debug session might look like this:
1.26      crook    11575: 
1.78      anton    11576: @example
                   11577: : badword 0 DO i . LOOP ;  ok
                   11578: 2 dbg badword 
                   11579: : badword  
                   11580: Scanning code...
1.44      crook    11581: 
1.78      anton    11582: Nesting debugger ready!
1.44      crook    11583: 
1.78      anton    11584: 400D4738  8049BC4 0              -> [ 2 ] 00002 00000 
                   11585: 400D4740  8049F68 DO             -> [ 0 ] 
                   11586: 400D4744  804A0C8 i              -> [ 1 ] 00000 
                   11587: 400D4748 400C5E60 .              -> 0 [ 0 ] 
                   11588: 400D474C  8049D0C LOOP           -> [ 0 ] 
                   11589: 400D4744  804A0C8 i              -> [ 1 ] 00001 
                   11590: 400D4748 400C5E60 .              -> 1 [ 0 ] 
                   11591: 400D474C  8049D0C LOOP           -> [ 0 ] 
                   11592: 400D4758  804B384 ;              ->  ok
                   11593: @end example
1.21      crook    11594: 
1.78      anton    11595: Each line displayed is one step. You always have to hit return to
                   11596: execute the next word that is displayed. If you don't want to execute
                   11597: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
                   11598: an overview what keys are available:
1.44      crook    11599: 
1.78      anton    11600: @table @i
1.44      crook    11601: 
1.78      anton    11602: @item @key{RET}
                   11603: Next; Execute the next word.
1.21      crook    11604: 
1.78      anton    11605: @item n
                   11606: Nest; Single step through next word.
1.44      crook    11607: 
1.78      anton    11608: @item u
                   11609: Unnest; Stop debugging and execute rest of word. If we got to this word
                   11610: with nest, continue debugging with the calling word.
1.44      crook    11611: 
1.78      anton    11612: @item d
                   11613: Done; Stop debugging and execute rest.
1.21      crook    11614: 
1.78      anton    11615: @item s
                   11616: Stop; Abort immediately.
1.44      crook    11617: 
1.78      anton    11618: @end table
1.44      crook    11619: 
1.78      anton    11620: Debugging large application with this mechanism is very difficult, because
                   11621: you have to nest very deeply into the program before the interesting part
                   11622: begins. This takes a lot of time. 
1.26      crook    11623: 
1.78      anton    11624: To do it more directly put a @code{BREAK:} command into your source code.
                   11625: When program execution reaches @code{BREAK:} the single step debugger is
                   11626: invoked and you have all the features described above.
1.44      crook    11627: 
1.78      anton    11628: If you have more than one part to debug it is useful to know where the
                   11629: program has stopped at the moment. You can do this by the 
                   11630: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
                   11631: string is typed out when the ``breakpoint'' is reached.
1.44      crook    11632: 
1.26      crook    11633: 
1.78      anton    11634: doc-dbg
                   11635: doc-break:
                   11636: doc-break"
1.44      crook    11637: 
                   11638: 
1.26      crook    11639: 
1.78      anton    11640: @c -------------------------------------------------------------
                   11641: @node Assembler and Code Words, Threading Words, Programming Tools, Words
                   11642: @section Assembler and Code Words
                   11643: @cindex assembler
                   11644: @cindex code words
1.44      crook    11645: 
1.78      anton    11646: @menu
                   11647: * Code and ;code::              
                   11648: * Common Assembler::            Assembler Syntax
                   11649: * Common Disassembler::         
                   11650: * 386 Assembler::               Deviations and special cases
                   11651: * Alpha Assembler::             Deviations and special cases
                   11652: * MIPS assembler::              Deviations and special cases
                   11653: * Other assemblers::            How to write them
                   11654: @end menu
1.21      crook    11655: 
1.78      anton    11656: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
                   11657: @subsection @code{Code} and @code{;code}
1.26      crook    11658: 
1.78      anton    11659: Gforth provides some words for defining primitives (words written in
                   11660: machine code), and for defining the machine-code equivalent of
                   11661: @code{DOES>}-based defining words. However, the machine-independent
                   11662: nature of Gforth poses a few problems: First of all, Gforth runs on
                   11663: several architectures, so it can provide no standard assembler. What's
                   11664: worse is that the register allocation not only depends on the processor,
                   11665: but also on the @code{gcc} version and options used.
1.44      crook    11666: 
1.78      anton    11667: The words that Gforth offers encapsulate some system dependences (e.g.,
                   11668: the header structure), so a system-independent assembler may be used in
                   11669: Gforth. If you do not have an assembler, you can compile machine code
                   11670: directly with @code{,} and @code{c,}@footnote{This isn't portable,
                   11671: because these words emit stuff in @i{data} space; it works because
                   11672: Gforth has unified code/data spaces. Assembler isn't likely to be
                   11673: portable anyway.}.
1.21      crook    11674: 
1.44      crook    11675: 
1.78      anton    11676: doc-assembler
                   11677: doc-init-asm
                   11678: doc-code
                   11679: doc-end-code
                   11680: doc-;code
                   11681: doc-flush-icache
1.44      crook    11682: 
1.21      crook    11683: 
1.78      anton    11684: If @code{flush-icache} does not work correctly, @code{code} words
                   11685: etc. will not work (reliably), either.
1.44      crook    11686: 
1.78      anton    11687: The typical usage of these @code{code} words can be shown most easily by
                   11688: analogy to the equivalent high-level defining words:
1.44      crook    11689: 
1.78      anton    11690: @example
                   11691: : foo                              code foo
                   11692:    <high-level Forth words>              <assembler>
                   11693: ;                                  end-code
                   11694:                                 
                   11695: : bar                              : bar
                   11696:    <high-level Forth words>           <high-level Forth words>
                   11697:    CREATE                             CREATE
                   11698:       <high-level Forth words>           <high-level Forth words>
                   11699:    DOES>                              ;code
                   11700:       <high-level Forth words>           <assembler>
                   11701: ;                                  end-code
                   11702: @end example
1.21      crook    11703: 
1.78      anton    11704: @c anton: the following stuff is also in "Common Assembler", in less detail.
1.44      crook    11705: 
1.78      anton    11706: @cindex registers of the inner interpreter
                   11707: In the assembly code you will want to refer to the inner interpreter's
                   11708: registers (e.g., the data stack pointer) and you may want to use other
                   11709: registers for temporary storage. Unfortunately, the register allocation
                   11710: is installation-dependent.
1.44      crook    11711: 
1.78      anton    11712: In particular, @code{ip} (Forth instruction pointer) and @code{rp}
1.100     anton    11713: (return stack pointer) may be in different places in @code{gforth} and
                   11714: @code{gforth-fast}, or different installations.  This means that you
                   11715: cannot write a @code{NEXT} routine that works reliably on both versions
                   11716: or different installations; so for doing @code{NEXT}, I recommend
                   11717: jumping to @code{' noop >code-address}, which contains nothing but a
                   11718: @code{NEXT}.
1.21      crook    11719: 
1.78      anton    11720: For general accesses to the inner interpreter's registers, the easiest
                   11721: solution is to use explicit register declarations (@pxref{Explicit Reg
                   11722: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) for
                   11723: all of the inner interpreter's registers: You have to compile Gforth
                   11724: with @code{-DFORCE_REG} (configure option @code{--enable-force-reg}) and
                   11725: the appropriate declarations must be present in the @code{machine.h}
                   11726: file (see @code{mips.h} for an example; you can find a full list of all
                   11727: declarable register symbols with @code{grep register engine.c}). If you
                   11728: give explicit registers to all variables that are declared at the
                   11729: beginning of @code{engine()}, you should be able to use the other
                   11730: caller-saved registers for temporary storage. Alternatively, you can use
                   11731: the @code{gcc} option @code{-ffixed-REG} (@pxref{Code Gen Options, ,
                   11732: Options for Code Generation Conventions, gcc.info, GNU C Manual}) to
                   11733: reserve a register (however, this restriction on register allocation may
                   11734: slow Gforth significantly).
1.44      crook    11735: 
1.78      anton    11736: If this solution is not viable (e.g., because @code{gcc} does not allow
                   11737: you to explicitly declare all the registers you need), you have to find
                   11738: out by looking at the code where the inner interpreter's registers
                   11739: reside and which registers can be used for temporary storage. You can
                   11740: get an assembly listing of the engine's code with @code{make engine.s}.
1.44      crook    11741: 
1.78      anton    11742: In any case, it is good practice to abstract your assembly code from the
                   11743: actual register allocation. E.g., if the data stack pointer resides in
                   11744: register @code{$17}, create an alias for this register called @code{sp},
                   11745: and use that in your assembly code.
1.21      crook    11746: 
1.78      anton    11747: @cindex code words, portable
                   11748: Another option for implementing normal and defining words efficiently
                   11749: is to add the desired functionality to the source of Gforth. For normal
                   11750: words you just have to edit @file{primitives} (@pxref{Automatic
                   11751: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
                   11752: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
                   11753: @file{prims2x.fs}, and possibly @file{cross.fs}.
1.44      crook    11754: 
1.78      anton    11755: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
                   11756: @subsection Common Assembler
1.44      crook    11757: 
1.78      anton    11758: The assemblers in Gforth generally use a postfix syntax, i.e., the
                   11759: instruction name follows the operands.
1.21      crook    11760: 
1.78      anton    11761: The operands are passed in the usual order (the same that is used in the
                   11762: manual of the architecture).  Since they all are Forth words, they have
                   11763: to be separated by spaces; you can also use Forth words to compute the
                   11764: operands.
1.44      crook    11765: 
1.78      anton    11766: The instruction names usually end with a @code{,}.  This makes it easier
                   11767: to visually separate instructions if you put several of them on one
                   11768: line; it also avoids shadowing other Forth words (e.g., @code{and}).
1.21      crook    11769: 
1.78      anton    11770: Registers are usually specified by number; e.g., (decimal) @code{11}
                   11771: specifies registers R11 and F11 on the Alpha architecture (which one,
                   11772: depends on the instruction).  The usual names are also available, e.g.,
                   11773: @code{s2} for R11 on Alpha.
1.21      crook    11774: 
1.78      anton    11775: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
                   11776: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
                   11777: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
                   11778: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}.  The
                   11779: conditions are specified in a way specific to each assembler.
1.1       anton    11780: 
1.78      anton    11781: Note that the register assignments of the Gforth engine can change
                   11782: between Gforth versions, or even between different compilations of the
                   11783: same Gforth version (e.g., if you use a different GCC version).  So if
                   11784: you want to refer to Gforth's registers (e.g., the stack pointer or
                   11785: TOS), I recommend defining your own words for refering to these
                   11786: registers, and using them later on; then you can easily adapt to a
                   11787: changed register assignment.  The stability of the register assignment
                   11788: is usually better if you build Gforth with @code{--enable-force-reg}.
1.1       anton    11789: 
1.100     anton    11790: The most common use of these registers is to dispatch to the next word
                   11791: (the @code{next} routine).  A portable way to do this is to jump to
                   11792: @code{' noop >code-address} (of course, this is less efficient than
                   11793: integrating the @code{next} code and scheduling it well).
1.1       anton    11794: 
1.96      anton    11795: Another difference between Gforth version is that the top of stack is
                   11796: kept in memory in @code{gforth} and, on most platforms, in a register in
                   11797: @code{gforth-fast}.
                   11798: 
1.78      anton    11799: @node  Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
                   11800: @subsection Common Disassembler
1.127     anton    11801: @cindex disassembler, general
                   11802: @cindex gdb disassembler
1.1       anton    11803: 
1.78      anton    11804: You can disassemble a @code{code} word with @code{see}
                   11805: (@pxref{Debugging}).  You can disassemble a section of memory with
1.1       anton    11806: 
1.127     anton    11807: doc-discode
1.44      crook    11808: 
1.127     anton    11809: There are two kinds of disassembler for Gforth: The Forth disassembler
                   11810: (available on some CPUs) and the gdb disassembler (available on
                   11811: platforms with @command{gdb} and @command{mktemp}).  If both are
                   11812: available, the Forth disassembler is used by default.  If you prefer
                   11813: the gdb disassembler, say
                   11814: 
                   11815: @example
                   11816: ' disasm-gdb is discode
                   11817: @end example
                   11818: 
                   11819: If neither is available, @code{discode} performs @code{dump}.
                   11820: 
                   11821: The Forth disassembler generally produces output that can be fed into the
1.78      anton    11822: assembler (i.e., same syntax, etc.).  It also includes additional
                   11823: information in comments.  In particular, the address of the instruction
                   11824: is given in a comment before the instruction.
1.1       anton    11825: 
1.127     anton    11826: The gdb disassembler produces output in the same format as the gdb
                   11827: @code{disassemble} command (@pxref{Machine Code,,Source and machine
                   11828: code,gdb,Debugging with GDB}), in the default flavour (AT&T syntax for
                   11829: the 386 and AMD64 architectures).
                   11830: 
1.78      anton    11831: @code{See} may display more or less than the actual code of the word,
                   11832: because the recognition of the end of the code is unreliable.  You can
1.127     anton    11833: use @code{discode} if it did not display enough.  It may display more, if
1.78      anton    11834: the code word is not immediately followed by a named word.  If you have
1.116     anton    11835: something else there, you can follow the word with @code{align latest ,}
1.78      anton    11836: to ensure that the end is recognized.
1.21      crook    11837: 
1.78      anton    11838: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
                   11839: @subsection 386 Assembler
1.44      crook    11840: 
1.78      anton    11841: The 386 assembler included in Gforth was written by Bernd Paysan, it's
                   11842: available under GPL, and originally part of bigFORTH.
1.21      crook    11843: 
1.78      anton    11844: The 386 disassembler included in Gforth was written by Andrew McKewan
                   11845: and is in the public domain.
1.21      crook    11846: 
1.91      anton    11847: The disassembler displays code in an Intel-like prefix syntax.
1.21      crook    11848: 
1.78      anton    11849: The assembler uses a postfix syntax with reversed parameters.
1.1       anton    11850: 
1.78      anton    11851: The assembler includes all instruction of the Athlon, i.e. 486 core
                   11852: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
                   11853: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
                   11854: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
1.1       anton    11855: 
1.78      anton    11856: There are several prefixes to switch between different operation sizes,
                   11857: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
                   11858: double-word accesses. Addressing modes can be switched with @code{.wa}
                   11859: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
                   11860: need a prefix for byte register names (@code{AL} et al).
1.1       anton    11861: 
1.78      anton    11862: For floating point operations, the prefixes are @code{.fs} (IEEE
                   11863: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
                   11864: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
1.21      crook    11865: 
1.78      anton    11866: The MMX opcodes don't have size prefixes, they are spelled out like in
                   11867: the Intel assembler. Instead of move from and to memory, there are
                   11868: PLDQ/PLDD and PSTQ/PSTD.
1.21      crook    11869: 
1.78      anton    11870: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
                   11871: ax.  Immediate values are indicated by postfixing them with @code{#},
1.91      anton    11872: e.g., @code{3 #}.  Here are some examples of addressing modes in various
                   11873: syntaxes:
1.21      crook    11874: 
1.26      crook    11875: @example
1.91      anton    11876: Gforth          Intel (NASM)   AT&T (gas)      Name
                   11877: .w ax           ax             %ax             register (16 bit)
                   11878: ax              eax            %eax            register (32 bit)
                   11879: 3 #             offset 3       $3              immediate
                   11880: 1000 #)         byte ptr 1000  1000            displacement
                   11881: bx )            [ebx]          (%ebx)          base
                   11882: 100 di d)       100[edi]       100(%edi)       base+displacement
                   11883: 20 ax *4 i#)    20[eax*4]      20(,%eax,4)     (index*scale)+displacement
                   11884: di ax *4 i)     [edi][eax*4]   (%edi,%eax,4)   base+(index*scale)
                   11885: 4 bx cx di)     4[ebx][ecx]    4(%ebx,%ecx)    base+index+displacement
                   11886: 12 sp ax *2 di) 12[esp][eax*2] 12(%esp,%eax,2) base+(index*scale)+displacement
                   11887: @end example
                   11888: 
                   11889: You can use @code{L)} and @code{LI)} instead of @code{D)} and
                   11890: @code{DI)} to enforce 32-bit displacement fields (useful for
                   11891: later patching).
1.21      crook    11892: 
1.78      anton    11893: Some example of instructions are:
1.1       anton    11894: 
                   11895: @example
1.78      anton    11896: ax bx mov             \ move ebx,eax
                   11897: 3 # ax mov            \ mov eax,3
1.137     pazsan   11898: 100 di d) ax mov      \ mov eax,100[edi]
1.78      anton    11899: 4 bx cx di) ax mov    \ mov eax,4[ebx][ecx]
                   11900: .w ax bx mov          \ mov bx,ax
1.1       anton    11901: @end example
                   11902: 
1.78      anton    11903: The following forms are supported for binary instructions:
1.1       anton    11904: 
                   11905: @example
1.78      anton    11906: <reg> <reg> <inst>
                   11907: <n> # <reg> <inst>
                   11908: <mem> <reg> <inst>
                   11909: <reg> <mem> <inst>
1.136     pazsan   11910: <n> # <mem> <inst>
1.1       anton    11911: @end example
                   11912: 
1.136     pazsan   11913: The shift/rotate syntax is:
1.1       anton    11914: 
1.26      crook    11915: @example
1.78      anton    11916: <reg/mem> 1 # shl \ shortens to shift without immediate
                   11917: <reg/mem> 4 # shl
                   11918: <reg/mem> cl shl
1.26      crook    11919: @end example
1.1       anton    11920: 
1.78      anton    11921: Precede string instructions (@code{movs} etc.) with @code{.b} to get
                   11922: the byte version.
1.1       anton    11923: 
1.78      anton    11924: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
                   11925: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
                   11926: pc < >= <= >}. (Note that most of these words shadow some Forth words
                   11927: when @code{assembler} is in front of @code{forth} in the search path,
                   11928: e.g., in @code{code} words).  Currently the control structure words use
                   11929: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
                   11930: to shuffle them (you can also use @code{swap} etc.).
1.21      crook    11931: 
1.78      anton    11932: Here is an example of a @code{code} word (assumes that the stack pointer
                   11933: is in esi and the TOS is in ebx):
1.21      crook    11934: 
1.26      crook    11935: @example
1.78      anton    11936: code my+ ( n1 n2 -- n )
                   11937:     4 si D) bx add
                   11938:     4 # si add
                   11939:     Next
                   11940: end-code
1.26      crook    11941: @end example
1.21      crook    11942: 
1.78      anton    11943: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
                   11944: @subsection Alpha Assembler
1.21      crook    11945: 
1.78      anton    11946: The Alpha assembler and disassembler were originally written by Bernd
                   11947: Thallner.
1.26      crook    11948: 
1.78      anton    11949: The register names @code{a0}--@code{a5} are not available to avoid
                   11950: shadowing hex numbers.
1.2       jwilke   11951: 
1.78      anton    11952: Immediate forms of arithmetic instructions are distinguished by a
                   11953: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
                   11954: does not count as arithmetic instruction).
1.2       jwilke   11955: 
1.78      anton    11956: You have to specify all operands to an instruction, even those that
                   11957: other assemblers consider optional, e.g., the destination register for
                   11958: @code{br,}, or the destination register and hint for @code{jmp,}.
1.2       jwilke   11959: 
1.78      anton    11960: You can specify conditions for @code{if,} by removing the first @code{b}
                   11961: and the trailing @code{,} from a branch with a corresponding name; e.g.,
1.2       jwilke   11962: 
1.26      crook    11963: @example
1.78      anton    11964: 11 fgt if, \ if F11>0e
                   11965:   ...
                   11966: endif,
1.26      crook    11967: @end example
1.2       jwilke   11968: 
1.78      anton    11969: @code{fbgt,} gives @code{fgt}.  
                   11970: 
                   11971: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
                   11972: @subsection MIPS assembler
1.2       jwilke   11973: 
1.78      anton    11974: The MIPS assembler was originally written by Christian Pirker.
1.2       jwilke   11975: 
1.78      anton    11976: Currently the assembler and disassembler only cover the MIPS-I
                   11977: architecture (R3000), and don't support FP instructions.
1.2       jwilke   11978: 
1.78      anton    11979: The register names @code{$a0}--@code{$a3} are not available to avoid
                   11980: shadowing hex numbers.
1.2       jwilke   11981: 
1.78      anton    11982: Because there is no way to distinguish registers from immediate values,
                   11983: you have to explicitly use the immediate forms of instructions, i.e.,
                   11984: @code{addiu,}, not just @code{addu,} (@command{as} does this
                   11985: implicitly).
1.2       jwilke   11986: 
1.78      anton    11987: If the architecture manual specifies several formats for the instruction
                   11988: (e.g., for @code{jalr,}), you usually have to use the one with more
                   11989: arguments (i.e., two for @code{jalr,}).  When in doubt, see
                   11990: @code{arch/mips/testasm.fs} for an example of correct use.
1.2       jwilke   11991: 
1.78      anton    11992: Branches and jumps in the MIPS architecture have a delay slot.  You have
                   11993: to fill it yourself (the simplest way is to use @code{nop,}), the
                   11994: assembler does not do it for you (unlike @command{as}).  Even
                   11995: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
                   11996: @code{else,} and @code{repeat,} need a delay slot.  Since @code{begin,}
                   11997: and @code{then,} just specify branch targets, they are not affected.
1.2       jwilke   11998: 
1.78      anton    11999: Note that you must not put branches, jumps, or @code{li,} into the delay
                   12000: slot: @code{li,} may expand to several instructions, and control flow
                   12001: instructions may not be put into the branch delay slot in any case.
1.2       jwilke   12002: 
1.78      anton    12003: For branches the argument specifying the target is a relative address;
                   12004: You have to add the address of the delay slot to get the absolute
                   12005: address.
1.1       anton    12006: 
1.78      anton    12007: The MIPS architecture also has load delay slots and restrictions on
                   12008: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
                   12009: yourself to satisfy these restrictions, the assembler does not do it for
                   12010: you.
1.1       anton    12011: 
1.78      anton    12012: You can specify the conditions for @code{if,} etc. by taking a
                   12013: conditional branch and leaving away the @code{b} at the start and the
                   12014: @code{,} at the end.  E.g.,
1.1       anton    12015: 
1.26      crook    12016: @example
1.78      anton    12017: 4 5 eq if,
                   12018:   ... \ do something if $4 equals $5
                   12019: then,
1.26      crook    12020: @end example
1.1       anton    12021: 
1.78      anton    12022: @node Other assemblers,  , MIPS assembler, Assembler and Code Words
                   12023: @subsection Other assemblers
                   12024: 
                   12025: If you want to contribute another assembler/disassembler, please contact
1.103     anton    12026: us (@email{anton@@mips.complang.tuwien.ac.at}) to check if we have such
                   12027: an assembler already.  If you are writing them from scratch, please use
                   12028: a similar syntax style as the one we use (i.e., postfix, commas at the
                   12029: end of the instruction names, @pxref{Common Assembler}); make the output
                   12030: of the disassembler be valid input for the assembler, and keep the style
1.78      anton    12031: similar to the style we used.
                   12032: 
                   12033: Hints on implementation: The most important part is to have a good test
                   12034: suite that contains all instructions.  Once you have that, the rest is
                   12035: easy.  For actual coding you can take a look at
                   12036: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
                   12037: the assembler and disassembler, avoiding redundancy and some potential
                   12038: bugs.  You can also look at that file (and @pxref{Advanced does> usage
                   12039: example}) to get ideas how to factor a disassembler.
                   12040: 
                   12041: Start with the disassembler, because it's easier to reuse data from the
                   12042: disassembler for the assembler than the other way round.
1.1       anton    12043: 
1.78      anton    12044: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
                   12045: how simple it can be.
1.1       anton    12046: 
1.78      anton    12047: @c -------------------------------------------------------------
                   12048: @node Threading Words, Passing Commands to the OS, Assembler and Code Words, Words
                   12049: @section Threading Words
                   12050: @cindex threading words
1.1       anton    12051: 
1.78      anton    12052: @cindex code address
                   12053: These words provide access to code addresses and other threading stuff
                   12054: in Gforth (and, possibly, other interpretive Forths). It more or less
                   12055: abstracts away the differences between direct and indirect threading
                   12056: (and, for direct threading, the machine dependences). However, at
                   12057: present this wordset is still incomplete. It is also pretty low-level;
                   12058: some day it will hopefully be made unnecessary by an internals wordset
                   12059: that abstracts implementation details away completely.
1.1       anton    12060: 
1.78      anton    12061: The terminology used here stems from indirect threaded Forth systems; in
                   12062: such a system, the XT of a word is represented by the CFA (code field
                   12063: address) of a word; the CFA points to a cell that contains the code
                   12064: address.  The code address is the address of some machine code that
                   12065: performs the run-time action of invoking the word (e.g., the
                   12066: @code{dovar:} routine pushes the address of the body of the word (a
                   12067: variable) on the stack
                   12068: ).
1.1       anton    12069: 
1.78      anton    12070: @cindex code address
                   12071: @cindex code field address
                   12072: In an indirect threaded Forth, you can get the code address of @i{name}
                   12073: with @code{' @i{name} @@}; in Gforth you can get it with @code{' @i{name}
                   12074: >code-address}, independent of the threading method.
1.1       anton    12075: 
1.78      anton    12076: doc-threading-method
                   12077: doc->code-address
                   12078: doc-code-address!
1.1       anton    12079: 
1.78      anton    12080: @cindex @code{does>}-handler
                   12081: @cindex @code{does>}-code
                   12082: For a word defined with @code{DOES>}, the code address usually points to
                   12083: a jump instruction (the @dfn{does-handler}) that jumps to the dodoes
                   12084: routine (in Gforth on some platforms, it can also point to the dodoes
                   12085: routine itself).  What you are typically interested in, though, is
                   12086: whether a word is a @code{DOES>}-defined word, and what Forth code it
                   12087: executes; @code{>does-code} tells you that.
1.1       anton    12088: 
1.78      anton    12089: doc->does-code
1.1       anton    12090: 
1.78      anton    12091: To create a @code{DOES>}-defined word with the following basic words,
                   12092: you have to set up a @code{DOES>}-handler with @code{does-handler!};
                   12093: @code{/does-handler} aus behind you have to place your executable Forth
                   12094: code.  Finally you have to create a word and modify its behaviour with
                   12095: @code{does-handler!}.
1.1       anton    12096: 
1.78      anton    12097: doc-does-code!
                   12098: doc-does-handler!
                   12099: doc-/does-handler
1.1       anton    12100: 
1.78      anton    12101: The code addresses produced by various defining words are produced by
                   12102: the following words:
1.1       anton    12103: 
1.78      anton    12104: doc-docol:
                   12105: doc-docon:
                   12106: doc-dovar:
                   12107: doc-douser:
                   12108: doc-dodefer:
                   12109: doc-dofield:
1.1       anton    12110: 
1.99      anton    12111: @cindex definer
                   12112: The following two words generalize @code{>code-address},
                   12113: @code{>does-code}, @code{code-address!}, and @code{does-code!}:
                   12114: 
                   12115: doc->definer
                   12116: doc-definer!
                   12117: 
1.26      crook    12118: @c -------------------------------------------------------------
1.78      anton    12119: @node Passing Commands to the OS, Keeping track of Time, Threading Words, Words
1.21      crook    12120: @section Passing Commands to the Operating System
                   12121: @cindex operating system - passing commands
                   12122: @cindex shell commands
                   12123: 
                   12124: Gforth allows you to pass an arbitrary string to the host operating
                   12125: system shell (if such a thing exists) for execution.
                   12126: 
                   12127: doc-sh
                   12128: doc-system
                   12129: doc-$?
1.23      crook    12130: doc-getenv
1.44      crook    12131: 
1.26      crook    12132: @c -------------------------------------------------------------
1.47      crook    12133: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
                   12134: @section Keeping track of Time
                   12135: @cindex time-related words
                   12136: 
                   12137: doc-ms
                   12138: doc-time&date
1.79      anton    12139: doc-utime
                   12140: doc-cputime
1.47      crook    12141: 
                   12142: 
                   12143: @c -------------------------------------------------------------
                   12144: @node Miscellaneous Words,  , Keeping track of Time, Words
1.21      crook    12145: @section Miscellaneous Words
                   12146: @cindex miscellaneous words
                   12147: 
1.29      crook    12148: @comment TODO find homes for these
                   12149: 
1.26      crook    12150: These section lists the ANS Forth words that are not documented
1.21      crook    12151: elsewhere in this manual. Ultimately, they all need proper homes.
                   12152: 
1.68      anton    12153: doc-quit
1.44      crook    12154: 
1.26      crook    12155: The following ANS Forth words are not currently supported by Gforth 
1.27      crook    12156: (@pxref{ANS conformance}):
1.21      crook    12157: 
                   12158: @code{EDITOR} 
                   12159: @code{EMIT?} 
                   12160: @code{FORGET} 
                   12161: 
1.24      anton    12162: @c ******************************************************************
                   12163: @node Error messages, Tools, Words, Top
                   12164: @chapter Error messages
                   12165: @cindex error messages
                   12166: @cindex backtrace
                   12167: 
                   12168: A typical Gforth error message looks like this:
                   12169: 
                   12170: @example
1.86      anton    12171: in file included from \evaluated string/:-1
1.24      anton    12172: in file included from ./yyy.fs:1
                   12173: ./xxx.fs:4: Invalid memory address
1.134     anton    12174: >>>bar<<<
1.79      anton    12175: Backtrace:
1.25      anton    12176: $400E664C @@
                   12177: $400E6664 foo
1.24      anton    12178: @end example
                   12179: 
                   12180: The message identifying the error is @code{Invalid memory address}.  The
                   12181: error happened when text-interpreting line 4 of the file
                   12182: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
                   12183: word on the line where the error happened, is pointed out (with
1.134     anton    12184: @code{>>>} and @code{<<<}).
1.24      anton    12185: 
                   12186: The file containing the error was included in line 1 of @file{./yyy.fs},
                   12187: and @file{yyy.fs} was included from a non-file (in this case, by giving
                   12188: @file{yyy.fs} as command-line parameter to Gforth).
                   12189: 
                   12190: At the end of the error message you find a return stack dump that can be
                   12191: interpreted as a backtrace (possibly empty). On top you find the top of
                   12192: the return stack when the @code{throw} happened, and at the bottom you
                   12193: find the return stack entry just above the return stack of the topmost
                   12194: text interpreter.
                   12195: 
                   12196: To the right of most return stack entries you see a guess for the word
                   12197: that pushed that return stack entry as its return address. This gives a
                   12198: backtrace. In our case we see that @code{bar} called @code{foo}, and
                   12199: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
                   12200: address} exception).
                   12201: 
                   12202: Note that the backtrace is not perfect: We don't know which return stack
                   12203: entries are return addresses (so we may get false positives); and in
                   12204: some cases (e.g., for @code{abort"}) we cannot determine from the return
                   12205: address the word that pushed the return address, so for some return
                   12206: addresses you see no names in the return stack dump.
1.25      anton    12207: 
                   12208: @cindex @code{catch} and backtraces
                   12209: The return stack dump represents the return stack at the time when a
                   12210: specific @code{throw} was executed.  In programs that make use of
                   12211: @code{catch}, it is not necessarily clear which @code{throw} should be
                   12212: used for the return stack dump (e.g., consider one @code{throw} that
                   12213: indicates an error, which is caught, and during recovery another error
1.42      anton    12214: happens; which @code{throw} should be used for the stack dump?).  Gforth
1.25      anton    12215: presents the return stack dump for the first @code{throw} after the last
                   12216: executed (not returned-to) @code{catch}; this works well in the usual
                   12217: case.
                   12218: 
                   12219: @cindex @code{gforth-fast} and backtraces
                   12220: @cindex @code{gforth-fast}, difference from @code{gforth}
                   12221: @cindex backtraces with @code{gforth-fast}
                   12222: @cindex return stack dump with @code{gforth-fast}
1.79      anton    12223: @code{Gforth} is able to do a return stack dump for throws generated
1.25      anton    12224: from primitives (e.g., invalid memory address, stack empty etc.);
                   12225: @code{gforth-fast} is only able to do a return stack dump from a
1.96      anton    12226: directly called @code{throw} (including @code{abort} etc.).  Given an
1.30      anton    12227: exception caused by a primitive in @code{gforth-fast}, you will
                   12228: typically see no return stack dump at all; however, if the exception is
                   12229: caught by @code{catch} (e.g., for restoring some state), and then
                   12230: @code{throw}n again, the return stack dump will be for the first such
                   12231: @code{throw}.
1.2       jwilke   12232: 
1.5       anton    12233: @c ******************************************************************
1.24      anton    12234: @node Tools, ANS conformance, Error messages, Top
1.1       anton    12235: @chapter Tools
                   12236: 
                   12237: @menu
                   12238: * ANS Report::                  Report the words used, sorted by wordset.
1.127     anton    12239: * Stack depth changes::         Where does this stack item come from?
1.1       anton    12240: @end menu
                   12241: 
                   12242: See also @ref{Emacs and Gforth}.
                   12243: 
1.126     pazsan   12244: @node ANS Report, Stack depth changes, Tools, Tools
1.1       anton    12245: @section @file{ans-report.fs}: Report the words used, sorted by wordset
                   12246: @cindex @file{ans-report.fs}
                   12247: @cindex report the words used in your program
                   12248: @cindex words used in your program
                   12249: 
                   12250: If you want to label a Forth program as ANS Forth Program, you must
                   12251: document which wordsets the program uses; for extension wordsets, it is
                   12252: helpful to list the words the program requires from these wordsets
                   12253: (because Forth systems are allowed to provide only some words of them).
                   12254: 
                   12255: The @file{ans-report.fs} tool makes it easy for you to determine which
                   12256: words from which wordset and which non-ANS words your application
                   12257: uses. You simply have to include @file{ans-report.fs} before loading the
                   12258: program you want to check. After loading your program, you can get the
                   12259: report with @code{print-ans-report}. A typical use is to run this as
                   12260: batch job like this:
                   12261: @example
                   12262: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
                   12263: @end example
                   12264: 
                   12265: The output looks like this (for @file{compat/control.fs}):
                   12266: @example
                   12267: The program uses the following words
                   12268: from CORE :
                   12269: : POSTPONE THEN ; immediate ?dup IF 0= 
                   12270: from BLOCK-EXT :
                   12271: \ 
                   12272: from FILE :
                   12273: ( 
                   12274: @end example
                   12275: 
                   12276: @subsection Caveats
                   12277: 
                   12278: Note that @file{ans-report.fs} just checks which words are used, not whether
                   12279: they are used in an ANS Forth conforming way!
                   12280: 
                   12281: Some words are defined in several wordsets in the
                   12282: standard. @file{ans-report.fs} reports them for only one of the
                   12283: wordsets, and not necessarily the one you expect. It depends on usage
                   12284: which wordset is the right one to specify. E.g., if you only use the
                   12285: compilation semantics of @code{S"}, it is a Core word; if you also use
                   12286: its interpretation semantics, it is a File word.
1.124     anton    12287: 
                   12288: 
1.127     anton    12289: @node Stack depth changes,  , ANS Report, Tools
1.124     anton    12290: @section Stack depth changes during interpretation
                   12291: @cindex @file{depth-changes.fs}
                   12292: @cindex depth changes during interpretation
                   12293: @cindex stack depth changes during interpretation
                   12294: @cindex items on the stack after interpretation
                   12295: 
                   12296: Sometimes you notice that, after loading a file, there are items left
                   12297: on the stack.  The tool @file{depth-changes.fs} helps you find out
                   12298: quickly where in the file these stack items are coming from.
                   12299: 
                   12300: The simplest way of using @file{depth-changes.fs} is to include it
                   12301: before the file(s) you want to check, e.g.:
                   12302: 
                   12303: @example
                   12304: gforth depth-changes.fs my-file.fs
                   12305: @end example
                   12306: 
                   12307: This will compare the stack depths of the data and FP stack at every
                   12308: empty line (in interpretation state) against these depths at the last
                   12309: empty line (in interpretation state).  If the depths are not equal,
                   12310: the position in the file and the stack contents are printed with
                   12311: @code{~~} (@pxref{Debugging}).  This indicates that a stack depth
                   12312: change has occured in the paragraph of non-empty lines before the
                   12313: indicated line.  It is a good idea to leave an empty line at the end
                   12314: of the file, so the last paragraph is checked, too.
                   12315: 
                   12316: Checking only at empty lines usually works well, but sometimes you
                   12317: have big blocks of non-empty lines (e.g., when building a big table),
                   12318: and you want to know where in this block the stack depth changed.  You
                   12319: can check all interpreted lines with
                   12320: 
                   12321: @example
                   12322: gforth depth-changes.fs -e "' all-lines is depth-changes-filter" my-file.fs
                   12323: @end example
                   12324: 
                   12325: This checks the stack depth at every end-of-line.  So the depth change
                   12326: occured in the line reported by the @code{~~} (not in the line
                   12327: before).
                   12328: 
                   12329: Note that, while this offers better accuracy in indicating where the
                   12330: stack depth changes, it will often report many intentional stack depth
                   12331: changes (e.g., when an interpreted computation stretches across
                   12332: several lines).  You can suppress the checking of some lines by
                   12333: putting backslashes at the end of these lines (not followed by white
                   12334: space), and using
                   12335: 
                   12336: @example
                   12337: gforth depth-changes.fs -e "' most-lines is depth-changes-filter" my-file.fs
                   12338: @end example
1.1       anton    12339: 
                   12340: @c ******************************************************************
1.65      anton    12341: @node ANS conformance, Standard vs Extensions, Tools, Top
1.1       anton    12342: @chapter ANS conformance
                   12343: @cindex ANS conformance of Gforth
                   12344: 
                   12345: To the best of our knowledge, Gforth is an
                   12346: 
                   12347: ANS Forth System
                   12348: @itemize @bullet
                   12349: @item providing the Core Extensions word set
                   12350: @item providing the Block word set
                   12351: @item providing the Block Extensions word set
                   12352: @item providing the Double-Number word set
                   12353: @item providing the Double-Number Extensions word set
                   12354: @item providing the Exception word set
                   12355: @item providing the Exception Extensions word set
                   12356: @item providing the Facility word set
1.40      anton    12357: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
1.1       anton    12358: @item providing the File Access word set
                   12359: @item providing the File Access Extensions word set
                   12360: @item providing the Floating-Point word set
                   12361: @item providing the Floating-Point Extensions word set
                   12362: @item providing the Locals word set
                   12363: @item providing the Locals Extensions word set
                   12364: @item providing the Memory-Allocation word set
                   12365: @item providing the Memory-Allocation Extensions word set (that one's easy)
                   12366: @item providing the Programming-Tools word set
                   12367: @item providing @code{;CODE}, @code{AHEAD}, @code{ASSEMBLER}, @code{BYE}, @code{CODE}, @code{CS-PICK}, @code{CS-ROLL}, @code{STATE}, @code{[ELSE]}, @code{[IF]}, @code{[THEN]} from the Programming-Tools Extensions word set
                   12368: @item providing the Search-Order word set
                   12369: @item providing the Search-Order Extensions word set
                   12370: @item providing the String word set
                   12371: @item providing the String Extensions word set (another easy one)
                   12372: @end itemize
                   12373: 
1.118     anton    12374: Gforth has the following environmental restrictions:
                   12375: 
                   12376: @cindex environmental restrictions
                   12377: @itemize @bullet
                   12378: @item
                   12379: While processing the OS command line, if an exception is not caught,
                   12380: Gforth exits with a non-zero exit code instyead of performing QUIT.
                   12381: 
                   12382: @item
                   12383: When an @code{throw} is performed after a @code{query}, Gforth does not
                   12384: allways restore the input source specification in effect at the
                   12385: corresponding catch.
                   12386: 
                   12387: @end itemize
                   12388: 
                   12389: 
1.1       anton    12390: @cindex system documentation
                   12391: In addition, ANS Forth systems are required to document certain
                   12392: implementation choices. This chapter tries to meet these
                   12393: requirements. In many cases it gives a way to ask the system for the
                   12394: information instead of providing the information directly, in
                   12395: particular, if the information depends on the processor, the operating
                   12396: system or the installation options chosen, or if they are likely to
                   12397: change during the maintenance of Gforth.
                   12398: 
                   12399: @comment The framework for the rest has been taken from pfe.
                   12400: 
                   12401: @menu
                   12402: * The Core Words::              
                   12403: * The optional Block word set::  
                   12404: * The optional Double Number word set::  
                   12405: * The optional Exception word set::  
                   12406: * The optional Facility word set::  
                   12407: * The optional File-Access word set::  
                   12408: * The optional Floating-Point word set::  
                   12409: * The optional Locals word set::  
                   12410: * The optional Memory-Allocation word set::  
                   12411: * The optional Programming-Tools word set::  
                   12412: * The optional Search-Order word set::  
                   12413: @end menu
                   12414: 
                   12415: 
                   12416: @c =====================================================================
                   12417: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
                   12418: @comment  node-name,  next,  previous,  up
                   12419: @section The Core Words
                   12420: @c =====================================================================
                   12421: @cindex core words, system documentation
                   12422: @cindex system documentation, core words
                   12423: 
                   12424: @menu
                   12425: * core-idef::                   Implementation Defined Options                   
                   12426: * core-ambcond::                Ambiguous Conditions                
                   12427: * core-other::                  Other System Documentation                  
                   12428: @end menu
                   12429: 
                   12430: @c ---------------------------------------------------------------------
                   12431: @node core-idef, core-ambcond, The Core Words, The Core Words
                   12432: @subsection Implementation Defined Options
                   12433: @c ---------------------------------------------------------------------
                   12434: @cindex core words, implementation-defined options
                   12435: @cindex implementation-defined options, core words
                   12436: 
                   12437: 
                   12438: @table @i
                   12439: @item (Cell) aligned addresses:
                   12440: @cindex cell-aligned addresses
                   12441: @cindex aligned addresses
                   12442: processor-dependent. Gforth's alignment words perform natural alignment
                   12443: (e.g., an address aligned for a datum of size 8 is divisible by
                   12444: 8). Unaligned accesses usually result in a @code{-23 THROW}.
                   12445: 
                   12446: @item @code{EMIT} and non-graphic characters:
                   12447: @cindex @code{EMIT} and non-graphic characters
                   12448: @cindex non-graphic characters and @code{EMIT}
                   12449: The character is output using the C library function (actually, macro)
                   12450: @code{putc}.
                   12451: 
                   12452: @item character editing of @code{ACCEPT} and @code{EXPECT}:
                   12453: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
                   12454: @cindex editing in @code{ACCEPT} and @code{EXPECT}
                   12455: @cindex @code{ACCEPT}, editing
                   12456: @cindex @code{EXPECT}, editing
                   12457: This is modeled on the GNU readline library (@pxref{Readline
                   12458: Interaction, , Command Line Editing, readline, The GNU Readline
                   12459: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
                   12460: producing a full word completion every time you type it (instead of
1.28      crook    12461: producing the common prefix of all completions). @xref{Command-line editing}.
1.1       anton    12462: 
                   12463: @item character set:
                   12464: @cindex character set
                   12465: The character set of your computer and display device. Gforth is
                   12466: 8-bit-clean (but some other component in your system may make trouble).
                   12467: 
                   12468: @item Character-aligned address requirements:
                   12469: @cindex character-aligned address requirements
                   12470: installation-dependent. Currently a character is represented by a C
                   12471: @code{unsigned char}; in the future we might switch to @code{wchar_t}
                   12472: (Comments on that requested).
                   12473: 
                   12474: @item character-set extensions and matching of names:
                   12475: @cindex character-set extensions and matching of names
1.26      crook    12476: @cindex case-sensitivity for name lookup
                   12477: @cindex name lookup, case-sensitivity
                   12478: @cindex locale and case-sensitivity
1.21      crook    12479: Any character except the ASCII NUL character can be used in a
1.1       anton    12480: name. Matching is case-insensitive (except in @code{TABLE}s). The
1.47      crook    12481: matching is performed using the C library function @code{strncasecmp}, whose
1.1       anton    12482: function is probably influenced by the locale. E.g., the @code{C} locale
                   12483: does not know about accents and umlauts, so they are matched
                   12484: case-sensitively in that locale. For portability reasons it is best to
                   12485: write programs such that they work in the @code{C} locale. Then one can
                   12486: use libraries written by a Polish programmer (who might use words
                   12487: containing ISO Latin-2 encoded characters) and by a French programmer
                   12488: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
                   12489: funny results for some of the words (which ones, depends on the font you
                   12490: are using)). Also, the locale you prefer may not be available in other
                   12491: operating systems. Hopefully, Unicode will solve these problems one day.
                   12492: 
                   12493: @item conditions under which control characters match a space delimiter:
                   12494: @cindex space delimiters
                   12495: @cindex control characters as delimiters
1.117     anton    12496: If @code{word} is called with the space character as a delimiter, all
1.1       anton    12497: white-space characters (as identified by the C macro @code{isspace()})
1.117     anton    12498: are delimiters. @code{Parse}, on the other hand, treats space like other
1.138     anton    12499: delimiters.  @code{Parse-name}, which is used by the outer
1.1       anton    12500: interpreter (aka text interpreter) by default, treats all white-space
                   12501: characters as delimiters.
                   12502: 
1.26      crook    12503: @item format of the control-flow stack:
                   12504: @cindex control-flow stack, format
                   12505: The data stack is used as control-flow stack. The size of a control-flow
1.1       anton    12506: stack item in cells is given by the constant @code{cs-item-size}. At the
                   12507: time of this writing, an item consists of a (pointer to a) locals list
                   12508: (third), an address in the code (second), and a tag for identifying the
                   12509: item (TOS). The following tags are used: @code{defstart},
                   12510: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
                   12511: @code{scopestart}.
                   12512: 
                   12513: @item conversion of digits > 35
                   12514: @cindex digits > 35
                   12515: The characters @code{[\]^_'} are the digits with the decimal value
                   12516: 36@minus{}41. There is no way to input many of the larger digits.
                   12517: 
                   12518: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
                   12519: @cindex @code{EXPECT}, display after end of input
                   12520: @cindex @code{ACCEPT}, display after end of input
                   12521: The cursor is moved to the end of the entered string. If the input is
                   12522: terminated using the @kbd{Return} key, a space is typed.
                   12523: 
                   12524: @item exception abort sequence of @code{ABORT"}:
                   12525: @cindex exception abort sequence of @code{ABORT"}
                   12526: @cindex @code{ABORT"}, exception abort sequence
                   12527: The error string is stored into the variable @code{"error} and a
                   12528: @code{-2 throw} is performed.
                   12529: 
                   12530: @item input line terminator:
                   12531: @cindex input line terminator
                   12532: @cindex line terminator on input
1.26      crook    12533: @cindex newline character on input
1.1       anton    12534: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
                   12535: lines. One of these characters is typically produced when you type the
                   12536: @kbd{Enter} or @kbd{Return} key.
                   12537: 
                   12538: @item maximum size of a counted string:
                   12539: @cindex maximum size of a counted string
                   12540: @cindex counted string, maximum size
                   12541: @code{s" /counted-string" environment? drop .}. Currently 255 characters
1.79      anton    12542: on all platforms, but this may change.
1.1       anton    12543: 
                   12544: @item maximum size of a parsed string:
                   12545: @cindex maximum size of a parsed string
                   12546: @cindex parsed string, maximum size
                   12547: Given by the constant @code{/line}. Currently 255 characters.
                   12548: 
                   12549: @item maximum size of a definition name, in characters:
                   12550: @cindex maximum size of a definition name, in characters
                   12551: @cindex name, maximum length
1.113     anton    12552: MAXU/8
1.1       anton    12553: 
                   12554: @item maximum string length for @code{ENVIRONMENT?}, in characters:
                   12555: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
                   12556: @cindex @code{ENVIRONMENT?} string length, maximum
1.113     anton    12557: MAXU/8
1.1       anton    12558: 
                   12559: @item method of selecting the user input device:
                   12560: @cindex user input device, method of selecting
                   12561: The user input device is the standard input. There is currently no way to
                   12562: change it from within Gforth. However, the input can typically be
                   12563: redirected in the command line that starts Gforth.
                   12564: 
                   12565: @item method of selecting the user output device:
                   12566: @cindex user output device, method of selecting
                   12567: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
1.10      anton    12568: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
                   12569: output when the user output device is a terminal, otherwise the output
                   12570: is buffered.
1.1       anton    12571: 
                   12572: @item methods of dictionary compilation:
                   12573: What are we expected to document here?
                   12574: 
                   12575: @item number of bits in one address unit:
                   12576: @cindex number of bits in one address unit
                   12577: @cindex address unit, size in bits
                   12578: @code{s" address-units-bits" environment? drop .}. 8 in all current
1.79      anton    12579: platforms.
1.1       anton    12580: 
                   12581: @item number representation and arithmetic:
                   12582: @cindex number representation and arithmetic
1.79      anton    12583: Processor-dependent. Binary two's complement on all current platforms.
1.1       anton    12584: 
                   12585: @item ranges for integer types:
                   12586: @cindex ranges for integer types
                   12587: @cindex integer types, ranges
                   12588: Installation-dependent. Make environmental queries for @code{MAX-N},
                   12589: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
                   12590: unsigned (and positive) types is 0. The lower bound for signed types on
                   12591: two's complement and one's complement machines machines can be computed
                   12592: by adding 1 to the upper bound.
                   12593: 
                   12594: @item read-only data space regions:
                   12595: @cindex read-only data space regions
                   12596: @cindex data-space, read-only regions
                   12597: The whole Forth data space is writable.
                   12598: 
                   12599: @item size of buffer at @code{WORD}:
                   12600: @cindex size of buffer at @code{WORD}
                   12601: @cindex @code{WORD} buffer size
                   12602: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
                   12603: shared with the pictured numeric output string. If overwriting
                   12604: @code{PAD} is acceptable, it is as large as the remaining dictionary
                   12605: space, although only as much can be sensibly used as fits in a counted
                   12606: string.
                   12607: 
                   12608: @item size of one cell in address units:
                   12609: @cindex cell size
                   12610: @code{1 cells .}.
                   12611: 
                   12612: @item size of one character in address units:
                   12613: @cindex char size
1.79      anton    12614: @code{1 chars .}. 1 on all current platforms.
1.1       anton    12615: 
                   12616: @item size of the keyboard terminal buffer:
                   12617: @cindex size of the keyboard terminal buffer
                   12618: @cindex terminal buffer, size
                   12619: Varies. You can determine the size at a specific time using @code{lp@@
                   12620: tib - .}. It is shared with the locals stack and TIBs of files that
                   12621: include the current file. You can change the amount of space for TIBs
                   12622: and locals stack at Gforth startup with the command line option
                   12623: @code{-l}.
                   12624: 
                   12625: @item size of the pictured numeric output buffer:
                   12626: @cindex size of the pictured numeric output buffer
                   12627: @cindex pictured numeric output buffer, size
                   12628: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
                   12629: shared with @code{WORD}.
                   12630: 
                   12631: @item size of the scratch area returned by @code{PAD}:
                   12632: @cindex size of the scratch area returned by @code{PAD}
                   12633: @cindex @code{PAD} size
                   12634: The remainder of dictionary space. @code{unused pad here - - .}.
                   12635: 
                   12636: @item system case-sensitivity characteristics:
                   12637: @cindex case-sensitivity characteristics
1.26      crook    12638: Dictionary searches are case-insensitive (except in
1.1       anton    12639: @code{TABLE}s). However, as explained above under @i{character-set
                   12640: extensions}, the matching for non-ASCII characters is determined by the
                   12641: locale you are using. In the default @code{C} locale all non-ASCII
                   12642: characters are matched case-sensitively.
                   12643: 
                   12644: @item system prompt:
                   12645: @cindex system prompt
                   12646: @cindex prompt
                   12647: @code{ ok} in interpret state, @code{ compiled} in compile state.
                   12648: 
                   12649: @item division rounding:
                   12650: @cindex division rounding
                   12651: installation dependent. @code{s" floored" environment? drop .}. We leave
                   12652: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
                   12653: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
                   12654: 
                   12655: @item values of @code{STATE} when true:
                   12656: @cindex @code{STATE} values
                   12657: -1.
                   12658: 
                   12659: @item values returned after arithmetic overflow:
                   12660: On two's complement machines, arithmetic is performed modulo
                   12661: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
                   12662: arithmetic (with appropriate mapping for signed types). Division by zero
                   12663: typically results in a @code{-55 throw} (Floating-point unidentified
1.80      anton    12664: fault) or @code{-10 throw} (divide by zero).
1.1       anton    12665: 
                   12666: @item whether the current definition can be found after @t{DOES>}:
                   12667: @cindex @t{DOES>}, visibility of current definition
                   12668: No.
                   12669: 
                   12670: @end table
                   12671: 
                   12672: @c ---------------------------------------------------------------------
                   12673: @node core-ambcond, core-other, core-idef, The Core Words
                   12674: @subsection Ambiguous conditions
                   12675: @c ---------------------------------------------------------------------
                   12676: @cindex core words, ambiguous conditions
                   12677: @cindex ambiguous conditions, core words
                   12678: 
                   12679: @table @i
                   12680: 
                   12681: @item a name is neither a word nor a number:
                   12682: @cindex name not found
1.26      crook    12683: @cindex undefined word
1.80      anton    12684: @code{-13 throw} (Undefined word).
1.1       anton    12685: 
                   12686: @item a definition name exceeds the maximum length allowed:
1.26      crook    12687: @cindex word name too long
1.1       anton    12688: @code{-19 throw} (Word name too long)
                   12689: 
                   12690: @item addressing a region not inside the various data spaces of the forth system:
                   12691: @cindex Invalid memory address
1.32      anton    12692: The stacks, code space and header space are accessible. Machine code space is
1.1       anton    12693: typically readable. Accessing other addresses gives results dependent on
                   12694: the operating system. On decent systems: @code{-9 throw} (Invalid memory
                   12695: address).
                   12696: 
                   12697: @item argument type incompatible with parameter:
1.26      crook    12698: @cindex argument type mismatch
1.1       anton    12699: This is usually not caught. Some words perform checks, e.g., the control
                   12700: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
                   12701: mismatch).
                   12702: 
                   12703: @item attempting to obtain the execution token of a word with undefined execution semantics:
                   12704: @cindex Interpreting a compile-only word, for @code{'} etc.
                   12705: @cindex execution token of words with undefined execution semantics
                   12706: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
                   12707: get an execution token for @code{compile-only-error} (which performs a
                   12708: @code{-14 throw} when executed).
                   12709: 
                   12710: @item dividing by zero:
                   12711: @cindex dividing by zero
                   12712: @cindex floating point unidentified fault, integer division
1.80      anton    12713: On some platforms, this produces a @code{-10 throw} (Division by
1.24      anton    12714: zero); on other systems, this typically results in a @code{-55 throw}
                   12715: (Floating-point unidentified fault).
1.1       anton    12716: 
                   12717: @item insufficient data stack or return stack space:
                   12718: @cindex insufficient data stack or return stack space
                   12719: @cindex stack overflow
1.26      crook    12720: @cindex address alignment exception, stack overflow
1.1       anton    12721: @cindex Invalid memory address, stack overflow
                   12722: Depending on the operating system, the installation, and the invocation
                   12723: of Gforth, this is either checked by the memory management hardware, or
1.24      anton    12724: it is not checked. If it is checked, you typically get a @code{-3 throw}
                   12725: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
                   12726: throw} (Invalid memory address) (depending on the platform and how you
                   12727: achieved the overflow) as soon as the overflow happens. If it is not
                   12728: checked, overflows typically result in mysterious illegal memory
                   12729: accesses, producing @code{-9 throw} (Invalid memory address) or
                   12730: @code{-23 throw} (Address alignment exception); they might also destroy
                   12731: the internal data structure of @code{ALLOCATE} and friends, resulting in
                   12732: various errors in these words.
1.1       anton    12733: 
                   12734: @item insufficient space for loop control parameters:
                   12735: @cindex insufficient space for loop control parameters
1.80      anton    12736: Like other return stack overflows.
1.1       anton    12737: 
                   12738: @item insufficient space in the dictionary:
                   12739: @cindex insufficient space in the dictionary
                   12740: @cindex dictionary overflow
1.12      anton    12741: If you try to allot (either directly with @code{allot}, or indirectly
                   12742: with @code{,}, @code{create} etc.) more memory than available in the
                   12743: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
                   12744: to access memory beyond the end of the dictionary, the results are
                   12745: similar to stack overflows.
1.1       anton    12746: 
                   12747: @item interpreting a word with undefined interpretation semantics:
                   12748: @cindex interpreting a word with undefined interpretation semantics
                   12749: @cindex Interpreting a compile-only word
                   12750: For some words, we have defined interpretation semantics. For the
                   12751: others: @code{-14 throw} (Interpreting a compile-only word).
                   12752: 
                   12753: @item modifying the contents of the input buffer or a string literal:
                   12754: @cindex modifying the contents of the input buffer or a string literal
                   12755: These are located in writable memory and can be modified.
                   12756: 
                   12757: @item overflow of the pictured numeric output string:
                   12758: @cindex overflow of the pictured numeric output string
                   12759: @cindex pictured numeric output string, overflow
1.24      anton    12760: @code{-17 throw} (Pictured numeric ouput string overflow).
1.1       anton    12761: 
                   12762: @item parsed string overflow:
                   12763: @cindex parsed string overflow
                   12764: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
                   12765: 
                   12766: @item producing a result out of range:
                   12767: @cindex result out of range
                   12768: On two's complement machines, arithmetic is performed modulo
                   12769: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
                   12770: arithmetic (with appropriate mapping for signed types). Division by zero
1.24      anton    12771: typically results in a @code{-10 throw} (divide by zero) or @code{-55
                   12772: throw} (floating point unidentified fault). @code{convert} and
                   12773: @code{>number} currently overflow silently.
1.1       anton    12774: 
                   12775: @item reading from an empty data or return stack:
                   12776: @cindex stack empty
                   12777: @cindex stack underflow
1.24      anton    12778: @cindex return stack underflow
1.1       anton    12779: The data stack is checked by the outer (aka text) interpreter after
                   12780: every word executed. If it has underflowed, a @code{-4 throw} (Stack
                   12781: underflow) is performed. Apart from that, stacks may be checked or not,
1.24      anton    12782: depending on operating system, installation, and invocation. If they are
                   12783: caught by a check, they typically result in @code{-4 throw} (Stack
                   12784: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
                   12785: (Invalid memory address), depending on the platform and which stack
                   12786: underflows and by how much. Note that even if the system uses checking
                   12787: (through the MMU), your program may have to underflow by a significant
                   12788: number of stack items to trigger the reaction (the reason for this is
                   12789: that the MMU, and therefore the checking, works with a page-size
                   12790: granularity).  If there is no checking, the symptoms resulting from an
                   12791: underflow are similar to those from an overflow.  Unbalanced return
1.80      anton    12792: stack errors can result in a variety of symptoms, including @code{-9 throw}
1.24      anton    12793: (Invalid memory address) and Illegal Instruction (typically @code{-260
                   12794: throw}).
1.1       anton    12795: 
                   12796: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
                   12797: @cindex unexpected end of the input buffer
                   12798: @cindex zero-length string as a name
                   12799: @cindex Attempt to use zero-length string as a name
                   12800: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
                   12801: use zero-length string as a name). Words like @code{'} probably will not
                   12802: find what they search. Note that it is possible to create zero-length
                   12803: names with @code{nextname} (should it not?).
                   12804: 
                   12805: @item @code{>IN} greater than input buffer:
                   12806: @cindex @code{>IN} greater than input buffer
                   12807: The next invocation of a parsing word returns a string with length 0.
                   12808: 
                   12809: @item @code{RECURSE} appears after @code{DOES>}:
                   12810: @cindex @code{RECURSE} appears after @code{DOES>}
                   12811: Compiles a recursive call to the defining word, not to the defined word.
                   12812: 
                   12813: @item argument input source different than current input source for @code{RESTORE-INPUT}:
                   12814: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
1.26      crook    12815: @cindex argument type mismatch, @code{RESTORE-INPUT}
1.1       anton    12816: @cindex @code{RESTORE-INPUT}, Argument type mismatch
                   12817: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
                   12818: the end of the file was reached), its source-id may be
                   12819: reused. Therefore, restoring an input source specification referencing a
                   12820: closed file may lead to unpredictable results instead of a @code{-12
                   12821: THROW}.
                   12822: 
                   12823: In the future, Gforth may be able to restore input source specifications
                   12824: from other than the current input source.
                   12825: 
                   12826: @item data space containing definitions gets de-allocated:
                   12827: @cindex data space containing definitions gets de-allocated
                   12828: Deallocation with @code{allot} is not checked. This typically results in
                   12829: memory access faults or execution of illegal instructions.
                   12830: 
                   12831: @item data space read/write with incorrect alignment:
                   12832: @cindex data space read/write with incorrect alignment
                   12833: @cindex alignment faults
1.26      crook    12834: @cindex address alignment exception
1.1       anton    12835: Processor-dependent. Typically results in a @code{-23 throw} (Address
1.12      anton    12836: alignment exception). Under Linux-Intel on a 486 or later processor with
1.1       anton    12837: alignment turned on, incorrect alignment results in a @code{-9 throw}
                   12838: (Invalid memory address). There are reportedly some processors with
1.12      anton    12839: alignment restrictions that do not report violations.
1.1       anton    12840: 
                   12841: @item data space pointer not properly aligned, @code{,}, @code{C,}:
                   12842: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
                   12843: Like other alignment errors.
                   12844: 
                   12845: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
                   12846: Like other stack underflows.
                   12847: 
                   12848: @item loop control parameters not available:
                   12849: @cindex loop control parameters not available
                   12850: Not checked. The counted loop words simply assume that the top of return
                   12851: stack items are loop control parameters and behave accordingly.
                   12852: 
                   12853: @item most recent definition does not have a name (@code{IMMEDIATE}):
                   12854: @cindex most recent definition does not have a name (@code{IMMEDIATE})
                   12855: @cindex last word was headerless
                   12856: @code{abort" last word was headerless"}.
                   12857: 
                   12858: @item name not defined by @code{VALUE} used by @code{TO}:
                   12859: @cindex name not defined by @code{VALUE} used by @code{TO}
                   12860: @cindex @code{TO} on non-@code{VALUE}s
                   12861: @cindex Invalid name argument, @code{TO}
                   12862: @code{-32 throw} (Invalid name argument) (unless name is a local or was
                   12863: defined by @code{CONSTANT}; in the latter case it just changes the constant).
                   12864: 
                   12865: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
                   12866: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
1.26      crook    12867: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
1.1       anton    12868: @code{-13 throw} (Undefined word)
                   12869: 
                   12870: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
                   12871: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
                   12872: Gforth behaves as if they were of the same type. I.e., you can predict
                   12873: the behaviour by interpreting all parameters as, e.g., signed.
                   12874: 
                   12875: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
                   12876: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
                   12877: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
                   12878: compilation semantics of @code{TO}.
                   12879: 
                   12880: @item String longer than a counted string returned by @code{WORD}:
1.26      crook    12881: @cindex string longer than a counted string returned by @code{WORD}
1.1       anton    12882: @cindex @code{WORD}, string overflow
                   12883: Not checked. The string will be ok, but the count will, of course,
                   12884: contain only the least significant bits of the length.
                   12885: 
                   12886: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
                   12887: @cindex @code{LSHIFT}, large shift counts
                   12888: @cindex @code{RSHIFT}, large shift counts
                   12889: Processor-dependent. Typical behaviours are returning 0 and using only
                   12890: the low bits of the shift count.
                   12891: 
                   12892: @item word not defined via @code{CREATE}:
                   12893: @cindex @code{>BODY} of non-@code{CREATE}d words
                   12894: @code{>BODY} produces the PFA of the word no matter how it was defined.
                   12895: 
                   12896: @cindex @code{DOES>} of non-@code{CREATE}d words
                   12897: @code{DOES>} changes the execution semantics of the last defined word no
                   12898: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
                   12899: @code{CREATE , DOES>}.
                   12900: 
                   12901: @item words improperly used outside @code{<#} and @code{#>}:
                   12902: Not checked. As usual, you can expect memory faults.
                   12903: 
                   12904: @end table
                   12905: 
                   12906: 
                   12907: @c ---------------------------------------------------------------------
                   12908: @node core-other,  , core-ambcond, The Core Words
                   12909: @subsection Other system documentation
                   12910: @c ---------------------------------------------------------------------
                   12911: @cindex other system documentation, core words
                   12912: @cindex core words, other system documentation
                   12913: 
                   12914: @table @i
                   12915: @item nonstandard words using @code{PAD}:
                   12916: @cindex @code{PAD} use by nonstandard words
                   12917: None.
                   12918: 
                   12919: @item operator's terminal facilities available:
                   12920: @cindex operator's terminal facilities available
1.80      anton    12921: After processing the OS's command line, Gforth goes into interactive mode,
1.1       anton    12922: and you can give commands to Gforth interactively. The actual facilities
                   12923: available depend on how you invoke Gforth.
                   12924: 
                   12925: @item program data space available:
                   12926: @cindex program data space available
                   12927: @cindex data space available
                   12928: @code{UNUSED .} gives the remaining dictionary space. The total
                   12929: dictionary space can be specified with the @code{-m} switch
                   12930: (@pxref{Invoking Gforth}) when Gforth starts up.
                   12931: 
                   12932: @item return stack space available:
                   12933: @cindex return stack space available
                   12934: You can compute the total return stack space in cells with
                   12935: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
                   12936: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
                   12937: 
                   12938: @item stack space available:
                   12939: @cindex stack space available
                   12940: You can compute the total data stack space in cells with
                   12941: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
                   12942: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
                   12943: 
                   12944: @item system dictionary space required, in address units:
                   12945: @cindex system dictionary space required, in address units
                   12946: Type @code{here forthstart - .} after startup. At the time of this
                   12947: writing, this gives 80080 (bytes) on a 32-bit system.
                   12948: @end table
                   12949: 
                   12950: 
                   12951: @c =====================================================================
                   12952: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
                   12953: @section The optional Block word set
                   12954: @c =====================================================================
                   12955: @cindex system documentation, block words
                   12956: @cindex block words, system documentation
                   12957: 
                   12958: @menu
                   12959: * block-idef::                  Implementation Defined Options
                   12960: * block-ambcond::               Ambiguous Conditions               
                   12961: * block-other::                 Other System Documentation                 
                   12962: @end menu
                   12963: 
                   12964: 
                   12965: @c ---------------------------------------------------------------------
                   12966: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
                   12967: @subsection Implementation Defined Options
                   12968: @c ---------------------------------------------------------------------
                   12969: @cindex implementation-defined options, block words
                   12970: @cindex block words, implementation-defined options
                   12971: 
                   12972: @table @i
                   12973: @item the format for display by @code{LIST}:
                   12974: @cindex @code{LIST} display format
                   12975: First the screen number is displayed, then 16 lines of 64 characters,
                   12976: each line preceded by the line number.
                   12977: 
                   12978: @item the length of a line affected by @code{\}:
                   12979: @cindex length of a line affected by @code{\}
                   12980: @cindex @code{\}, line length in blocks
                   12981: 64 characters.
                   12982: @end table
                   12983: 
                   12984: 
                   12985: @c ---------------------------------------------------------------------
                   12986: @node block-ambcond, block-other, block-idef, The optional Block word set
                   12987: @subsection Ambiguous conditions
                   12988: @c ---------------------------------------------------------------------
                   12989: @cindex block words, ambiguous conditions
                   12990: @cindex ambiguous conditions, block words
                   12991: 
                   12992: @table @i
                   12993: @item correct block read was not possible:
                   12994: @cindex block read not possible
                   12995: Typically results in a @code{throw} of some OS-derived value (between
                   12996: -512 and -2048). If the blocks file was just not long enough, blanks are
                   12997: supplied for the missing portion.
                   12998: 
                   12999: @item I/O exception in block transfer:
                   13000: @cindex I/O exception in block transfer
                   13001: @cindex block transfer, I/O exception
                   13002: Typically results in a @code{throw} of some OS-derived value (between
                   13003: -512 and -2048).
                   13004: 
                   13005: @item invalid block number:
                   13006: @cindex invalid block number
                   13007: @cindex block number invalid
                   13008: @code{-35 throw} (Invalid block number)
                   13009: 
                   13010: @item a program directly alters the contents of @code{BLK}:
                   13011: @cindex @code{BLK}, altering @code{BLK}
                   13012: The input stream is switched to that other block, at the same
                   13013: position. If the storing to @code{BLK} happens when interpreting
                   13014: non-block input, the system will get quite confused when the block ends.
                   13015: 
                   13016: @item no current block buffer for @code{UPDATE}:
                   13017: @cindex @code{UPDATE}, no current block buffer
                   13018: @code{UPDATE} has no effect.
                   13019: 
                   13020: @end table
                   13021: 
                   13022: @c ---------------------------------------------------------------------
                   13023: @node block-other,  , block-ambcond, The optional Block word set
                   13024: @subsection Other system documentation
                   13025: @c ---------------------------------------------------------------------
                   13026: @cindex other system documentation, block words
                   13027: @cindex block words, other system documentation
                   13028: 
                   13029: @table @i
                   13030: @item any restrictions a multiprogramming system places on the use of buffer addresses:
                   13031: No restrictions (yet).
                   13032: 
                   13033: @item the number of blocks available for source and data:
                   13034: depends on your disk space.
                   13035: 
                   13036: @end table
                   13037: 
                   13038: 
                   13039: @c =====================================================================
                   13040: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
                   13041: @section The optional Double Number word set
                   13042: @c =====================================================================
                   13043: @cindex system documentation, double words
                   13044: @cindex double words, system documentation
                   13045: 
                   13046: @menu
                   13047: * double-ambcond::              Ambiguous Conditions              
                   13048: @end menu
                   13049: 
                   13050: 
                   13051: @c ---------------------------------------------------------------------
                   13052: @node double-ambcond,  , The optional Double Number word set, The optional Double Number word set
                   13053: @subsection Ambiguous conditions
                   13054: @c ---------------------------------------------------------------------
                   13055: @cindex double words, ambiguous conditions
                   13056: @cindex ambiguous conditions, double words
                   13057: 
                   13058: @table @i
1.29      crook    13059: @item @i{d} outside of range of @i{n} in @code{D>S}:
                   13060: @cindex @code{D>S}, @i{d} out of range of @i{n} 
                   13061: The least significant cell of @i{d} is produced.
1.1       anton    13062: 
                   13063: @end table
                   13064: 
                   13065: 
                   13066: @c =====================================================================
                   13067: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
                   13068: @section The optional Exception word set
                   13069: @c =====================================================================
                   13070: @cindex system documentation, exception words
                   13071: @cindex exception words, system documentation
                   13072: 
                   13073: @menu
                   13074: * exception-idef::              Implementation Defined Options              
                   13075: @end menu
                   13076: 
                   13077: 
                   13078: @c ---------------------------------------------------------------------
                   13079: @node exception-idef,  , The optional Exception word set, The optional Exception word set
                   13080: @subsection Implementation Defined Options
                   13081: @c ---------------------------------------------------------------------
                   13082: @cindex implementation-defined options, exception words
                   13083: @cindex exception words, implementation-defined options
                   13084: 
                   13085: @table @i
                   13086: @item @code{THROW}-codes used in the system:
                   13087: @cindex @code{THROW}-codes used in the system
                   13088: The codes -256@minus{}-511 are used for reporting signals. The mapping
1.29      crook    13089: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
1.1       anton    13090: codes -512@minus{}-2047 are used for OS errors (for file and memory
                   13091: allocation operations). The mapping from OS error numbers to throw codes
                   13092: is -512@minus{}@code{errno}. One side effect of this mapping is that
                   13093: undefined OS errors produce a message with a strange number; e.g.,
                   13094: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
                   13095: @end table
                   13096: 
                   13097: @c =====================================================================
                   13098: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
                   13099: @section The optional Facility word set
                   13100: @c =====================================================================
                   13101: @cindex system documentation, facility words
                   13102: @cindex facility words, system documentation
                   13103: 
                   13104: @menu
                   13105: * facility-idef::               Implementation Defined Options               
                   13106: * facility-ambcond::            Ambiguous Conditions            
                   13107: @end menu
                   13108: 
                   13109: 
                   13110: @c ---------------------------------------------------------------------
                   13111: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
                   13112: @subsection Implementation Defined Options
                   13113: @c ---------------------------------------------------------------------
                   13114: @cindex implementation-defined options, facility words
                   13115: @cindex facility words, implementation-defined options
                   13116: 
                   13117: @table @i
                   13118: @item encoding of keyboard events (@code{EKEY}):
                   13119: @cindex keyboard events, encoding in @code{EKEY}
                   13120: @cindex @code{EKEY}, encoding of keyboard events
1.40      anton    13121: Keys corresponding to ASCII characters are encoded as ASCII characters.
1.41      anton    13122: Other keys are encoded with the constants @code{k-left}, @code{k-right},
                   13123: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
                   13124: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
                   13125: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
1.40      anton    13126: 
1.1       anton    13127: 
                   13128: @item duration of a system clock tick:
                   13129: @cindex duration of a system clock tick
                   13130: @cindex clock tick duration
                   13131: System dependent. With respect to @code{MS}, the time is specified in
                   13132: microseconds. How well the OS and the hardware implement this, is
                   13133: another question.
                   13134: 
                   13135: @item repeatability to be expected from the execution of @code{MS}:
                   13136: @cindex repeatability to be expected from the execution of @code{MS}
                   13137: @cindex @code{MS}, repeatability to be expected
                   13138: System dependent. On Unix, a lot depends on load. If the system is
                   13139: lightly loaded, and the delay is short enough that Gforth does not get
                   13140: swapped out, the performance should be acceptable. Under MS-DOS and
                   13141: other single-tasking systems, it should be good.
                   13142: 
                   13143: @end table
                   13144: 
                   13145: 
                   13146: @c ---------------------------------------------------------------------
                   13147: @node facility-ambcond,  , facility-idef, The optional Facility word set
                   13148: @subsection Ambiguous conditions
                   13149: @c ---------------------------------------------------------------------
                   13150: @cindex facility words, ambiguous conditions
                   13151: @cindex ambiguous conditions, facility words
                   13152: 
                   13153: @table @i
                   13154: @item @code{AT-XY} can't be performed on user output device:
                   13155: @cindex @code{AT-XY} can't be performed on user output device
                   13156: Largely terminal dependent. No range checks are done on the arguments.
                   13157: No errors are reported. You may see some garbage appearing, you may see
                   13158: simply nothing happen.
                   13159: 
                   13160: @end table
                   13161: 
                   13162: 
                   13163: @c =====================================================================
                   13164: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
                   13165: @section The optional File-Access word set
                   13166: @c =====================================================================
                   13167: @cindex system documentation, file words
                   13168: @cindex file words, system documentation
                   13169: 
                   13170: @menu
                   13171: * file-idef::                   Implementation Defined Options
                   13172: * file-ambcond::                Ambiguous Conditions                
                   13173: @end menu
                   13174: 
                   13175: @c ---------------------------------------------------------------------
                   13176: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
                   13177: @subsection Implementation Defined Options
                   13178: @c ---------------------------------------------------------------------
                   13179: @cindex implementation-defined options, file words
                   13180: @cindex file words, implementation-defined options
                   13181: 
                   13182: @table @i
                   13183: @item file access methods used:
                   13184: @cindex file access methods used
                   13185: @code{R/O}, @code{R/W} and @code{BIN} work as you would
                   13186: expect. @code{W/O} translates into the C file opening mode @code{w} (or
                   13187: @code{wb}): The file is cleared, if it exists, and created, if it does
                   13188: not (with both @code{open-file} and @code{create-file}).  Under Unix
                   13189: @code{create-file} creates a file with 666 permissions modified by your
                   13190: umask.
                   13191: 
                   13192: @item file exceptions:
                   13193: @cindex file exceptions
                   13194: The file words do not raise exceptions (except, perhaps, memory access
                   13195: faults when you pass illegal addresses or file-ids).
                   13196: 
                   13197: @item file line terminator:
                   13198: @cindex file line terminator
                   13199: System-dependent. Gforth uses C's newline character as line
                   13200: terminator. What the actual character code(s) of this are is
                   13201: system-dependent.
                   13202: 
                   13203: @item file name format:
                   13204: @cindex file name format
                   13205: System dependent. Gforth just uses the file name format of your OS.
                   13206: 
                   13207: @item information returned by @code{FILE-STATUS}:
                   13208: @cindex @code{FILE-STATUS}, returned information
                   13209: @code{FILE-STATUS} returns the most powerful file access mode allowed
                   13210: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
                   13211: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
                   13212: along with the returned mode.
                   13213: 
                   13214: @item input file state after an exception when including source:
                   13215: @cindex exception when including source
                   13216: All files that are left via the exception are closed.
                   13217: 
1.29      crook    13218: @item @i{ior} values and meaning:
                   13219: @cindex @i{ior} values and meaning
1.68      anton    13220: @cindex @i{wior} values and meaning
1.29      crook    13221: The @i{ior}s returned by the file and memory allocation words are
1.1       anton    13222: intended as throw codes. They typically are in the range
                   13223: -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
1.29      crook    13224: @i{ior}s is -512@minus{}@i{errno}.
1.1       anton    13225: 
                   13226: @item maximum depth of file input nesting:
                   13227: @cindex maximum depth of file input nesting
                   13228: @cindex file input nesting, maximum depth
                   13229: limited by the amount of return stack, locals/TIB stack, and the number
                   13230: of open files available. This should not give you troubles.
                   13231: 
                   13232: @item maximum size of input line:
                   13233: @cindex maximum size of input line
                   13234: @cindex input line size, maximum
                   13235: @code{/line}. Currently 255.
                   13236: 
                   13237: @item methods of mapping block ranges to files:
                   13238: @cindex mapping block ranges to files
                   13239: @cindex files containing blocks
                   13240: @cindex blocks in files
                   13241: By default, blocks are accessed in the file @file{blocks.fb} in the
                   13242: current working directory. The file can be switched with @code{USE}.
                   13243: 
                   13244: @item number of string buffers provided by @code{S"}:
                   13245: @cindex @code{S"}, number of string buffers
                   13246: 1
                   13247: 
                   13248: @item size of string buffer used by @code{S"}:
                   13249: @cindex @code{S"}, size of string buffer
                   13250: @code{/line}. currently 255.
                   13251: 
                   13252: @end table
                   13253: 
                   13254: @c ---------------------------------------------------------------------
                   13255: @node file-ambcond,  , file-idef, The optional File-Access word set
                   13256: @subsection Ambiguous conditions
                   13257: @c ---------------------------------------------------------------------
                   13258: @cindex file words, ambiguous conditions
                   13259: @cindex ambiguous conditions, file words
                   13260: 
                   13261: @table @i
                   13262: @item attempting to position a file outside its boundaries:
                   13263: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
                   13264: @code{REPOSITION-FILE} is performed as usual: Afterwards,
                   13265: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
                   13266: 
                   13267: @item attempting to read from file positions not yet written:
                   13268: @cindex reading from file positions not yet written
                   13269: End-of-file, i.e., zero characters are read and no error is reported.
                   13270: 
1.29      crook    13271: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
                   13272: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid 
1.1       anton    13273: An appropriate exception may be thrown, but a memory fault or other
                   13274: problem is more probable.
                   13275: 
1.29      crook    13276: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
                   13277: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
                   13278: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
                   13279: The @i{ior} produced by the operation, that discovered the problem, is
1.1       anton    13280: thrown.
                   13281: 
                   13282: @item named file cannot be opened (@code{INCLUDED}):
                   13283: @cindex @code{INCLUDED}, named file cannot be opened
1.29      crook    13284: The @i{ior} produced by @code{open-file} is thrown.
1.1       anton    13285: 
                   13286: @item requesting an unmapped block number:
                   13287: @cindex unmapped block numbers
                   13288: There are no unmapped legal block numbers. On some operating systems,
                   13289: writing a block with a large number may overflow the file system and
                   13290: have an error message as consequence.
                   13291: 
                   13292: @item using @code{source-id} when @code{blk} is non-zero:
                   13293: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
                   13294: @code{source-id} performs its function. Typically it will give the id of
                   13295: the source which loaded the block. (Better ideas?)
                   13296: 
                   13297: @end table
                   13298: 
                   13299: 
                   13300: @c =====================================================================
                   13301: @node  The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
                   13302: @section The optional Floating-Point word set
                   13303: @c =====================================================================
                   13304: @cindex system documentation, floating-point words
                   13305: @cindex floating-point words, system documentation
                   13306: 
                   13307: @menu
                   13308: * floating-idef::               Implementation Defined Options
                   13309: * floating-ambcond::            Ambiguous Conditions            
                   13310: @end menu
                   13311: 
                   13312: 
                   13313: @c ---------------------------------------------------------------------
                   13314: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
                   13315: @subsection Implementation Defined Options
                   13316: @c ---------------------------------------------------------------------
                   13317: @cindex implementation-defined options, floating-point words
                   13318: @cindex floating-point words, implementation-defined options
                   13319: 
                   13320: @table @i
                   13321: @item format and range of floating point numbers:
                   13322: @cindex format and range of floating point numbers
                   13323: @cindex floating point numbers, format and range
                   13324: System-dependent; the @code{double} type of C.
                   13325: 
1.29      crook    13326: @item results of @code{REPRESENT} when @i{float} is out of range:
                   13327: @cindex  @code{REPRESENT}, results when @i{float} is out of range
1.1       anton    13328: System dependent; @code{REPRESENT} is implemented using the C library
                   13329: function @code{ecvt()} and inherits its behaviour in this respect.
                   13330: 
                   13331: @item rounding or truncation of floating-point numbers:
                   13332: @cindex rounding of floating-point numbers
                   13333: @cindex truncation of floating-point numbers
                   13334: @cindex floating-point numbers, rounding or truncation
                   13335: System dependent; the rounding behaviour is inherited from the hosting C
                   13336: compiler. IEEE-FP-based (i.e., most) systems by default round to
                   13337: nearest, and break ties by rounding to even (i.e., such that the last
                   13338: bit of the mantissa is 0).
                   13339: 
                   13340: @item size of floating-point stack:
                   13341: @cindex floating-point stack size
                   13342: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
                   13343: the floating-point stack (in floats). You can specify this on startup
                   13344: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
                   13345: 
                   13346: @item width of floating-point stack:
                   13347: @cindex floating-point stack width 
                   13348: @code{1 floats}.
                   13349: 
                   13350: @end table
                   13351: 
                   13352: 
                   13353: @c ---------------------------------------------------------------------
                   13354: @node floating-ambcond,  , floating-idef, The optional Floating-Point word set
                   13355: @subsection Ambiguous conditions
                   13356: @c ---------------------------------------------------------------------
                   13357: @cindex floating-point words, ambiguous conditions
                   13358: @cindex ambiguous conditions, floating-point words
                   13359: 
                   13360: @table @i
                   13361: @item @code{df@@} or @code{df!} used with an address that is not double-float  aligned:
                   13362: @cindex @code{df@@} or @code{df!} used with an address that is not double-float  aligned
                   13363: System-dependent. Typically results in a @code{-23 THROW} like other
                   13364: alignment violations.
                   13365: 
                   13366: @item @code{f@@} or @code{f!} used with an address that is not float  aligned:
                   13367: @cindex @code{f@@} used with an address that is not float aligned
                   13368: @cindex @code{f!} used with an address that is not float aligned
                   13369: System-dependent. Typically results in a @code{-23 THROW} like other
                   13370: alignment violations.
                   13371: 
                   13372: @item floating-point result out of range:
                   13373: @cindex floating-point result out of range
1.80      anton    13374: System-dependent. Can result in a @code{-43 throw} (floating point
                   13375: overflow), @code{-54 throw} (floating point underflow), @code{-41 throw}
                   13376: (floating point inexact result), @code{-55 THROW} (Floating-point
1.1       anton    13377: unidentified fault), or can produce a special value representing, e.g.,
                   13378: Infinity.
                   13379: 
                   13380: @item @code{sf@@} or @code{sf!} used with an address that is not single-float  aligned:
                   13381: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float  aligned
                   13382: System-dependent. Typically results in an alignment fault like other
                   13383: alignment violations.
                   13384: 
1.35      anton    13385: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
                   13386: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
1.1       anton    13387: The floating-point number is converted into decimal nonetheless.
                   13388: 
                   13389: @item Both arguments are equal to zero (@code{FATAN2}):
                   13390: @cindex @code{FATAN2}, both arguments are equal to zero
                   13391: System-dependent. @code{FATAN2} is implemented using the C library
                   13392: function @code{atan2()}.
                   13393: 
1.29      crook    13394: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
                   13395: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
                   13396: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
1.1       anton    13397: because of small errors and the tan will be a very large (or very small)
                   13398: but finite number.
                   13399: 
1.29      crook    13400: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
                   13401: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
1.1       anton    13402: The result is rounded to the nearest float.
                   13403: 
                   13404: @item dividing by zero:
                   13405: @cindex dividing by zero, floating-point
                   13406: @cindex floating-point dividing by zero
                   13407: @cindex floating-point unidentified fault, FP divide-by-zero
1.80      anton    13408: Platform-dependent; can produce an Infinity, NaN, @code{-42 throw}
                   13409: (floating point divide by zero) or @code{-55 throw} (Floating-point
                   13410: unidentified fault).
1.1       anton    13411: 
                   13412: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
                   13413: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
                   13414: System dependent. On IEEE-FP based systems the number is converted into
                   13415: an infinity.
                   13416: 
1.29      crook    13417: @item @i{float}<1 (@code{FACOSH}):
                   13418: @cindex @code{FACOSH}, @i{float}<1
1.1       anton    13419: @cindex floating-point unidentified fault, @code{FACOSH}
1.80      anton    13420: Platform-dependent; on IEEE-FP systems typically produces a NaN.
1.1       anton    13421: 
1.29      crook    13422: @item @i{float}=<-1 (@code{FLNP1}):
                   13423: @cindex @code{FLNP1}, @i{float}=<-1
1.1       anton    13424: @cindex floating-point unidentified fault, @code{FLNP1}
1.80      anton    13425: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
                   13426: negative infinity for @i{float}=-1).
1.1       anton    13427: 
1.29      crook    13428: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
                   13429: @cindex @code{FLN}, @i{float}=<0
                   13430: @cindex @code{FLOG}, @i{float}=<0
1.1       anton    13431: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
1.80      anton    13432: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
                   13433: negative infinity for @i{float}=0).
1.1       anton    13434: 
1.29      crook    13435: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
                   13436: @cindex @code{FASINH}, @i{float}<0
                   13437: @cindex @code{FSQRT}, @i{float}<0
1.1       anton    13438: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
1.80      anton    13439: Platform-dependent; for @code{fsqrt} this typically gives a NaN, for
                   13440: @code{fasinh} some platforms produce a NaN, others a number (bug in the
                   13441: C library?).
1.1       anton    13442: 
1.29      crook    13443: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
                   13444: @cindex @code{FACOS}, |@i{float}|>1
                   13445: @cindex @code{FASIN}, |@i{float}|>1
                   13446: @cindex @code{FATANH}, |@i{float}|>1
1.1       anton    13447: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
1.80      anton    13448: Platform-dependent; IEEE-FP systems typically produce a NaN.
1.1       anton    13449: 
1.29      crook    13450: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
                   13451: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
1.1       anton    13452: @cindex floating-point unidentified fault, @code{F>D}
1.80      anton    13453: Platform-dependent; typically, some double number is produced and no
                   13454: error is reported.
1.1       anton    13455: 
                   13456: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
                   13457: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
1.80      anton    13458: @code{Precision} characters of the numeric output area are used.  If
                   13459: @code{precision} is too high, these words will smash the data or code
                   13460: close to @code{here}.
1.1       anton    13461: @end table
                   13462: 
                   13463: @c =====================================================================
                   13464: @node  The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
                   13465: @section The optional Locals word set
                   13466: @c =====================================================================
                   13467: @cindex system documentation, locals words
                   13468: @cindex locals words, system documentation
                   13469: 
                   13470: @menu
                   13471: * locals-idef::                 Implementation Defined Options                 
                   13472: * locals-ambcond::              Ambiguous Conditions              
                   13473: @end menu
                   13474: 
                   13475: 
                   13476: @c ---------------------------------------------------------------------
                   13477: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
                   13478: @subsection Implementation Defined Options
                   13479: @c ---------------------------------------------------------------------
                   13480: @cindex implementation-defined options, locals words
                   13481: @cindex locals words, implementation-defined options
                   13482: 
                   13483: @table @i
                   13484: @item maximum number of locals in a definition:
                   13485: @cindex maximum number of locals in a definition
                   13486: @cindex locals, maximum number in a definition
                   13487: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
                   13488: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
                   13489: characters. The number of locals in a definition is bounded by the size
                   13490: of locals-buffer, which contains the names of the locals.
                   13491: 
                   13492: @end table
                   13493: 
                   13494: 
                   13495: @c ---------------------------------------------------------------------
                   13496: @node locals-ambcond,  , locals-idef, The optional Locals word set
                   13497: @subsection Ambiguous conditions
                   13498: @c ---------------------------------------------------------------------
                   13499: @cindex locals words, ambiguous conditions
                   13500: @cindex ambiguous conditions, locals words
                   13501: 
                   13502: @table @i
                   13503: @item executing a named local in interpretation state:
                   13504: @cindex local in interpretation state
                   13505: @cindex Interpreting a compile-only word, for a local
                   13506: Locals have no interpretation semantics. If you try to perform the
                   13507: interpretation semantics, you will get a @code{-14 throw} somewhere
                   13508: (Interpreting a compile-only word). If you perform the compilation
                   13509: semantics, the locals access will be compiled (irrespective of state).
                   13510: 
1.29      crook    13511: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
1.1       anton    13512: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
                   13513: @cindex @code{TO} on non-@code{VALUE}s and non-locals
                   13514: @cindex Invalid name argument, @code{TO}
                   13515: @code{-32 throw} (Invalid name argument)
                   13516: 
                   13517: @end table
                   13518: 
                   13519: 
                   13520: @c =====================================================================
                   13521: @node  The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
                   13522: @section The optional Memory-Allocation word set
                   13523: @c =====================================================================
                   13524: @cindex system documentation, memory-allocation words
                   13525: @cindex memory-allocation words, system documentation
                   13526: 
                   13527: @menu
                   13528: * memory-idef::                 Implementation Defined Options                 
                   13529: @end menu
                   13530: 
                   13531: 
                   13532: @c ---------------------------------------------------------------------
                   13533: @node memory-idef,  , The optional Memory-Allocation word set, The optional Memory-Allocation word set
                   13534: @subsection Implementation Defined Options
                   13535: @c ---------------------------------------------------------------------
                   13536: @cindex implementation-defined options, memory-allocation words
                   13537: @cindex memory-allocation words, implementation-defined options
                   13538: 
                   13539: @table @i
1.29      crook    13540: @item values and meaning of @i{ior}:
                   13541: @cindex  @i{ior} values and meaning
                   13542: The @i{ior}s returned by the file and memory allocation words are
1.1       anton    13543: intended as throw codes. They typically are in the range
                   13544: -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
1.29      crook    13545: @i{ior}s is -512@minus{}@i{errno}.
1.1       anton    13546: 
                   13547: @end table
                   13548: 
                   13549: @c =====================================================================
                   13550: @node  The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
                   13551: @section The optional Programming-Tools word set
                   13552: @c =====================================================================
                   13553: @cindex system documentation, programming-tools words
                   13554: @cindex programming-tools words, system documentation
                   13555: 
                   13556: @menu
                   13557: * programming-idef::            Implementation Defined Options            
                   13558: * programming-ambcond::         Ambiguous Conditions         
                   13559: @end menu
                   13560: 
                   13561: 
                   13562: @c ---------------------------------------------------------------------
                   13563: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
                   13564: @subsection Implementation Defined Options
                   13565: @c ---------------------------------------------------------------------
                   13566: @cindex implementation-defined options, programming-tools words
                   13567: @cindex programming-tools words, implementation-defined options
                   13568: 
                   13569: @table @i
                   13570: @item ending sequence for input following @code{;CODE} and @code{CODE}:
                   13571: @cindex @code{;CODE} ending sequence
                   13572: @cindex @code{CODE} ending sequence
                   13573: @code{END-CODE}
                   13574: 
                   13575: @item manner of processing input following @code{;CODE} and @code{CODE}:
                   13576: @cindex @code{;CODE}, processing input
                   13577: @cindex @code{CODE}, processing input
                   13578: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
                   13579: the input is processed by the text interpreter, (starting) in interpret
                   13580: state.
                   13581: 
                   13582: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
                   13583: @cindex @code{ASSEMBLER}, search order capability
                   13584: The ANS Forth search order word set.
                   13585: 
                   13586: @item source and format of display by @code{SEE}:
                   13587: @cindex @code{SEE}, source and format of output
1.80      anton    13588: The source for @code{see} is the executable code used by the inner
1.1       anton    13589: interpreter.  The current @code{see} tries to output Forth source code
1.80      anton    13590: (and on some platforms, assembly code for primitives) as well as
                   13591: possible.
1.1       anton    13592: 
                   13593: @end table
                   13594: 
                   13595: @c ---------------------------------------------------------------------
                   13596: @node programming-ambcond,  , programming-idef, The optional Programming-Tools word set
                   13597: @subsection Ambiguous conditions
                   13598: @c ---------------------------------------------------------------------
                   13599: @cindex programming-tools words, ambiguous conditions
                   13600: @cindex ambiguous conditions, programming-tools words
                   13601: 
                   13602: @table @i
                   13603: 
1.21      crook    13604: @item deleting the compilation word list (@code{FORGET}):
                   13605: @cindex @code{FORGET}, deleting the compilation word list
1.1       anton    13606: Not implemented (yet).
                   13607: 
1.29      crook    13608: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
                   13609: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
                   13610: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
1.1       anton    13611: @cindex control-flow stack underflow
                   13612: This typically results in an @code{abort"} with a descriptive error
                   13613: message (may change into a @code{-22 throw} (Control structure mismatch)
                   13614: in the future). You may also get a memory access error. If you are
                   13615: unlucky, this ambiguous condition is not caught.
                   13616: 
1.29      crook    13617: @item @i{name} can't be found (@code{FORGET}):
                   13618: @cindex @code{FORGET}, @i{name} can't be found
1.1       anton    13619: Not implemented (yet).
                   13620: 
1.29      crook    13621: @item @i{name} not defined via @code{CREATE}:
                   13622: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
1.1       anton    13623: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
                   13624: the execution semantics of the last defined word no matter how it was
                   13625: defined.
                   13626: 
                   13627: @item @code{POSTPONE} applied to @code{[IF]}:
                   13628: @cindex @code{POSTPONE} applied to @code{[IF]}
                   13629: @cindex @code{[IF]} and @code{POSTPONE}
                   13630: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
                   13631: equivalent to @code{[IF]}.
                   13632: 
                   13633: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
                   13634: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
                   13635: Continue in the same state of conditional compilation in the next outer
                   13636: input source. Currently there is no warning to the user about this.
                   13637: 
                   13638: @item removing a needed definition (@code{FORGET}):
                   13639: @cindex @code{FORGET}, removing a needed definition
                   13640: Not implemented (yet).
                   13641: 
                   13642: @end table
                   13643: 
                   13644: 
                   13645: @c =====================================================================
                   13646: @node  The optional Search-Order word set,  , The optional Programming-Tools word set, ANS conformance
                   13647: @section The optional Search-Order word set
                   13648: @c =====================================================================
                   13649: @cindex system documentation, search-order words
                   13650: @cindex search-order words, system documentation
                   13651: 
                   13652: @menu
                   13653: * search-idef::                 Implementation Defined Options                 
                   13654: * search-ambcond::              Ambiguous Conditions              
                   13655: @end menu
                   13656: 
                   13657: 
                   13658: @c ---------------------------------------------------------------------
                   13659: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
                   13660: @subsection Implementation Defined Options
                   13661: @c ---------------------------------------------------------------------
                   13662: @cindex implementation-defined options, search-order words
                   13663: @cindex search-order words, implementation-defined options
                   13664: 
                   13665: @table @i
                   13666: @item maximum number of word lists in search order:
                   13667: @cindex maximum number of word lists in search order
                   13668: @cindex search order, maximum depth
                   13669: @code{s" wordlists" environment? drop .}. Currently 16.
                   13670: 
                   13671: @item minimum search order:
                   13672: @cindex minimum search order
                   13673: @cindex search order, minimum
                   13674: @code{root root}.
                   13675: 
                   13676: @end table
                   13677: 
                   13678: @c ---------------------------------------------------------------------
                   13679: @node search-ambcond,  , search-idef, The optional Search-Order word set
                   13680: @subsection Ambiguous conditions
                   13681: @c ---------------------------------------------------------------------
                   13682: @cindex search-order words, ambiguous conditions
                   13683: @cindex ambiguous conditions, search-order words
                   13684: 
                   13685: @table @i
1.21      crook    13686: @item changing the compilation word list (during compilation):
                   13687: @cindex changing the compilation word list (during compilation)
                   13688: @cindex compilation word list, change before definition ends
                   13689: The word is entered into the word list that was the compilation word list
1.1       anton    13690: at the start of the definition. Any changes to the name field (e.g.,
                   13691: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
1.116     anton    13692: are applied to the latest defined word (as reported by @code{latest} or
                   13693: @code{latestxt}), if possible, irrespective of the compilation word list.
1.1       anton    13694: 
                   13695: @item search order empty (@code{previous}):
                   13696: @cindex @code{previous}, search order empty
1.26      crook    13697: @cindex vocstack empty, @code{previous}
1.1       anton    13698: @code{abort" Vocstack empty"}.
                   13699: 
                   13700: @item too many word lists in search order (@code{also}):
                   13701: @cindex @code{also}, too many word lists in search order
1.26      crook    13702: @cindex vocstack full, @code{also}
1.1       anton    13703: @code{abort" Vocstack full"}.
                   13704: 
                   13705: @end table
                   13706: 
                   13707: @c ***************************************************************
1.65      anton    13708: @node Standard vs Extensions, Model, ANS conformance, Top
                   13709: @chapter Should I use Gforth extensions?
                   13710: @cindex Gforth extensions
                   13711: 
                   13712: As you read through the rest of this manual, you will see documentation
                   13713: for @i{Standard} words, and documentation for some appealing Gforth
                   13714: @i{extensions}. You might ask yourself the question: @i{``Should I
                   13715: restrict myself to the standard, or should I use the extensions?''}
                   13716: 
                   13717: The answer depends on the goals you have for the program you are working
                   13718: on:
                   13719: 
                   13720: @itemize @bullet
                   13721: 
                   13722: @item Is it just for yourself or do you want to share it with others?
                   13723: 
                   13724: @item
                   13725: If you want to share it, do the others all use Gforth?
                   13726: 
                   13727: @item
                   13728: If it is just for yourself, do you want to restrict yourself to Gforth?
                   13729: 
                   13730: @end itemize
                   13731: 
                   13732: If restricting the program to Gforth is ok, then there is no reason not
                   13733: to use extensions.  It is still a good idea to keep to the standard
                   13734: where it is easy, in case you want to reuse these parts in another
                   13735: program that you want to be portable.
                   13736: 
                   13737: If you want to be able to port the program to other Forth systems, there
                   13738: are the following points to consider:
                   13739: 
                   13740: @itemize @bullet
                   13741: 
                   13742: @item
                   13743: Most Forth systems that are being maintained support the ANS Forth
                   13744: standard.  So if your program complies with the standard, it will be
                   13745: portable among many systems.
                   13746: 
                   13747: @item
                   13748: A number of the Gforth extensions can be implemented in ANS Forth using
                   13749: public-domain files provided in the @file{compat/} directory. These are
                   13750: mentioned in the text in passing.  There is no reason not to use these
                   13751: extensions, your program will still be ANS Forth compliant; just include
                   13752: the appropriate compat files with your program.
                   13753: 
                   13754: @item
                   13755: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
                   13756: analyse your program and determine what non-Standard words it relies
                   13757: upon.  However, it does not check whether you use standard words in a
                   13758: non-standard way.
                   13759: 
                   13760: @item
                   13761: Some techniques are not standardized by ANS Forth, and are hard or
                   13762: impossible to implement in a standard way, but can be implemented in
                   13763: most Forth systems easily, and usually in similar ways (e.g., accessing
                   13764: word headers).  Forth has a rich historical precedent for programmers
                   13765: taking advantage of implementation-dependent features of their tools
                   13766: (for example, relying on a knowledge of the dictionary
                   13767: structure). Sometimes these techniques are necessary to extract every
                   13768: last bit of performance from the hardware, sometimes they are just a
                   13769: programming shorthand.
                   13770: 
                   13771: @item
                   13772: Does using a Gforth extension save more work than the porting this part
                   13773: to other Forth systems (if any) will cost?
                   13774: 
                   13775: @item
                   13776: Is the additional functionality worth the reduction in portability and
                   13777: the additional porting problems?
                   13778: 
                   13779: @end itemize
                   13780: 
                   13781: In order to perform these consideratios, you need to know what's
                   13782: standard and what's not.  This manual generally states if something is
1.81      anton    13783: non-standard, but the authoritative source is the
                   13784: @uref{http://www.taygeta.com/forth/dpans.html,standard document}.
1.65      anton    13785: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
                   13786: into the thought processes of the technical committee.
                   13787: 
                   13788: Note also that portability between Forth systems is not the only
                   13789: portability issue; there is also the issue of portability between
                   13790: different platforms (processor/OS combinations).
                   13791: 
                   13792: @c ***************************************************************
                   13793: @node Model, Integrating Gforth, Standard vs Extensions, Top
1.1       anton    13794: @chapter Model
                   13795: 
                   13796: This chapter has yet to be written. It will contain information, on
                   13797: which internal structures you can rely.
                   13798: 
                   13799: @c ***************************************************************
                   13800: @node Integrating Gforth, Emacs and Gforth, Model, Top
                   13801: @chapter Integrating Gforth into C programs
                   13802: 
                   13803: This is not yet implemented.
                   13804: 
                   13805: Several people like to use Forth as scripting language for applications
                   13806: that are otherwise written in C, C++, or some other language.
                   13807: 
                   13808: The Forth system ATLAST provides facilities for embedding it into
                   13809: applications; unfortunately it has several disadvantages: most
                   13810: importantly, it is not based on ANS Forth, and it is apparently dead
                   13811: (i.e., not developed further and not supported). The facilities
1.21      crook    13812: provided by Gforth in this area are inspired by ATLAST's facilities, so
1.1       anton    13813: making the switch should not be hard.
                   13814: 
                   13815: We also tried to design the interface such that it can easily be
                   13816: implemented by other Forth systems, so that we may one day arrive at a
                   13817: standardized interface. Such a standard interface would allow you to
                   13818: replace the Forth system without having to rewrite C code.
                   13819: 
                   13820: You embed the Gforth interpreter by linking with the library
                   13821: @code{libgforth.a} (give the compiler the option @code{-lgforth}).  All
                   13822: global symbols in this library that belong to the interface, have the
                   13823: prefix @code{forth_}. (Global symbols that are used internally have the
                   13824: prefix @code{gforth_}).
                   13825: 
                   13826: You can include the declarations of Forth types and the functions and
                   13827: variables of the interface with @code{#include <forth.h>}.
                   13828: 
                   13829: Types.
                   13830: 
                   13831: Variables.
                   13832: 
                   13833: Data and FP Stack pointer. Area sizes.
                   13834: 
                   13835: functions.
                   13836: 
                   13837: forth_init(imagefile)
                   13838: forth_evaluate(string) exceptions?
                   13839: forth_goto(address) (or forth_execute(xt)?)
                   13840: forth_continue() (a corountining mechanism)
                   13841: 
                   13842: Adding primitives.
                   13843: 
                   13844: No checking.
                   13845: 
                   13846: Signals?
                   13847: 
                   13848: Accessing the Stacks
                   13849: 
1.26      crook    13850: @c ******************************************************************
1.1       anton    13851: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
                   13852: @chapter Emacs and Gforth
                   13853: @cindex Emacs and Gforth
                   13854: 
                   13855: @cindex @file{gforth.el}
                   13856: @cindex @file{forth.el}
                   13857: @cindex Rydqvist, Goran
1.107     dvdkhlng 13858: @cindex Kuehling, David
1.1       anton    13859: @cindex comment editing commands
                   13860: @cindex @code{\}, editing with Emacs
                   13861: @cindex debug tracer editing commands
                   13862: @cindex @code{~~}, removal with Emacs
                   13863: @cindex Forth mode in Emacs
1.107     dvdkhlng 13864: 
1.1       anton    13865: Gforth comes with @file{gforth.el}, an improved version of
                   13866: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
1.26      crook    13867: improvements are:
                   13868: 
                   13869: @itemize @bullet
                   13870: @item
1.107     dvdkhlng 13871: A better handling of indentation.
                   13872: @item
                   13873: A custom hilighting engine for Forth-code.
1.26      crook    13874: @item
                   13875: Comment paragraph filling (@kbd{M-q})
                   13876: @item
                   13877: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
                   13878: @item
                   13879: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
1.41      anton    13880: @item
                   13881: Support of the @code{info-lookup} feature for looking up the
                   13882: documentation of a word.
1.107     dvdkhlng 13883: @item
                   13884: Support for reading and writing blocks files.
1.26      crook    13885: @end itemize
                   13886: 
1.107     dvdkhlng 13887: To get a basic description of these features, enter Forth mode and
                   13888: type @kbd{C-h m}.
1.1       anton    13889: 
                   13890: @cindex source location of error or debugging output in Emacs
                   13891: @cindex error output, finding the source location in Emacs
                   13892: @cindex debugging output, finding the source location in Emacs
                   13893: In addition, Gforth supports Emacs quite well: The source code locations
                   13894: given in error messages, debugging output (from @code{~~}) and failed
                   13895: assertion messages are in the right format for Emacs' compilation mode
                   13896: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
                   13897: Manual}) so the source location corresponding to an error or other
                   13898: message is only a few keystrokes away (@kbd{C-x `} for the next error,
                   13899: @kbd{C-c C-c} for the error under the cursor).
                   13900: 
1.107     dvdkhlng 13901: @cindex viewing the documentation of a word in Emacs
                   13902: @cindex context-sensitive help
                   13903: Moreover, for words documented in this manual, you can look up the
                   13904: glossary entry quickly by using @kbd{C-h TAB}
                   13905: (@code{info-lookup-symbol}, @pxref{Documentation, ,Documentation
                   13906: Commands, emacs, Emacs Manual}).  This feature requires Emacs 20.3 or
                   13907: later and does not work for words containing @code{:}.
                   13908: 
                   13909: @menu
                   13910: * Installing gforth.el::        Making Emacs aware of Forth.
                   13911: * Emacs Tags::                  Viewing the source of a word in Emacs.
                   13912: * Hilighting::                  Making Forth code look prettier.
                   13913: * Auto-Indentation::            Customizing auto-indentation.
                   13914: * Blocks Files::                Reading and writing blocks files.
                   13915: @end menu
                   13916: 
                   13917: @c ----------------------------------
1.109     anton    13918: @node Installing gforth.el, Emacs Tags, Emacs and Gforth, Emacs and Gforth
1.107     dvdkhlng 13919: @section Installing gforth.el
                   13920: @cindex @file{.emacs}
                   13921: @cindex @file{gforth.el}, installation
                   13922: To make the features from @file{gforth.el} available in Emacs, add
                   13923: the following lines to your @file{.emacs} file:
                   13924: 
                   13925: @example
                   13926: (autoload 'forth-mode "gforth.el")
                   13927: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) 
                   13928:                            auto-mode-alist))
                   13929: (autoload 'forth-block-mode "gforth.el")
                   13930: (setq auto-mode-alist (cons '("\\.fb\\'" . forth-block-mode) 
                   13931:                            auto-mode-alist))
                   13932: (add-hook 'forth-mode-hook (function (lambda ()
                   13933:    ;; customize variables here:
                   13934:    (setq forth-indent-level 4)
                   13935:    (setq forth-minor-indent-level 2)
                   13936:    (setq forth-hilight-level 3)
                   13937:    ;;; ...
                   13938: )))
                   13939: @end example
                   13940: 
                   13941: @c ----------------------------------
                   13942: @node Emacs Tags, Hilighting, Installing gforth.el, Emacs and Gforth
                   13943: @section Emacs Tags
1.1       anton    13944: @cindex @file{TAGS} file
                   13945: @cindex @file{etags.fs}
                   13946: @cindex viewing the source of a word in Emacs
1.43      anton    13947: @cindex @code{require}, placement in files
                   13948: @cindex @code{include}, placement in files
1.107     dvdkhlng 13949: If you @code{require} @file{etags.fs}, a new @file{TAGS} file will be
                   13950: produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
1.1       anton    13951: contains the definitions of all words defined afterwards. You can then
1.107     dvdkhlng 13952: find the source for a word using @kbd{M-.}. Note that Emacs can use
1.1       anton    13953: several tags files at the same time (e.g., one for the Gforth sources
                   13954: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
                   13955: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
                   13956: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
1.43      anton    13957: @file{/usr/local/share/gforth/0.2.0/TAGS}).  To get the best behaviour
                   13958: with @file{etags.fs}, you should avoid putting definitions both before
                   13959: and after @code{require} etc., otherwise you will see the same file
                   13960: visited several times by commands like @code{tags-search}.
1.1       anton    13961: 
1.107     dvdkhlng 13962: @c ----------------------------------
                   13963: @node Hilighting, Auto-Indentation, Emacs Tags, Emacs and Gforth
                   13964: @section Hilighting
                   13965: @cindex hilighting Forth code in Emacs
                   13966: @cindex highlighting Forth code in Emacs
                   13967: @file{gforth.el} comes with a custom source hilighting engine.  When
                   13968: you open a file in @code{forth-mode}, it will be completely parsed,
                   13969: assigning faces to keywords, comments, strings etc.  While you edit
                   13970: the file, modified regions get parsed and updated on-the-fly. 
                   13971: 
                   13972: Use the variable `forth-hilight-level' to change the level of
                   13973: decoration from 0 (no hilighting at all) to 3 (the default).  Even if
                   13974: you set the hilighting level to 0, the parser will still work in the
                   13975: background, collecting information about whether regions of text are
                   13976: ``compiled'' or ``interpreted''.  Those information are required for
                   13977: auto-indentation to work properly.  Set `forth-disable-parser' to
                   13978: non-nil if your computer is too slow to handle parsing.  This will
                   13979: have an impact on the smartness of the auto-indentation engine,
                   13980: though.
                   13981: 
                   13982: Sometimes Forth sources define new features that should be hilighted,
                   13983: new control structures, defining-words etc.  You can use the variable
                   13984: `forth-custom-words' to make @code{forth-mode} hilight additional
                   13985: words and constructs.  See the docstring of `forth-words' for details
                   13986: (in Emacs, type @kbd{C-h v forth-words}).
                   13987: 
                   13988: `forth-custom-words' is meant to be customized in your
                   13989: @file{.emacs} file.  To customize hilighing in a file-specific manner,
                   13990: set `forth-local-words' in a local-variables section at the end of
                   13991: your source file (@pxref{Local Variables in Files,, Variables, emacs, Emacs Manual}).
                   13992: 
                   13993: Example:
                   13994: @example
                   13995: 0 [IF]
                   13996:    Local Variables:
                   13997:    forth-local-words:
                   13998:       ((("t:") definition-starter (font-lock-keyword-face . 1)
                   13999:         "[ \t\n]" t name (font-lock-function-name-face . 3))
                   14000:        ((";t") definition-ender (font-lock-keyword-face . 1)))
                   14001:    End:
                   14002: [THEN]
                   14003: @end example
                   14004: 
                   14005: @c ----------------------------------
                   14006: @node Auto-Indentation, Blocks Files, Hilighting, Emacs and Gforth
                   14007: @section Auto-Indentation
                   14008: @cindex auto-indentation of Forth code in Emacs
                   14009: @cindex indentation of Forth code in Emacs
                   14010: @code{forth-mode} automatically tries to indent lines in a smart way,
                   14011: whenever you type @key{TAB} or break a line with @kbd{C-m}.
                   14012: 
                   14013: Simple customization can be achieved by setting
                   14014: `forth-indent-level' and `forth-minor-indent-level' in your
                   14015: @file{.emacs} file. For historical reasons @file{gforth.el} indents
                   14016: per default by multiples of 4 columns.  To use the more traditional
                   14017: 3-column indentation, add the following lines to your @file{.emacs}:
                   14018: 
                   14019: @example
                   14020: (add-hook 'forth-mode-hook (function (lambda ()
                   14021:    ;; customize variables here:
                   14022:    (setq forth-indent-level 3)
                   14023:    (setq forth-minor-indent-level 1)
                   14024: )))
                   14025: @end example
                   14026: 
                   14027: If you want indentation to recognize non-default words, customize it
                   14028: by setting `forth-custom-indent-words' in your @file{.emacs}.  See the
                   14029: docstring of `forth-indent-words' for details (in Emacs, type @kbd{C-h
                   14030: v forth-indent-words}).
                   14031: 
                   14032: To customize indentation in a file-specific manner, set
                   14033: `forth-local-indent-words' in a local-variables section at the end of
                   14034: your source file (@pxref{Local Variables in Files, Variables,,emacs,
                   14035: Emacs Manual}).
                   14036: 
                   14037: Example:
                   14038: @example
                   14039: 0 [IF]
                   14040:    Local Variables:
                   14041:    forth-local-indent-words:
                   14042:       ((("t:") (0 . 2) (0 . 2))
                   14043:        ((";t") (-2 . 0) (0 . -2)))
                   14044:    End:
                   14045: [THEN]
                   14046: @end example
                   14047: 
                   14048: @c ----------------------------------
1.109     anton    14049: @node Blocks Files,  , Auto-Indentation, Emacs and Gforth
1.107     dvdkhlng 14050: @section Blocks Files
                   14051: @cindex blocks files, use with Emacs
                   14052: @code{forth-mode} Autodetects blocks files by checking whether the
                   14053: length of the first line exceeds 1023 characters.  It then tries to
                   14054: convert the file into normal text format.  When you save the file, it
                   14055: will be written to disk as normal stream-source file.
                   14056: 
                   14057: If you want to write blocks files, use @code{forth-blocks-mode}.  It
                   14058: inherits all the features from @code{forth-mode}, plus some additions:
1.41      anton    14059: 
1.107     dvdkhlng 14060: @itemize @bullet
                   14061: @item
                   14062: Files are written to disk in blocks file format.
                   14063: @item
                   14064: Screen numbers are displayed in the mode line (enumerated beginning
                   14065: with the value of `forth-block-base')
                   14066: @item
                   14067: Warnings are displayed when lines exceed 64 characters.
                   14068: @item
                   14069: The beginning of the currently edited block is marked with an
                   14070: overlay-arrow. 
                   14071: @end itemize
1.41      anton    14072: 
1.107     dvdkhlng 14073: There are some restrictions you should be aware of.  When you open a
                   14074: blocks file that contains tabulator or newline characters, these
                   14075: characters will be translated into spaces when the file is written
                   14076: back to disk.  If tabs or newlines are encountered during blocks file
                   14077: reading, an error is output to the echo area. So have a look at the
                   14078: `*Messages*' buffer, when Emacs' bell rings during reading.
1.1       anton    14079: 
1.107     dvdkhlng 14080: Please consult the docstring of @code{forth-blocks-mode} for more
                   14081: information by typing @kbd{C-h v forth-blocks-mode}).
1.1       anton    14082: 
1.26      crook    14083: @c ******************************************************************
1.1       anton    14084: @node Image Files, Engine, Emacs and Gforth, Top
                   14085: @chapter Image Files
1.26      crook    14086: @cindex image file
                   14087: @cindex @file{.fi} files
1.1       anton    14088: @cindex precompiled Forth code
                   14089: @cindex dictionary in persistent form
                   14090: @cindex persistent form of dictionary
                   14091: 
                   14092: An image file is a file containing an image of the Forth dictionary,
                   14093: i.e., compiled Forth code and data residing in the dictionary.  By
                   14094: convention, we use the extension @code{.fi} for image files.
                   14095: 
                   14096: @menu
1.18      anton    14097: * Image Licensing Issues::      Distribution terms for images.
                   14098: * Image File Background::       Why have image files?
1.67      anton    14099: * Non-Relocatable Image Files::  don't always work.
1.18      anton    14100: * Data-Relocatable Image Files::  are better.
1.67      anton    14101: * Fully Relocatable Image Files::  better yet.
1.18      anton    14102: * Stack and Dictionary Sizes::  Setting the default sizes for an image.
1.29      crook    14103: * Running Image Files::         @code{gforth -i @i{file}} or @i{file}.
1.18      anton    14104: * Modifying the Startup Sequence::  and turnkey applications.
1.1       anton    14105: @end menu
                   14106: 
1.18      anton    14107: @node Image Licensing Issues, Image File Background, Image Files, Image Files
                   14108: @section Image Licensing Issues
                   14109: @cindex license for images
                   14110: @cindex image license
                   14111: 
                   14112: An image created with @code{gforthmi} (@pxref{gforthmi}) or
                   14113: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
                   14114: original image; i.e., according to copyright law it is a derived work of
                   14115: the original image.
                   14116: 
                   14117: Since Gforth is distributed under the GNU GPL, the newly created image
                   14118: falls under the GNU GPL, too. In particular, this means that if you
                   14119: distribute the image, you have to make all of the sources for the image
1.113     anton    14120: available, including those you wrote.  For details see @ref{Copying, ,
1.18      anton    14121: GNU General Public License (Section 3)}.
                   14122: 
                   14123: If you create an image with @code{cross} (@pxref{cross.fs}), the image
                   14124: contains only code compiled from the sources you gave it; if none of
                   14125: these sources is under the GPL, the terms discussed above do not apply
                   14126: to the image. However, if your image needs an engine (a gforth binary)
                   14127: that is under the GPL, you should make sure that you distribute both in
                   14128: a way that is at most a @emph{mere aggregation}, if you don't want the
                   14129: terms of the GPL to apply to the image.
                   14130: 
                   14131: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
1.1       anton    14132: @section Image File Background
                   14133: @cindex image file background
                   14134: 
1.80      anton    14135: Gforth consists not only of primitives (in the engine), but also of
1.1       anton    14136: definitions written in Forth. Since the Forth compiler itself belongs to
                   14137: those definitions, it is not possible to start the system with the
1.80      anton    14138: engine and the Forth source alone. Therefore we provide the Forth
1.26      crook    14139: code as an image file in nearly executable form. When Gforth starts up,
                   14140: a C routine loads the image file into memory, optionally relocates the
                   14141: addresses, then sets up the memory (stacks etc.) according to
                   14142: information in the image file, and (finally) starts executing Forth
                   14143: code.
1.1       anton    14144: 
                   14145: The image file variants represent different compromises between the
                   14146: goals of making it easy to generate image files and making them
                   14147: portable.
                   14148: 
                   14149: @cindex relocation at run-time
1.26      crook    14150: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
1.1       anton    14151: run-time. This avoids many of the complications discussed below (image
                   14152: files are data relocatable without further ado), but costs performance
                   14153: (one addition per memory access).
                   14154: 
                   14155: @cindex relocation at load-time
1.26      crook    14156: By contrast, the Gforth loader performs relocation at image load time. The
                   14157: loader also has to replace tokens that represent primitive calls with the
1.1       anton    14158: appropriate code-field addresses (or code addresses in the case of
                   14159: direct threading).
                   14160: 
                   14161: There are three kinds of image files, with different degrees of
                   14162: relocatability: non-relocatable, data-relocatable, and fully relocatable
                   14163: image files.
                   14164: 
                   14165: @cindex image file loader
                   14166: @cindex relocating loader
                   14167: @cindex loader for image files
                   14168: These image file variants have several restrictions in common; they are
                   14169: caused by the design of the image file loader:
                   14170: 
                   14171: @itemize @bullet
                   14172: @item
                   14173: There is only one segment; in particular, this means, that an image file
                   14174: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
1.26      crook    14175: them). The contents of the stacks are not represented, either.
1.1       anton    14176: 
                   14177: @item
                   14178: The only kinds of relocation supported are: adding the same offset to
                   14179: all cells that represent data addresses; and replacing special tokens
                   14180: with code addresses or with pieces of machine code.
                   14181: 
                   14182: If any complex computations involving addresses are performed, the
                   14183: results cannot be represented in the image file. Several applications that
                   14184: use such computations come to mind:
                   14185: @itemize @minus
                   14186: @item
                   14187: Hashing addresses (or data structures which contain addresses) for table
                   14188: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
                   14189: purpose, you will have no problem, because the hash tables are
                   14190: recomputed automatically when the system is started. If you use your own
                   14191: hash tables, you will have to do something similar.
                   14192: 
                   14193: @item
                   14194: There's a cute implementation of doubly-linked lists that uses
                   14195: @code{XOR}ed addresses. You could represent such lists as singly-linked
                   14196: in the image file, and restore the doubly-linked representation on
                   14197: startup.@footnote{In my opinion, though, you should think thrice before
                   14198: using a doubly-linked list (whatever implementation).}
                   14199: 
                   14200: @item
                   14201: The code addresses of run-time routines like @code{docol:} cannot be
                   14202: represented in the image file (because their tokens would be replaced by
                   14203: machine code in direct threaded implementations). As a workaround,
                   14204: compute these addresses at run-time with @code{>code-address} from the
                   14205: executions tokens of appropriate words (see the definitions of
1.80      anton    14206: @code{docol:} and friends in @file{kernel/getdoers.fs}).
1.1       anton    14207: 
                   14208: @item
                   14209: On many architectures addresses are represented in machine code in some
                   14210: shifted or mangled form. You cannot put @code{CODE} words that contain
                   14211: absolute addresses in this form in a relocatable image file. Workarounds
                   14212: are representing the address in some relative form (e.g., relative to
                   14213: the CFA, which is present in some register), or loading the address from
                   14214: a place where it is stored in a non-mangled form.
                   14215: @end itemize
                   14216: @end itemize
                   14217: 
                   14218: @node  Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
                   14219: @section Non-Relocatable Image Files
                   14220: @cindex non-relocatable image files
1.26      crook    14221: @cindex image file, non-relocatable
1.1       anton    14222: 
                   14223: These files are simple memory dumps of the dictionary. They are specific
                   14224: to the executable (i.e., @file{gforth} file) they were created
                   14225: with. What's worse, they are specific to the place on which the
                   14226: dictionary resided when the image was created. Now, there is no
                   14227: guarantee that the dictionary will reside at the same place the next
                   14228: time you start Gforth, so there's no guarantee that a non-relocatable
                   14229: image will work the next time (Gforth will complain instead of crashing,
                   14230: though).
                   14231: 
                   14232: You can create a non-relocatable image file with
                   14233: 
1.44      crook    14234: 
1.1       anton    14235: doc-savesystem
                   14236: 
1.44      crook    14237: 
1.1       anton    14238: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
                   14239: @section Data-Relocatable Image Files
                   14240: @cindex data-relocatable image files
1.26      crook    14241: @cindex image file, data-relocatable
1.1       anton    14242: 
                   14243: These files contain relocatable data addresses, but fixed code addresses
                   14244: (instead of tokens). They are specific to the executable (i.e.,
                   14245: @file{gforth} file) they were created with. For direct threading on some
                   14246: architectures (e.g., the i386), data-relocatable images do not work. You
                   14247: get a data-relocatable image, if you use @file{gforthmi} with a
                   14248: Gforth binary that is not doubly indirect threaded (@pxref{Fully
                   14249: Relocatable Image Files}).
                   14250: 
                   14251: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
                   14252: @section Fully Relocatable Image Files
                   14253: @cindex fully relocatable image files
1.26      crook    14254: @cindex image file, fully relocatable
1.1       anton    14255: 
                   14256: @cindex @file{kern*.fi}, relocatability
                   14257: @cindex @file{gforth.fi}, relocatability
                   14258: These image files have relocatable data addresses, and tokens for code
                   14259: addresses. They can be used with different binaries (e.g., with and
                   14260: without debugging) on the same machine, and even across machines with
                   14261: the same data formats (byte order, cell size, floating point
                   14262: format). However, they are usually specific to the version of Gforth
                   14263: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
                   14264: are fully relocatable.
                   14265: 
                   14266: There are two ways to create a fully relocatable image file:
                   14267: 
                   14268: @menu
1.29      crook    14269: * gforthmi::                    The normal way
1.1       anton    14270: * cross.fs::                    The hard way
                   14271: @end menu
                   14272: 
                   14273: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
                   14274: @subsection @file{gforthmi}
                   14275: @cindex @file{comp-i.fs}
                   14276: @cindex @file{gforthmi}
                   14277: 
                   14278: You will usually use @file{gforthmi}. If you want to create an
1.29      crook    14279: image @i{file} that contains everything you would load by invoking
                   14280: Gforth with @code{gforth @i{options}}, you simply say:
1.1       anton    14281: @example
1.29      crook    14282: gforthmi @i{file} @i{options}
1.1       anton    14283: @end example
                   14284: 
                   14285: E.g., if you want to create an image @file{asm.fi} that has the file
                   14286: @file{asm.fs} loaded in addition to the usual stuff, you could do it
                   14287: like this:
                   14288: 
                   14289: @example
                   14290: gforthmi asm.fi asm.fs
                   14291: @end example
                   14292: 
1.27      crook    14293: @file{gforthmi} is implemented as a sh script and works like this: It
                   14294: produces two non-relocatable images for different addresses and then
                   14295: compares them. Its output reflects this: first you see the output (if
1.62      crook    14296: any) of the two Gforth invocations that produce the non-relocatable image
1.27      crook    14297: files, then you see the output of the comparing program: It displays the
                   14298: offset used for data addresses and the offset used for code addresses;
1.1       anton    14299: moreover, for each cell that cannot be represented correctly in the
1.44      crook    14300: image files, it displays a line like this:
1.1       anton    14301: 
                   14302: @example
                   14303:      78DC         BFFFFA50         BFFFFA40
                   14304: @end example
                   14305: 
                   14306: This means that at offset $78dc from @code{forthstart}, one input image
                   14307: contains $bffffa50, and the other contains $bffffa40. Since these cells
                   14308: cannot be represented correctly in the output image, you should examine
                   14309: these places in the dictionary and verify that these cells are dead
                   14310: (i.e., not read before they are written).
1.39      anton    14311: 
                   14312: @cindex --application, @code{gforthmi} option
                   14313: If you insert the option @code{--application} in front of the image file
                   14314: name, you will get an image that uses the @code{--appl-image} option
                   14315: instead of the @code{--image-file} option (@pxref{Invoking
                   14316: Gforth}). When you execute such an image on Unix (by typing the image
                   14317: name as command), the Gforth engine will pass all options to the image
                   14318: instead of trying to interpret them as engine options.
1.1       anton    14319: 
1.27      crook    14320: If you type @file{gforthmi} with no arguments, it prints some usage
                   14321: instructions.
                   14322: 
1.1       anton    14323: @cindex @code{savesystem} during @file{gforthmi}
                   14324: @cindex @code{bye} during @file{gforthmi}
                   14325: @cindex doubly indirect threaded code
1.44      crook    14326: @cindex environment variables
                   14327: @cindex @code{GFORTHD} -- environment variable
                   14328: @cindex @code{GFORTH} -- environment variable
1.1       anton    14329: @cindex @code{gforth-ditc}
1.29      crook    14330: There are a few wrinkles: After processing the passed @i{options}, the
1.1       anton    14331: words @code{savesystem} and @code{bye} must be visible. A special doubly
                   14332: indirect threaded version of the @file{gforth} executable is used for
1.62      crook    14333: creating the non-relocatable images; you can pass the exact filename of
1.1       anton    14334: this executable through the environment variable @code{GFORTHD}
                   14335: (default: @file{gforth-ditc}); if you pass a version that is not doubly
                   14336: indirect threaded, you will not get a fully relocatable image, but a
1.27      crook    14337: data-relocatable image (because there is no code address offset). The
                   14338: normal @file{gforth} executable is used for creating the relocatable
                   14339: image; you can pass the exact filename of this executable through the
                   14340: environment variable @code{GFORTH}.
1.1       anton    14341: 
                   14342: @node cross.fs,  , gforthmi, Fully Relocatable Image Files
                   14343: @subsection @file{cross.fs}
                   14344: @cindex @file{cross.fs}
                   14345: @cindex cross-compiler
                   14346: @cindex metacompiler
1.47      crook    14347: @cindex target compiler
1.1       anton    14348: 
                   14349: You can also use @code{cross}, a batch compiler that accepts a Forth-like
1.47      crook    14350: programming language (@pxref{Cross Compiler}).
1.1       anton    14351: 
1.47      crook    14352: @code{cross} allows you to create image files for machines with
1.1       anton    14353: different data sizes and data formats than the one used for generating
                   14354: the image file. You can also use it to create an application image that
                   14355: does not contain a Forth compiler. These features are bought with
                   14356: restrictions and inconveniences in programming. E.g., addresses have to
                   14357: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
                   14358: order to make the code relocatable.
                   14359: 
                   14360: 
                   14361: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
                   14362: @section Stack and Dictionary Sizes
                   14363: @cindex image file, stack and dictionary sizes
                   14364: @cindex dictionary size default
                   14365: @cindex stack size default
                   14366: 
                   14367: If you invoke Gforth with a command line flag for the size
                   14368: (@pxref{Invoking Gforth}), the size you specify is stored in the
                   14369: dictionary. If you save the dictionary with @code{savesystem} or create
                   14370: an image with @file{gforthmi}, this size will become the default
                   14371: for the resulting image file. E.g., the following will create a
1.21      crook    14372: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
1.1       anton    14373: 
                   14374: @example
                   14375: gforthmi gforth.fi -m 1M
                   14376: @end example
                   14377: 
                   14378: In other words, if you want to set the default size for the dictionary
                   14379: and the stacks of an image, just invoke @file{gforthmi} with the
                   14380: appropriate options when creating the image.
                   14381: 
                   14382: @cindex stack size, cache-friendly
                   14383: Note: For cache-friendly behaviour (i.e., good performance), you should
                   14384: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
                   14385: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
                   14386: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
                   14387: 
                   14388: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
                   14389: @section Running Image Files
                   14390: @cindex running image files
                   14391: @cindex invoking image files
                   14392: @cindex image file invocation
                   14393: 
                   14394: @cindex -i, invoke image file
                   14395: @cindex --image file, invoke image file
1.29      crook    14396: You can invoke Gforth with an image file @i{image} instead of the
1.1       anton    14397: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
                   14398: @example
1.29      crook    14399: gforth -i @i{image}
1.1       anton    14400: @end example
                   14401: 
                   14402: @cindex executable image file
1.26      crook    14403: @cindex image file, executable
1.1       anton    14404: If your operating system supports starting scripts with a line of the
                   14405: form @code{#! ...}, you just have to type the image file name to start
                   14406: Gforth with this image file (note that the file extension @code{.fi} is
1.29      crook    14407: just a convention). I.e., to run Gforth with the image file @i{image},
                   14408: you can just type @i{image} instead of @code{gforth -i @i{image}}.
1.27      crook    14409: This works because every @code{.fi} file starts with a line of this
                   14410: format:
                   14411: 
                   14412: @example
                   14413: #! /usr/local/bin/gforth-0.4.0 -i
                   14414: @end example
                   14415: 
                   14416: The file and pathname for the Gforth engine specified on this line is
                   14417: the specific Gforth executable that it was built against; i.e. the value
                   14418: of the environment variable @code{GFORTH} at the time that
                   14419: @file{gforthmi} was executed.
1.1       anton    14420: 
1.27      crook    14421: You can make use of the same shell capability to make a Forth source
                   14422: file into an executable. For example, if you place this text in a file:
1.26      crook    14423: 
                   14424: @example
                   14425: #! /usr/local/bin/gforth
                   14426: 
                   14427: ." Hello, world" CR
                   14428: bye
                   14429: @end example
                   14430: 
                   14431: @noindent
1.27      crook    14432: and then make the file executable (chmod +x in Unix), you can run it
1.26      crook    14433: directly from the command line. The sequence @code{#!} is used in two
                   14434: ways; firstly, it is recognised as a ``magic sequence'' by the operating
1.29      crook    14435: system@footnote{The Unix kernel actually recognises two types of files:
                   14436: executable files and files of data, where the data is processed by an
                   14437: interpreter that is specified on the ``interpreter line'' -- the first
                   14438: line of the file, starting with the sequence #!. There may be a small
                   14439: limit (e.g., 32) on the number of characters that may be specified on
                   14440: the interpreter line.} secondly it is treated as a comment character by
                   14441: Gforth. Because of the second usage, a space is required between
1.80      anton    14442: @code{#!} and the path to the executable (moreover, some Unixes
                   14443: require the sequence @code{#! /}).
1.27      crook    14444: 
                   14445: The disadvantage of this latter technique, compared with using
1.80      anton    14446: @file{gforthmi}, is that it is slightly slower; the Forth source code is
                   14447: compiled on-the-fly, each time the program is invoked.
1.26      crook    14448: 
1.1       anton    14449: doc-#!
                   14450: 
1.44      crook    14451: 
1.1       anton    14452: @node Modifying the Startup Sequence,  , Running Image Files, Image Files
                   14453: @section Modifying the Startup Sequence
                   14454: @cindex startup sequence for image file
                   14455: @cindex image file initialization sequence
                   14456: @cindex initialization sequence of image file
                   14457: 
1.120     anton    14458: You can add your own initialization to the startup sequence of an image
                   14459: through the deferred word @code{'cold}. @code{'cold} is invoked just
                   14460: before the image-specific command line processing (i.e., loading files
                   14461: and evaluating (@code{-e}) strings) starts.
1.1       anton    14462: 
                   14463: A sequence for adding your initialization usually looks like this:
                   14464: 
                   14465: @example
                   14466: :noname
                   14467:     Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
                   14468:     ... \ your stuff
                   14469: ; IS 'cold
                   14470: @end example
                   14471: 
                   14472: @cindex turnkey image files
1.26      crook    14473: @cindex image file, turnkey applications
1.1       anton    14474: You can make a turnkey image by letting @code{'cold} execute a word
                   14475: (your turnkey application) that never returns; instead, it exits Gforth
                   14476: via @code{bye} or @code{throw}.
                   14477: 
1.121     anton    14478: You can access the (image-specific) command-line arguments through
                   14479: @code{argc}, @code{argv} and @code{arg} (@pxref{OS command line
                   14480: arguments}).
1.1       anton    14481: 
1.26      crook    14482: If @code{'cold} exits normally, Gforth processes the command-line
                   14483: arguments as files to be loaded and strings to be evaluated.  Therefore,
                   14484: @code{'cold} should remove the arguments it has used in this case.
                   14485: 
                   14486: doc-'cold
1.44      crook    14487: 
1.1       anton    14488: @c ******************************************************************
1.113     anton    14489: @node Engine, Cross Compiler, Image Files, Top
1.1       anton    14490: @chapter Engine
                   14491: @cindex engine
                   14492: @cindex virtual machine
                   14493: 
1.26      crook    14494: Reading this chapter is not necessary for programming with Gforth. It
1.1       anton    14495: may be helpful for finding your way in the Gforth sources.
                   14496: 
1.109     anton    14497: The ideas in this section have also been published in the following
                   14498: papers: Bernd Paysan, @cite{ANS fig/GNU/??? Forth} (in German),
                   14499: Forth-Tagung '93; M. Anton Ertl,
                   14500: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
                   14501: Portable Forth Engine}}, EuroForth '93; M. Anton Ertl,
                   14502: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl02.ps.gz,
                   14503: Threaded code variations and optimizations (extended version)}},
                   14504: Forth-Tagung '02.
1.1       anton    14505: 
                   14506: @menu
                   14507: * Portability::                 
                   14508: * Threading::                   
                   14509: * Primitives::                  
                   14510: * Performance::                 
                   14511: @end menu
                   14512: 
                   14513: @node Portability, Threading, Engine, Engine
                   14514: @section Portability
                   14515: @cindex engine portability
                   14516: 
1.26      crook    14517: An important goal of the Gforth Project is availability across a wide
                   14518: range of personal machines. fig-Forth, and, to a lesser extent, F83,
                   14519: achieved this goal by manually coding the engine in assembly language
                   14520: for several then-popular processors. This approach is very
                   14521: labor-intensive and the results are short-lived due to progress in
                   14522: computer architecture.
1.1       anton    14523: 
                   14524: @cindex C, using C for the engine
                   14525: Others have avoided this problem by coding in C, e.g., Mitch Bradley
                   14526: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
                   14527: particularly popular for UNIX-based Forths due to the large variety of
                   14528: architectures of UNIX machines. Unfortunately an implementation in C
                   14529: does not mix well with the goals of efficiency and with using
                   14530: traditional techniques: Indirect or direct threading cannot be expressed
                   14531: in C, and switch threading, the fastest technique available in C, is
                   14532: significantly slower. Another problem with C is that it is very
                   14533: cumbersome to express double integer arithmetic.
                   14534: 
                   14535: @cindex GNU C for the engine
                   14536: @cindex long long
                   14537: Fortunately, there is a portable language that does not have these
                   14538: limitations: GNU C, the version of C processed by the GNU C compiler
                   14539: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
                   14540: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
                   14541: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
                   14542: threading possible, its @code{long long} type (@pxref{Long Long, ,
                   14543: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
1.109     anton    14544: double numbers on many systems.  GNU C is freely available on all
1.1       anton    14545: important (and many unimportant) UNIX machines, VMS, 80386s running
                   14546: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
                   14547: on all these machines.
                   14548: 
                   14549: Writing in a portable language has the reputation of producing code that
                   14550: is slower than assembly. For our Forth engine we repeatedly looked at
                   14551: the code produced by the compiler and eliminated most compiler-induced
                   14552: inefficiencies by appropriate changes in the source code.
                   14553: 
                   14554: @cindex explicit register declarations
                   14555: @cindex --enable-force-reg, configuration flag
                   14556: @cindex -DFORCE_REG
                   14557: However, register allocation cannot be portably influenced by the
                   14558: programmer, leading to some inefficiencies on register-starved
                   14559: machines. We use explicit register declarations (@pxref{Explicit Reg
                   14560: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
                   14561: improve the speed on some machines. They are turned on by using the
                   14562: configuration flag @code{--enable-force-reg} (@code{gcc} switch
                   14563: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
                   14564: machine, but also on the compiler version: On some machines some
                   14565: compiler versions produce incorrect code when certain explicit register
                   14566: declarations are used. So by default @code{-DFORCE_REG} is not used.
                   14567: 
                   14568: @node Threading, Primitives, Portability, Engine
                   14569: @section Threading
                   14570: @cindex inner interpreter implementation
                   14571: @cindex threaded code implementation
                   14572: 
                   14573: @cindex labels as values
                   14574: GNU C's labels as values extension (available since @code{gcc-2.0},
                   14575: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
1.29      crook    14576: makes it possible to take the address of @i{label} by writing
                   14577: @code{&&@i{label}}.  This address can then be used in a statement like
                   14578: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
1.1       anton    14579: @code{goto x}.
                   14580: 
1.26      crook    14581: @cindex @code{NEXT}, indirect threaded
1.1       anton    14582: @cindex indirect threaded inner interpreter
                   14583: @cindex inner interpreter, indirect threaded
1.26      crook    14584: With this feature an indirect threaded @code{NEXT} looks like:
1.1       anton    14585: @example
                   14586: cfa = *ip++;
                   14587: ca = *cfa;
                   14588: goto *ca;
                   14589: @end example
                   14590: @cindex instruction pointer
                   14591: For those unfamiliar with the names: @code{ip} is the Forth instruction
                   14592: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
                   14593: execution token and points to the code field of the next word to be
                   14594: executed; The @code{ca} (code address) fetched from there points to some
                   14595: executable code, e.g., a primitive or the colon definition handler
                   14596: @code{docol}.
                   14597: 
1.26      crook    14598: @cindex @code{NEXT}, direct threaded
1.1       anton    14599: @cindex direct threaded inner interpreter
                   14600: @cindex inner interpreter, direct threaded
                   14601: Direct threading is even simpler:
                   14602: @example
                   14603: ca = *ip++;
                   14604: goto *ca;
                   14605: @end example
                   14606: 
                   14607: Of course we have packaged the whole thing neatly in macros called
1.26      crook    14608: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
1.1       anton    14609: 
                   14610: @menu
                   14611: * Scheduling::                  
                   14612: * Direct or Indirect Threaded?::  
1.109     anton    14613: * Dynamic Superinstructions::   
1.1       anton    14614: * DOES>::                       
                   14615: @end menu
                   14616: 
                   14617: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
                   14618: @subsection Scheduling
                   14619: @cindex inner interpreter optimization
                   14620: 
                   14621: There is a little complication: Pipelined and superscalar processors,
                   14622: i.e., RISC and some modern CISC machines can process independent
                   14623: instructions while waiting for the results of an instruction. The
                   14624: compiler usually reorders (schedules) the instructions in a way that
                   14625: achieves good usage of these delay slots. However, on our first tries
                   14626: the compiler did not do well on scheduling primitives. E.g., for
                   14627: @code{+} implemented as
                   14628: @example
                   14629: n=sp[0]+sp[1];
                   14630: sp++;
                   14631: sp[0]=n;
                   14632: NEXT;
                   14633: @end example
1.81      anton    14634: the @code{NEXT} comes strictly after the other code, i.e., there is
                   14635: nearly no scheduling. After a little thought the problem becomes clear:
                   14636: The compiler cannot know that @code{sp} and @code{ip} point to different
1.21      crook    14637: addresses (and the version of @code{gcc} we used would not know it even
                   14638: if it was possible), so it could not move the load of the cfa above the
                   14639: store to the TOS. Indeed the pointers could be the same, if code on or
                   14640: very near the top of stack were executed. In the interest of speed we
                   14641: chose to forbid this probably unused ``feature'' and helped the compiler
1.81      anton    14642: in scheduling: @code{NEXT} is divided into several parts:
                   14643: @code{NEXT_P0}, @code{NEXT_P1} and @code{NEXT_P2}). @code{+} now looks
                   14644: like:
1.1       anton    14645: @example
1.81      anton    14646: NEXT_P0;
1.1       anton    14647: n=sp[0]+sp[1];
                   14648: sp++;
                   14649: NEXT_P1;
                   14650: sp[0]=n;
                   14651: NEXT_P2;
                   14652: @end example
                   14653: 
1.81      anton    14654: There are various schemes that distribute the different operations of
                   14655: NEXT between these parts in several ways; in general, different schemes
                   14656: perform best on different processors.  We use a scheme for most
                   14657: architectures that performs well for most processors of this
1.109     anton    14658: architecture; in the future we may switch to benchmarking and chosing
1.81      anton    14659: the scheme on installation time.
                   14660: 
1.1       anton    14661: 
1.109     anton    14662: @node Direct or Indirect Threaded?, Dynamic Superinstructions, Scheduling, Threading
1.1       anton    14663: @subsection Direct or Indirect Threaded?
                   14664: @cindex threading, direct or indirect?
                   14665: 
1.109     anton    14666: Threaded forth code consists of references to primitives (simple machine
                   14667: code routines like @code{+}) and to non-primitives (e.g., colon
                   14668: definitions, variables, constants); for a specific class of
                   14669: non-primitives (e.g., variables) there is one code routine (e.g.,
                   14670: @code{dovar}), but each variable needs a separate reference to its data.
                   14671: 
                   14672: Traditionally Forth has been implemented as indirect threaded code,
                   14673: because this allows to use only one cell to reference a non-primitive
                   14674: (basically you point to the data, and find the code address there).
                   14675: 
                   14676: @cindex primitive-centric threaded code
                   14677: However, threaded code in Gforth (since 0.6.0) uses two cells for
                   14678: non-primitives, one for the code address, and one for the data address;
                   14679: the data pointer is an immediate argument for the virtual machine
                   14680: instruction represented by the code address.  We call this
                   14681: @emph{primitive-centric} threaded code, because all code addresses point
                   14682: to simple primitives.  E.g., for a variable, the code address is for
                   14683: @code{lit} (also used for integer literals like @code{99}).
                   14684: 
                   14685: Primitive-centric threaded code allows us to use (faster) direct
                   14686: threading as dispatch method, completely portably (direct threaded code
                   14687: in Gforth before 0.6.0 required architecture-specific code).  It also
                   14688: eliminates the performance problems related to I-cache consistency that
                   14689: 386 implementations have with direct threaded code, and allows
                   14690: additional optimizations.
                   14691: 
                   14692: @cindex hybrid direct/indirect threaded code
                   14693: There is a catch, however: the @var{xt} parameter of @code{execute} can
                   14694: occupy only one cell, so how do we pass non-primitives with their code
                   14695: @emph{and} data addresses to them?  Our answer is to use indirect
                   14696: threaded dispatch for @code{execute} and other words that use a
                   14697: single-cell xt.  So, normal threaded code in colon definitions uses
                   14698: direct threading, and @code{execute} and similar words, which dispatch
                   14699: to xts on the data stack, use indirect threaded code.  We call this
                   14700: @emph{hybrid direct/indirect} threaded code.
                   14701: 
                   14702: @cindex engines, gforth vs. gforth-fast vs. gforth-itc
                   14703: @cindex gforth engine
                   14704: @cindex gforth-fast engine
                   14705: The engines @command{gforth} and @command{gforth-fast} use hybrid
                   14706: direct/indirect threaded code.  This means that with these engines you
                   14707: cannot use @code{,} to compile an xt.  Instead, you have to use
                   14708: @code{compile,}.
                   14709: 
                   14710: @cindex gforth-itc engine
1.115     anton    14711: If you want to compile xts with @code{,}, use @command{gforth-itc}.
                   14712: This engine uses plain old indirect threaded code.  It still compiles in
                   14713: a primitive-centric style, so you cannot use @code{compile,} instead of
1.109     anton    14714: @code{,} (e.g., for producing tables of xts with @code{] word1 word2
1.115     anton    14715: ... [}).  If you want to do that, you have to use @command{gforth-itc}
1.109     anton    14716: and execute @code{' , is compile,}.  Your program can check if it is
                   14717: running on a hybrid direct/indirect threaded engine or a pure indirect
                   14718: threaded engine with @code{threading-method} (@pxref{Threading Words}).
                   14719: 
                   14720: 
                   14721: @node Dynamic Superinstructions, DOES>, Direct or Indirect Threaded?, Threading
                   14722: @subsection Dynamic Superinstructions
                   14723: @cindex Dynamic superinstructions with replication
                   14724: @cindex Superinstructions
                   14725: @cindex Replication
                   14726: 
                   14727: The engines @command{gforth} and @command{gforth-fast} use another
                   14728: optimization: Dynamic superinstructions with replication.  As an
                   14729: example, consider the following colon definition:
                   14730: 
                   14731: @example
                   14732: : squared ( n1 -- n2 )
                   14733:   dup * ;
                   14734: @end example
                   14735: 
                   14736: Gforth compiles this into the threaded code sequence
                   14737: 
                   14738: @example
                   14739: dup
                   14740: *
                   14741: ;s
                   14742: @end example
                   14743: 
                   14744: In normal direct threaded code there is a code address occupying one
                   14745: cell for each of these primitives.  Each code address points to a
                   14746: machine code routine, and the interpreter jumps to this machine code in
                   14747: order to execute the primitive.  The routines for these three
                   14748: primitives are (in @command{gforth-fast} on the 386):
                   14749: 
                   14750: @example
                   14751: Code dup  
                   14752: ( $804B950 )  add     esi , # -4  \ $83 $C6 $FC 
                   14753: ( $804B953 )  add     ebx , # 4  \ $83 $C3 $4 
                   14754: ( $804B956 )  mov     dword ptr 4 [esi] , ecx  \ $89 $4E $4 
                   14755: ( $804B959 )  jmp     dword ptr FC [ebx]  \ $FF $63 $FC 
                   14756: end-code
                   14757: Code *  
                   14758: ( $804ACC4 )  mov     eax , dword ptr 4 [esi]  \ $8B $46 $4 
                   14759: ( $804ACC7 )  add     esi , # 4  \ $83 $C6 $4 
                   14760: ( $804ACCA )  add     ebx , # 4  \ $83 $C3 $4 
                   14761: ( $804ACCD )  imul    ecx , eax  \ $F $AF $C8 
                   14762: ( $804ACD0 )  jmp     dword ptr FC [ebx]  \ $FF $63 $FC 
                   14763: end-code
                   14764: Code ;s  
                   14765: ( $804A693 )  mov     eax , dword ptr [edi]  \ $8B $7 
                   14766: ( $804A695 )  add     edi , # 4  \ $83 $C7 $4 
                   14767: ( $804A698 )  lea     ebx , dword ptr 4 [eax]  \ $8D $58 $4 
                   14768: ( $804A69B )  jmp     dword ptr FC [ebx]  \ $FF $63 $FC 
                   14769: end-code
                   14770: @end example
                   14771: 
                   14772: With dynamic superinstructions and replication the compiler does not
                   14773: just lay down the threaded code, but also copies the machine code
                   14774: fragments, usually without the jump at the end.
                   14775: 
                   14776: @example
                   14777: ( $4057D27D )  add     esi , # -4  \ $83 $C6 $FC 
                   14778: ( $4057D280 )  add     ebx , # 4  \ $83 $C3 $4 
                   14779: ( $4057D283 )  mov     dword ptr 4 [esi] , ecx  \ $89 $4E $4 
                   14780: ( $4057D286 )  mov     eax , dword ptr 4 [esi]  \ $8B $46 $4 
                   14781: ( $4057D289 )  add     esi , # 4  \ $83 $C6 $4 
                   14782: ( $4057D28C )  add     ebx , # 4  \ $83 $C3 $4 
                   14783: ( $4057D28F )  imul    ecx , eax  \ $F $AF $C8 
                   14784: ( $4057D292 )  mov     eax , dword ptr [edi]  \ $8B $7 
                   14785: ( $4057D294 )  add     edi , # 4  \ $83 $C7 $4 
                   14786: ( $4057D297 )  lea     ebx , dword ptr 4 [eax]  \ $8D $58 $4 
                   14787: ( $4057D29A )  jmp     dword ptr FC [ebx]  \ $FF $63 $FC 
                   14788: @end example
                   14789: 
                   14790: Only when a threaded-code control-flow change happens (e.g., in
                   14791: @code{;s}), the jump is appended.  This optimization eliminates many of
                   14792: these jumps and makes the rest much more predictable.  The speedup
                   14793: depends on the processor and the application; on the Athlon and Pentium
                   14794: III this optimization typically produces a speedup by a factor of 2.
                   14795: 
                   14796: The code addresses in the direct-threaded code are set to point to the
                   14797: appropriate points in the copied machine code, in this example like
                   14798: this:
1.1       anton    14799: 
1.109     anton    14800: @example
                   14801: primitive  code address
                   14802:    dup       $4057D27D
                   14803:    *         $4057D286
                   14804:    ;s        $4057D292
                   14805: @end example
                   14806: 
                   14807: Thus there can be threaded-code jumps to any place in this piece of
                   14808: code.  This also simplifies decompilation quite a bit.
                   14809: 
                   14810: @cindex --no-dynamic command-line option
                   14811: @cindex --no-super command-line option
                   14812: You can disable this optimization with @option{--no-dynamic}.  You can
                   14813: use the copying without eliminating the jumps (i.e., dynamic
                   14814: replication, but without superinstructions) with @option{--no-super};
                   14815: this gives the branch prediction benefit alone; the effect on
1.110     anton    14816: performance depends on the CPU; on the Athlon and Pentium III the
                   14817: speedup is a little less than for dynamic superinstructions with
                   14818: replication.
                   14819: 
                   14820: @cindex patching threaded code
                   14821: One use of these options is if you want to patch the threaded code.
                   14822: With superinstructions, many of the dispatch jumps are eliminated, so
                   14823: patching often has no effect.  These options preserve all the dispatch
                   14824: jumps.
1.109     anton    14825: 
                   14826: @cindex --dynamic command-line option
1.110     anton    14827: On some machines dynamic superinstructions are disabled by default,
                   14828: because it is unsafe on these machines.  However, if you feel
                   14829: adventurous, you can enable it with @option{--dynamic}.
1.109     anton    14830: 
                   14831: @node DOES>,  , Dynamic Superinstructions, Threading
1.1       anton    14832: @subsection DOES>
                   14833: @cindex @code{DOES>} implementation
                   14834: 
1.26      crook    14835: @cindex @code{dodoes} routine
                   14836: @cindex @code{DOES>}-code
1.1       anton    14837: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
                   14838: the chunk of code executed by every word defined by a
1.109     anton    14839: @code{CREATE}...@code{DOES>} pair; actually with primitive-centric code,
                   14840: this is only needed if the xt of the word is @code{execute}d. The main
                   14841: problem here is: How to find the Forth code to be executed, i.e. the
                   14842: code after the @code{DOES>} (the @code{DOES>}-code)? There are two
                   14843: solutions:
1.1       anton    14844: 
1.21      crook    14845: In fig-Forth the code field points directly to the @code{dodoes} and the
1.109     anton    14846: @code{DOES>}-code address is stored in the cell after the code address
                   14847: (i.e. at @code{@i{CFA} cell+}). It may seem that this solution is
                   14848: illegal in the Forth-79 and all later standards, because in fig-Forth
                   14849: this address lies in the body (which is illegal in these
                   14850: standards). However, by making the code field larger for all words this
                   14851: solution becomes legal again.  We use this approach.  Leaving a cell
                   14852: unused in most words is a bit wasteful, but on the machines we are
                   14853: targeting this is hardly a problem.
                   14854: 
1.1       anton    14855: 
                   14856: @node Primitives, Performance, Threading, Engine
                   14857: @section Primitives
                   14858: @cindex primitives, implementation
                   14859: @cindex virtual machine instructions, implementation
                   14860: 
                   14861: @menu
                   14862: * Automatic Generation::        
                   14863: * TOS Optimization::            
                   14864: * Produced code::               
                   14865: @end menu
                   14866: 
                   14867: @node Automatic Generation, TOS Optimization, Primitives, Primitives
                   14868: @subsection Automatic Generation
                   14869: @cindex primitives, automatic generation
                   14870: 
                   14871: @cindex @file{prims2x.fs}
1.109     anton    14872: 
1.1       anton    14873: Since the primitives are implemented in a portable language, there is no
                   14874: longer any need to minimize the number of primitives. On the contrary,
                   14875: having many primitives has an advantage: speed. In order to reduce the
                   14876: number of errors in primitives and to make programming them easier, we
1.109     anton    14877: provide a tool, the primitive generator (@file{prims2x.fs} aka Vmgen,
                   14878: @pxref{Top, Vmgen, Introduction, vmgen, Vmgen}), that automatically
                   14879: generates most (and sometimes all) of the C code for a primitive from
                   14880: the stack effect notation.  The source for a primitive has the following
                   14881: form:
1.1       anton    14882: 
                   14883: @cindex primitive source format
                   14884: @format
1.58      anton    14885: @i{Forth-name}  ( @i{stack-effect} )        @i{category}    [@i{pronounc.}]
1.29      crook    14886: [@code{""}@i{glossary entry}@code{""}]
                   14887: @i{C code}
1.1       anton    14888: [@code{:}
1.29      crook    14889: @i{Forth code}]
1.1       anton    14890: @end format
                   14891: 
                   14892: The items in brackets are optional. The category and glossary fields
                   14893: are there for generating the documentation, the Forth code is there
                   14894: for manual implementations on machines without GNU C. E.g., the source
                   14895: for the primitive @code{+} is:
                   14896: @example
1.58      anton    14897: +    ( n1 n2 -- n )   core    plus
1.1       anton    14898: n = n1+n2;
                   14899: @end example
                   14900: 
                   14901: This looks like a specification, but in fact @code{n = n1+n2} is C
                   14902: code. Our primitive generation tool extracts a lot of information from
                   14903: the stack effect notations@footnote{We use a one-stack notation, even
                   14904: though we have separate data and floating-point stacks; The separate
                   14905: notation can be generated easily from the unified notation.}: The number
                   14906: of items popped from and pushed on the stack, their type, and by what
                   14907: name they are referred to in the C code. It then generates a C code
                   14908: prelude and postlude for each primitive. The final C code for @code{+}
                   14909: looks like this:
                   14910: 
                   14911: @example
1.46      pazsan   14912: I_plus: /* + ( n1 n2 -- n ) */  /* label, stack effect */
1.1       anton    14913: /*  */                          /* documentation */
1.81      anton    14914: NAME("+")                       /* debugging output (with -DDEBUG) */
1.1       anton    14915: @{
                   14916: DEF_CA                          /* definition of variable ca (indirect threading) */
                   14917: Cell n1;                        /* definitions of variables */
                   14918: Cell n2;
                   14919: Cell n;
1.81      anton    14920: NEXT_P0;                        /* NEXT part 0 */
1.1       anton    14921: n1 = (Cell) sp[1];              /* input */
                   14922: n2 = (Cell) TOS;
                   14923: sp += 1;                        /* stack adjustment */
                   14924: @{
                   14925: n = n1+n2;                      /* C code taken from the source */
                   14926: @}
                   14927: NEXT_P1;                        /* NEXT part 1 */
                   14928: TOS = (Cell)n;                  /* output */
                   14929: NEXT_P2;                        /* NEXT part 2 */
                   14930: @}
                   14931: @end example
                   14932: 
                   14933: This looks long and inefficient, but the GNU C compiler optimizes quite
                   14934: well and produces optimal code for @code{+} on, e.g., the R3000 and the
                   14935: HP RISC machines: Defining the @code{n}s does not produce any code, and
                   14936: using them as intermediate storage also adds no cost.
                   14937: 
1.26      crook    14938: There are also other optimizations that are not illustrated by this
                   14939: example: assignments between simple variables are usually for free (copy
1.1       anton    14940: propagation). If one of the stack items is not used by the primitive
                   14941: (e.g.  in @code{drop}), the compiler eliminates the load from the stack
                   14942: (dead code elimination). On the other hand, there are some things that
                   14943: the compiler does not do, therefore they are performed by
                   14944: @file{prims2x.fs}: The compiler does not optimize code away that stores
                   14945: a stack item to the place where it just came from (e.g., @code{over}).
                   14946: 
                   14947: While programming a primitive is usually easy, there are a few cases
                   14948: where the programmer has to take the actions of the generator into
                   14949: account, most notably @code{?dup}, but also words that do not (always)
1.26      crook    14950: fall through to @code{NEXT}.
1.109     anton    14951: 
                   14952: For more information
1.1       anton    14953: 
                   14954: @node TOS Optimization, Produced code, Automatic Generation, Primitives
                   14955: @subsection TOS Optimization
                   14956: @cindex TOS optimization for primitives
                   14957: @cindex primitives, keeping the TOS in a register
                   14958: 
                   14959: An important optimization for stack machine emulators, e.g., Forth
                   14960: engines, is keeping  one or more of the top stack items in
1.29      crook    14961: registers.  If a word has the stack effect @i{in1}...@i{inx} @code{--}
                   14962: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
1.1       anton    14963: @itemize @bullet
                   14964: @item
1.29      crook    14965: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
1.1       anton    14966: due to fewer loads from and stores to the stack.
1.29      crook    14967: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
                   14968: @i{y<n}, due to additional moves between registers.
1.1       anton    14969: @end itemize
                   14970: 
                   14971: @cindex -DUSE_TOS
                   14972: @cindex -DUSE_NO_TOS
                   14973: In particular, keeping one item in a register is never a disadvantage,
                   14974: if there are enough registers. Keeping two items in registers is a
                   14975: disadvantage for frequent words like @code{?branch}, constants,
                   14976: variables, literals and @code{i}. Therefore our generator only produces
                   14977: code that keeps zero or one items in registers. The generated C code
                   14978: covers both cases; the selection between these alternatives is made at
                   14979: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
                   14980: code for @code{+} is just a simple variable name in the one-item case,
                   14981: otherwise it is a macro that expands into @code{sp[0]}. Note that the
                   14982: GNU C compiler tries to keep simple variables like @code{TOS} in
                   14983: registers, and it usually succeeds, if there are enough registers.
                   14984: 
                   14985: @cindex -DUSE_FTOS
                   14986: @cindex -DUSE_NO_FTOS
                   14987: The primitive generator performs the TOS optimization for the
                   14988: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
                   14989: operations the benefit of this optimization is even larger:
                   14990: floating-point operations take quite long on most processors, but can be
                   14991: performed in parallel with other operations as long as their results are
                   14992: not used. If the FP-TOS is kept in a register, this works. If
                   14993: it is kept on the stack, i.e., in memory, the store into memory has to
                   14994: wait for the result of the floating-point operation, lengthening the
                   14995: execution time of the primitive considerably.
                   14996: 
                   14997: The TOS optimization makes the automatic generation of primitives a
                   14998: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
                   14999: @code{TOS} is not sufficient. There are some special cases to
                   15000: consider:
                   15001: @itemize @bullet
                   15002: @item In the case of @code{dup ( w -- w w )} the generator must not
                   15003: eliminate the store to the original location of the item on the stack,
                   15004: if the TOS optimization is turned on.
                   15005: @item Primitives with stack effects of the form @code{--}
1.29      crook    15006: @i{out1}...@i{outy} must store the TOS to the stack at the start.
                   15007: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
1.1       anton    15008: must load the TOS from the stack at the end. But for the null stack
                   15009: effect @code{--} no stores or loads should be generated.
                   15010: @end itemize
                   15011: 
                   15012: @node Produced code,  , TOS Optimization, Primitives
                   15013: @subsection Produced code
                   15014: @cindex primitives, assembly code listing
                   15015: 
                   15016: @cindex @file{engine.s}
                   15017: To see what assembly code is produced for the primitives on your machine
                   15018: with your compiler and your flag settings, type @code{make engine.s} and
1.81      anton    15019: look at the resulting file @file{engine.s}.  Alternatively, you can also
                   15020: disassemble the code of primitives with @code{see} on some architectures.
1.1       anton    15021: 
                   15022: @node  Performance,  , Primitives, Engine
                   15023: @section Performance
                   15024: @cindex performance of some Forth interpreters
                   15025: @cindex engine performance
                   15026: @cindex benchmarking Forth systems
                   15027: @cindex Gforth performance
                   15028: 
                   15029: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
1.112     anton    15030: impossible to write a significantly faster threaded-code engine.
1.1       anton    15031: 
                   15032: On register-starved machines like the 386 architecture processors
                   15033: improvements are possible, because @code{gcc} does not utilize the
                   15034: registers as well as a human, even with explicit register declarations;
                   15035: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
                   15036: and hand-tuned it for the 486; this system is 1.19 times faster on the
                   15037: Sieve benchmark on a 486DX2/66 than Gforth compiled with
1.40      anton    15038: @code{gcc-2.6.3} with @code{-DFORCE_REG}.  The situation has improved
                   15039: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
                   15040: registers fit in real registers (and we can even afford to use the TOS
                   15041: optimization), resulting in a speedup of 1.14 on the sieve over the
1.112     anton    15042: earlier results.  And dynamic superinstructions provide another speedup
                   15043: (but only around a factor 1.2 on the 486).
1.1       anton    15044: 
                   15045: @cindex Win32Forth performance
                   15046: @cindex NT Forth performance
                   15047: @cindex eforth performance
                   15048: @cindex ThisForth performance
                   15049: @cindex PFE performance
                   15050: @cindex TILE performance
1.81      anton    15051: The potential advantage of assembly language implementations is not
1.112     anton    15052: necessarily realized in complete Forth systems: We compared Gforth-0.5.9
1.81      anton    15053: (direct threaded, compiled with @code{gcc-2.95.1} and
                   15054: @code{-DFORCE_REG}) with Win32Forth 1.2093 (newer versions are
                   15055: reportedly much faster), LMI's NT Forth (Beta, May 1994) and Eforth
                   15056: (with and without peephole (aka pinhole) optimization of the threaded
                   15057: code); all these systems were written in assembly language. We also
                   15058: compared Gforth with three systems written in C: PFE-0.9.14 (compiled
                   15059: with @code{gcc-2.6.3} with the default configuration for Linux:
                   15060: @code{-O2 -fomit-frame-pointer -DUSE_REGS -DUNROLL_NEXT}), ThisForth
                   15061: Beta (compiled with @code{gcc-2.6.3 -O3 -fomit-frame-pointer}; ThisForth
                   15062: employs peephole optimization of the threaded code) and TILE (compiled
                   15063: with @code{make opt}). We benchmarked Gforth, PFE, ThisForth and TILE on
                   15064: a 486DX2/66 under Linux. Kenneth O'Heskin kindly provided the results
                   15065: for Win32Forth and NT Forth on a 486DX2/66 with similar memory
                   15066: performance under Windows NT. Marcel Hendrix ported Eforth to Linux,
                   15067: then extended it to run the benchmarks, added the peephole optimizer,
                   15068: ran the benchmarks and reported the results.
1.40      anton    15069: 
1.1       anton    15070: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
                   15071: matrix multiplication come from the Stanford integer benchmarks and have
                   15072: been translated into Forth by Martin Fraeman; we used the versions
                   15073: included in the TILE Forth package, but with bigger data set sizes; and
                   15074: a recursive Fibonacci number computation for benchmarking calling
                   15075: performance. The following table shows the time taken for the benchmarks
                   15076: scaled by the time taken by Gforth (in other words, it shows the speedup
                   15077: factor that Gforth achieved over the other systems).
                   15078: 
                   15079: @example
1.112     anton    15080: relative       Win32-    NT       eforth       This-      
                   15081: time     Gforth Forth Forth eforth  +opt   PFE Forth  TILE
                   15082: sieve      1.00  2.16  1.78   2.16  1.32  2.46  4.96 13.37
                   15083: bubble     1.00  1.93  2.07   2.18  1.29  2.21        5.70
                   15084: matmul     1.00  1.92  1.76   1.90  0.96  2.06        5.32
                   15085: fib        1.00  2.32  2.03   1.86  1.31  2.64  4.55  6.54
1.1       anton    15086: @end example
                   15087: 
1.26      crook    15088: You may be quite surprised by the good performance of Gforth when
                   15089: compared with systems written in assembly language. One important reason
                   15090: for the disappointing performance of these other systems is probably
                   15091: that they are not written optimally for the 486 (e.g., they use the
                   15092: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
                   15093: but costly method for relocating the Forth image: like @code{cforth}, it
                   15094: computes the actual addresses at run time, resulting in two address
                   15095: computations per @code{NEXT} (@pxref{Image File Background}).
                   15096: 
1.1       anton    15097: The speedup of Gforth over PFE, ThisForth and TILE can be easily
                   15098: explained with the self-imposed restriction of the latter systems to
                   15099: standard C, which makes efficient threading impossible (however, the
1.4       anton    15100: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
1.1       anton    15101: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
                   15102: Moreover, current C compilers have a hard time optimizing other aspects
                   15103: of the ThisForth and the TILE source.
                   15104: 
1.26      crook    15105: The performance of Gforth on 386 architecture processors varies widely
                   15106: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
                   15107: allocate any of the virtual machine registers into real machine
                   15108: registers by itself and would not work correctly with explicit register
1.112     anton    15109: declarations, giving a significantly slower engine (on a 486DX2/66
                   15110: running the Sieve) than the one measured above.
1.1       anton    15111: 
1.26      crook    15112: Note that there have been several releases of Win32Forth since the
                   15113: release presented here, so the results presented above may have little
1.40      anton    15114: predictive value for the performance of Win32Forth today (results for
                   15115: the current release on an i486DX2/66 are welcome).
1.1       anton    15116: 
                   15117: @cindex @file{Benchres}
1.66      anton    15118: In
                   15119: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
                   15120: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
1.1       anton    15121: Maierhofer (presented at EuroForth '95), an indirect threaded version of
1.66      anton    15122: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
                   15123: several native code systems; that version of Gforth is slower on a 486
1.112     anton    15124: than the version used here. You can find a newer version of these
                   15125: measurements at
1.47      crook    15126: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
1.1       anton    15127: find numbers for Gforth on various machines in @file{Benchres}.
                   15128: 
1.26      crook    15129: @c ******************************************************************
1.113     anton    15130: @c @node Binding to System Library, Cross Compiler, Engine, Top
                   15131: @c @chapter Binding to System Library
1.13      pazsan   15132: 
1.113     anton    15133: @c ****************************************************************
                   15134: @node Cross Compiler, Bugs, Engine, Top
1.14      pazsan   15135: @chapter Cross Compiler
1.47      crook    15136: @cindex @file{cross.fs}
                   15137: @cindex cross-compiler
                   15138: @cindex metacompiler
                   15139: @cindex target compiler
1.13      pazsan   15140: 
1.46      pazsan   15141: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
                   15142: mostly written in Forth, including crucial parts like the outer
                   15143: interpreter and compiler, it needs compiled Forth code to get
                   15144: started. The cross compiler allows to create new images for other
                   15145: architectures, even running under another Forth system.
1.13      pazsan   15146: 
                   15147: @menu
1.67      anton    15148: * Using the Cross Compiler::    
                   15149: * How the Cross Compiler Works::  
1.13      pazsan   15150: @end menu
                   15151: 
1.21      crook    15152: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
1.14      pazsan   15153: @section Using the Cross Compiler
1.46      pazsan   15154: 
                   15155: The cross compiler uses a language that resembles Forth, but isn't. The
                   15156: main difference is that you can execute Forth code after definition,
                   15157: while you usually can't execute the code compiled by cross, because the
                   15158: code you are compiling is typically for a different computer than the
                   15159: one you are compiling on.
                   15160: 
1.81      anton    15161: @c anton: This chapter is somewhat different from waht I would expect: I
                   15162: @c would expect an explanation of the cross language and how to create an
                   15163: @c application image with it.  The section explains some aspects of
                   15164: @c creating a Gforth kernel.
                   15165: 
1.46      pazsan   15166: The Makefile is already set up to allow you to create kernels for new
                   15167: architectures with a simple make command. The generic kernels using the
                   15168: GCC compiled virtual machine are created in the normal build process
                   15169: with @code{make}. To create a embedded Gforth executable for e.g. the
                   15170: 8086 processor (running on a DOS machine), type
                   15171: 
                   15172: @example
                   15173: make kernl-8086.fi
                   15174: @end example
                   15175: 
                   15176: This will use the machine description from the @file{arch/8086}
                   15177: directory to create a new kernel. A machine file may look like that:
                   15178: 
                   15179: @example
                   15180: \ Parameter for target systems                         06oct92py
                   15181: 
                   15182:     4 Constant cell             \ cell size in bytes
                   15183:     2 Constant cell<<           \ cell shift to bytes
                   15184:     5 Constant cell>bit         \ cell shift to bits
                   15185:     8 Constant bits/char        \ bits per character
                   15186:     8 Constant bits/byte        \ bits per byte [default: 8]
                   15187:     8 Constant float            \ bytes per float
                   15188:     8 Constant /maxalign        \ maximum alignment in bytes
                   15189: false Constant bigendian        \ byte order
                   15190: ( true=big, false=little )
                   15191: 
                   15192: include machpc.fs               \ feature list
                   15193: @end example
                   15194: 
                   15195: This part is obligatory for the cross compiler itself, the feature list
                   15196: is used by the kernel to conditionally compile some features in and out,
                   15197: depending on whether the target supports these features.
                   15198: 
                   15199: There are some optional features, if you define your own primitives,
                   15200: have an assembler, or need special, nonstandard preparation to make the
1.81      anton    15201: boot process work. @code{asm-include} includes an assembler,
1.46      pazsan   15202: @code{prims-include} includes primitives, and @code{>boot} prepares for
                   15203: booting.
                   15204: 
                   15205: @example
                   15206: : asm-include    ." Include assembler" cr
                   15207:   s" arch/8086/asm.fs" included ;
                   15208: 
                   15209: : prims-include  ." Include primitives" cr
                   15210:   s" arch/8086/prim.fs" included ;
                   15211: 
                   15212: : >boot          ." Prepare booting" cr
                   15213:   s" ' boot >body into-forth 1+ !" evaluate ;
                   15214: @end example
                   15215: 
                   15216: These words are used as sort of macro during the cross compilation in
1.81      anton    15217: the file @file{kernel/main.fs}. Instead of using these macros, it would
1.46      pazsan   15218: be possible --- but more complicated --- to write a new kernel project
                   15219: file, too.
                   15220: 
                   15221: @file{kernel/main.fs} expects the machine description file name on the
                   15222: stack; the cross compiler itself (@file{cross.fs}) assumes that either
                   15223: @code{mach-file} leaves a counted string on the stack, or
                   15224: @code{machine-file} leaves an address, count pair of the filename on the
                   15225: stack.
                   15226: 
                   15227: The feature list is typically controlled using @code{SetValue}, generic
                   15228: files that are used by several projects can use @code{DefaultValue}
                   15229: instead. Both functions work like @code{Value}, when the value isn't
                   15230: defined, but @code{SetValue} works like @code{to} if the value is
                   15231: defined, and @code{DefaultValue} doesn't set anything, if the value is
                   15232: defined.
                   15233: 
                   15234: @example
                   15235: \ generic mach file for pc gforth                       03sep97jaw
                   15236: 
                   15237: true DefaultValue NIL  \ relocating
                   15238: 
                   15239: >ENVIRON
                   15240: 
                   15241: true DefaultValue file          \ controls the presence of the
                   15242:                                 \ file access wordset
                   15243: true DefaultValue OS            \ flag to indicate a operating system
                   15244: 
                   15245: true DefaultValue prims         \ true: primitives are c-code
                   15246: 
                   15247: true DefaultValue floating      \ floating point wordset is present
                   15248: 
                   15249: true DefaultValue glocals       \ gforth locals are present
                   15250:                                 \ will be loaded
                   15251: true DefaultValue dcomps        \ double number comparisons
                   15252: 
                   15253: true DefaultValue hash          \ hashing primitives are loaded/present
                   15254: 
                   15255: true DefaultValue xconds        \ used together with glocals,
                   15256:                                 \ special conditionals supporting gforths'
                   15257:                                 \ local variables
                   15258: true DefaultValue header        \ save a header information
                   15259: 
                   15260: true DefaultValue backtrace     \ enables backtrace code
                   15261: 
                   15262: false DefaultValue ec
                   15263: false DefaultValue crlf
                   15264: 
                   15265: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
                   15266: 
                   15267: &16 KB          DefaultValue stack-size
                   15268: &15 KB &512 +   DefaultValue fstack-size
                   15269: &15 KB          DefaultValue rstack-size
                   15270: &14 KB &512 +   DefaultValue lstack-size
                   15271: @end example
1.13      pazsan   15272: 
1.48      anton    15273: @node How the Cross Compiler Works,  , Using the Cross Compiler, Cross Compiler
1.14      pazsan   15274: @section How the Cross Compiler Works
1.13      pazsan   15275: 
                   15276: @node Bugs, Origin, Cross Compiler, Top
1.21      crook    15277: @appendix Bugs
1.1       anton    15278: @cindex bug reporting
                   15279: 
1.21      crook    15280: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
1.1       anton    15281: 
1.103     anton    15282: If you find a bug, please submit a bug report through
                   15283: @uref{https://savannah.gnu.org/bugs/?func=addbug&group=gforth}.
1.21      crook    15284: 
                   15285: @itemize @bullet
                   15286: @item
1.81      anton    15287: A program (or a sequence of keyboard commands) that reproduces the bug.
                   15288: @item
                   15289: A description of what you think constitutes the buggy behaviour.
                   15290: @item
1.21      crook    15291: The Gforth version used (it is announced at the start of an
                   15292: interactive Gforth session).
                   15293: @item
                   15294: The machine and operating system (on Unix
                   15295: systems @code{uname -a} will report this information).
                   15296: @item
1.81      anton    15297: The installation options (you can find the configure options at the
                   15298: start of @file{config.status}) and configuration (@code{configure}
                   15299: output or @file{config.cache}).
1.21      crook    15300: @item
                   15301: A complete list of changes (if any) you (or your installer) have made to the
                   15302: Gforth sources.
                   15303: @end itemize
1.1       anton    15304: 
                   15305: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
                   15306: to Report Bugs, gcc.info, GNU C Manual}.
                   15307: 
                   15308: 
1.21      crook    15309: @node Origin, Forth-related information, Bugs, Top
                   15310: @appendix Authors and Ancestors of Gforth
1.1       anton    15311: 
                   15312: @section Authors and Contributors
                   15313: @cindex authors of Gforth
                   15314: @cindex contributors to Gforth
                   15315: 
                   15316: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
1.81      anton    15317: Ertl. The third major author was Jens Wilke.  Neal Crook contributed a
                   15318: lot to the manual.  Assemblers and disassemblers were contributed by
                   15319: Andrew McKewan, Christian Pirker, and Bernd Thallner.  Lennart Benschop
                   15320: (who was one of Gforth's first users, in mid-1993) and Stuart Ramsden
                   15321: inspired us with their continuous feedback. Lennart Benshop contributed
1.1       anton    15322: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
                   15323: support for calling C libraries. Helpful comments also came from Paul
                   15324: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
1.113     anton    15325: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, Robert
                   15326: Epprecht, Dennis Ruffer and David N. Williams. Since the release of
                   15327: Gforth-0.2.1 there were also helpful comments from many others; thank
                   15328: you all, sorry for not listing you here (but digging through my mailbox
                   15329: to extract your names is on my to-do list).
1.1       anton    15330: 
                   15331: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
                   15332: and autoconf, among others), and to the creators of the Internet: Gforth
1.21      crook    15333: was developed across the Internet, and its authors did not meet
1.20      pazsan   15334: physically for the first 4 years of development.
1.1       anton    15335: 
                   15336: @section Pedigree
1.26      crook    15337: @cindex pedigree of Gforth
1.1       anton    15338: 
1.81      anton    15339: Gforth descends from bigFORTH (1993) and fig-Forth.  Of course, a
                   15340: significant part of the design of Gforth was prescribed by ANS Forth.
1.1       anton    15341: 
1.20      pazsan   15342: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
1.1       anton    15343: 32 bit native code version of VolksForth for the Atari ST, written
                   15344: mostly by Dietrich Weineck.
                   15345: 
1.81      anton    15346: VolksForth was written by Klaus Schleisiek, Bernd Pennemann, Georg
                   15347: Rehfeld and Dietrich Weineck for the C64 (called UltraForth there) in
                   15348: the mid-80s and ported to the Atari ST in 1986.  It descends from F83.
1.1       anton    15349: 
                   15350: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
                   15351: Forth-83 standard. !! Pedigree? When?
                   15352: 
                   15353: A team led by Bill Ragsdale implemented fig-Forth on many processors in
                   15354: 1979. Robert Selzer and Bill Ragsdale developed the original
                   15355: implementation of fig-Forth for the 6502 based on microForth.
                   15356: 
                   15357: The principal architect of microForth was Dean Sanderson. microForth was
                   15358: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
                   15359: the 1802, and subsequently implemented on the 8080, the 6800 and the
                   15360: Z80.
                   15361: 
                   15362: All earlier Forth systems were custom-made, usually by Charles Moore,
                   15363: who discovered (as he puts it) Forth during the late 60s. The first full
                   15364: Forth existed in 1971.
                   15365: 
1.81      anton    15366: A part of the information in this section comes from
                   15367: @cite{@uref{http://www.forth.com/Content/History/History1.htm,The
                   15368: Evolution of Forth}} by Elizabeth D. Rather, Donald R. Colburn and
                   15369: Charles H. Moore, presented at the HOPL-II conference and preprinted in
                   15370: SIGPLAN Notices 28(3), 1993.  You can find more historical and
                   15371: genealogical information about Forth there.
1.1       anton    15372: 
1.81      anton    15373: @c ------------------------------------------------------------------
1.113     anton    15374: @node Forth-related information, Licenses, Origin, Top
1.21      crook    15375: @appendix Other Forth-related information
                   15376: @cindex Forth-related information
                   15377: 
1.81      anton    15378: @c anton: I threw most of this stuff out, because it can be found through
                   15379: @c the FAQ and the FAQ is more likely to be up-to-date.
1.21      crook    15380: 
                   15381: @cindex comp.lang.forth
                   15382: @cindex frequently asked questions
1.81      anton    15383: There is an active news group (comp.lang.forth) discussing Forth
                   15384: (including Gforth) and Forth-related issues. Its
                   15385: @uref{http://www.complang.tuwien.ac.at/forth/faq/faq-general-2.html,FAQs}
                   15386: (frequently asked questions and their answers) contains a lot of
                   15387: information on Forth.  You should read it before posting to
                   15388: comp.lang.forth.
1.21      crook    15389: 
1.81      anton    15390: The ANS Forth standard is most usable in its
                   15391: @uref{http://www.taygeta.com/forth/dpans.html, HTML form}.
1.21      crook    15392: 
1.113     anton    15393: @c ---------------------------------------------------
                   15394: @node  Licenses, Word Index, Forth-related information, Top
                   15395: @appendix Licenses
                   15396: 
                   15397: @menu
                   15398: * GNU Free Documentation License::  License for copying this manual.
                   15399: * Copying::                         GPL (for copying this software).
                   15400: @end menu
                   15401: 
                   15402: @include fdl.texi
                   15403: 
                   15404: @include gpl.texi
                   15405: 
                   15406: 
                   15407: 
1.81      anton    15408: @c ------------------------------------------------------------------
1.113     anton    15409: @node Word Index, Concept Index, Licenses, Top
1.1       anton    15410: @unnumbered Word Index
                   15411: 
1.26      crook    15412: This index is a list of Forth words that have ``glossary'' entries
                   15413: within this manual. Each word is listed with its stack effect and
                   15414: wordset.
1.1       anton    15415: 
                   15416: @printindex fn
                   15417: 
1.81      anton    15418: @c anton: the name index seems superfluous given the word and concept indices.
                   15419: 
                   15420: @c @node Name Index, Concept Index, Word Index, Top
                   15421: @c @unnumbered Name Index
1.41      anton    15422: 
1.81      anton    15423: @c This index is a list of Forth words that have ``glossary'' entries
                   15424: @c within this manual.
1.41      anton    15425: 
1.81      anton    15426: @c @printindex ky
1.41      anton    15427: 
1.113     anton    15428: @c -------------------------------------------------------
1.81      anton    15429: @node Concept Index,  , Word Index, Top
1.1       anton    15430: @unnumbered Concept and Word Index
                   15431: 
1.26      crook    15432: Not all entries listed in this index are present verbatim in the
                   15433: text. This index also duplicates, in abbreviated form, all of the words
                   15434: listed in the Word Index (only the names are listed for the words here).
1.1       anton    15435: 
                   15436: @printindex cp
                   15437: 
                   15438: @bye
1.81      anton    15439: 
                   15440: 
1.1       anton    15441: 

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>