Annotation of gforth/doc/gforth.ds, revision 1.91

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
1.29      crook      14: @comment .. would be useful to have a word that identified all deferred words
                     15: @comment should semantics stuff in intro be moved to another section
                     16: 
1.66      anton      17: @c POSTPONE, COMPILE, [COMPILE], LITERAL should have their own section
1.28      crook      18: 
1.1       anton      19: @comment %**start of header (This is for running Texinfo on a region.)
                     20: @setfilename gforth.info
                     21: @settitle Gforth Manual
                     22: @dircategory GNU programming tools
                     23: @direntry
                     24: * Gforth: (gforth).             A fast interpreter for the Forth language.
                     25: @end direntry
1.49      anton      26: @c The Texinfo manual also recommends doing this, but for Gforth it may
                     27: @c  not make much sense
                     28: @c @dircategory Individual utilities
                     29: @c @direntry
                     30: @c * Gforth: (gforth)Invoking Gforth.      gforth, gforth-fast, gforthmi
                     31: @c @end direntry
                     32: 
1.1       anton      33: @comment @setchapternewpage odd
1.29      crook      34: @comment TODO this gets left in by HTML converter
1.12      anton      35: @macro progstyle {}
                     36: Programming style note:
1.3       anton      37: @end macro
1.48      anton      38: 
                     39: @macro assignment {}
                     40: @table @i
                     41: @item Assignment:
                     42: @end macro
                     43: @macro endassignment {}
                     44: @end table
                     45: @end macro
                     46: 
1.1       anton      47: @comment %**end of header (This is for running Texinfo on a region.)
                     48: 
1.29      crook      49: 
                     50: @comment ----------------------------------------------------------
                     51: @comment macros for beautifying glossary entries
                     52: @comment if these are used, need to strip them out for HTML converter
                     53: @comment else they get repeated verbatim in HTML output.
                     54: @comment .. not working yet.
                     55: 
                     56: @macro GLOSS-START {}
                     57: @iftex
                     58: @ninerm
                     59: @end iftex
                     60: @end macro
                     61: 
                     62: @macro GLOSS-END {}
                     63: @iftex
                     64: @rm
                     65: @end iftex
                     66: @end macro
                     67: 
                     68: @comment ----------------------------------------------------------
                     69: 
                     70: 
1.10      anton      71: @include version.texi
                     72: 
1.49      anton      73: @ifnottex
1.11      anton      74: This file documents Gforth @value{VERSION}
1.1       anton      75: 
1.62      crook      76: Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
1.1       anton      77: 
                     78:      Permission is granted to make and distribute verbatim copies of
                     79:      this manual provided the copyright notice and this permission notice
                     80:      are preserved on all copies.
                     81:      
                     82: @ignore
                     83:      Permission is granted to process this file through TeX and print the
                     84:      results, provided the printed document carries a copying permission
                     85:      notice identical to this one except for the removal of this paragraph
                     86:      (this paragraph not being relevant to the printed manual).
                     87:      
                     88: @end ignore
                     89:      Permission is granted to copy and distribute modified versions of this
                     90:      manual under the conditions for verbatim copying, provided also that the
                     91:      sections entitled "Distribution" and "General Public License" are
                     92:      included exactly as in the original, and provided that the entire
                     93:      resulting derived work is distributed under the terms of a permission
                     94:      notice identical to this one.
                     95:      
                     96:      Permission is granted to copy and distribute translations of this manual
                     97:      into another language, under the above conditions for modified versions,
                     98:      except that the sections entitled "Distribution" and "General Public
                     99:      License" may be included in a translation approved by the author instead
                    100:      of in the original English.
1.49      anton     101: @end ifnottex
1.1       anton     102: 
                    103: @finalout
                    104: @titlepage
                    105: @sp 10
                    106: @center @titlefont{Gforth Manual}
                    107: @sp 2
1.11      anton     108: @center for version @value{VERSION}
1.1       anton     109: @sp 2
1.34      anton     110: @center Neal Crook
1.1       anton     111: @center Anton Ertl
1.6       pazsan    112: @center Bernd Paysan
1.5       anton     113: @center Jens Wilke
1.1       anton     114: @sp 3
1.47      crook     115: @center This manual is permanently under construction and was last updated on 15-Mar-2000
1.1       anton     116: 
                    117: @comment  The following two commands start the copyright page.
                    118: @page
                    119: @vskip 0pt plus 1filll
1.62      crook     120: Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
1.1       anton     121: 
                    122: @comment !! Published by ... or You can get a copy of this manual ...
                    123: 
                    124:      Permission is granted to make and distribute verbatim copies of
                    125:      this manual provided the copyright notice and this permission notice
                    126:      are preserved on all copies.
                    127:      
                    128:      Permission is granted to copy and distribute modified versions of this
                    129:      manual under the conditions for verbatim copying, provided also that the
                    130:      sections entitled "Distribution" and "General Public License" are
                    131:      included exactly as in the original, and provided that the entire
                    132:      resulting derived work is distributed under the terms of a permission
                    133:      notice identical to this one.
                    134:      
                    135:      Permission is granted to copy and distribute translations of this manual
                    136:      into another language, under the above conditions for modified versions,
                    137:      except that the sections entitled "Distribution" and "General Public
                    138:      License" may be included in a translation approved by the author instead
                    139:      of in the original English.
                    140: @end titlepage
                    141: 
                    142: @node Top, License, (dir), (dir)
1.49      anton     143: @ifnottex
1.1       anton     144: Gforth is a free implementation of ANS Forth available on many
1.11      anton     145: personal machines. This manual corresponds to version @value{VERSION}.
1.49      anton     146: @end ifnottex
1.1       anton     147: 
                    148: @menu
1.21      crook     149: * License::                     The GPL
1.26      crook     150: * Goals::                       About the Gforth Project
1.29      crook     151: * Gforth Environment::          Starting (and exiting) Gforth
1.48      anton     152: * Tutorial::                    Hands-on Forth Tutorial
1.21      crook     153: * Introduction::                An introduction to ANS Forth
1.1       anton     154: * Words::                       Forth words available in Gforth
1.24      anton     155: * Error messages::              How to interpret them
1.1       anton     156: * Tools::                       Programming tools
                    157: * ANS conformance::             Implementation-defined options etc.
1.65      anton     158: * Standard vs Extensions::      Should I use extensions?
1.1       anton     159: * Model::                       The abstract machine of Gforth
                    160: * Integrating Gforth::          Forth as scripting language for applications
                    161: * Emacs and Gforth::            The Gforth Mode
                    162: * Image Files::                 @code{.fi} files contain compiled code
                    163: * Engine::                      The inner interpreter and the primitives
1.24      anton     164: * Binding to System Library::   
1.13      pazsan    165: * Cross Compiler::              The Cross Compiler
1.1       anton     166: * Bugs::                        How to report them
                    167: * Origin::                      Authors and ancestors of Gforth
1.21      crook     168: * Forth-related information::   Books and places to look on the WWW
1.1       anton     169: * Word Index::                  An item for each Forth word
                    170: * Concept Index::               A menu covering many topics
1.12      anton     171: 
1.91    ! anton     172: @detailmenu
        !           173:  --- The Detailed Node Listing ---
1.12      anton     174: 
1.29      crook     175: Gforth Environment
                    176: 
1.32      anton     177: * Invoking Gforth::             Getting in
                    178: * Leaving Gforth::              Getting out
                    179: * Command-line editing::        
1.48      anton     180: * Environment variables::       that affect how Gforth starts up
1.32      anton     181: * Gforth Files::                What gets installed and where
1.48      anton     182: * Startup speed::               When 35ms is not fast enough ...
                    183: 
                    184: Forth Tutorial
                    185: 
                    186: * Starting Gforth Tutorial::    
                    187: * Syntax Tutorial::             
                    188: * Crash Course Tutorial::       
                    189: * Stack Tutorial::              
                    190: * Arithmetics Tutorial::        
                    191: * Stack Manipulation Tutorial::  
                    192: * Using files for Forth code Tutorial::  
                    193: * Comments Tutorial::           
                    194: * Colon Definitions Tutorial::  
                    195: * Decompilation Tutorial::      
                    196: * Stack-Effect Comments Tutorial::  
                    197: * Types Tutorial::              
                    198: * Factoring Tutorial::          
                    199: * Designing the stack effect Tutorial::  
                    200: * Local Variables Tutorial::    
                    201: * Conditional execution Tutorial::  
                    202: * Flags and Comparisons Tutorial::  
                    203: * General Loops Tutorial::      
                    204: * Counted loops Tutorial::      
                    205: * Recursion Tutorial::          
                    206: * Leaving definitions or loops Tutorial::  
                    207: * Return Stack Tutorial::       
                    208: * Memory Tutorial::             
                    209: * Characters and Strings Tutorial::  
                    210: * Alignment Tutorial::          
1.87      anton     211: * Files Tutorial::              
1.48      anton     212: * Interpretation and Compilation Semantics and Immediacy Tutorial::  
                    213: * Execution Tokens Tutorial::   
                    214: * Exceptions Tutorial::         
                    215: * Defining Words Tutorial::     
                    216: * Arrays and Records Tutorial::  
                    217: * POSTPONE Tutorial::           
                    218: * Literal Tutorial::            
                    219: * Advanced macros Tutorial::    
                    220: * Compilation Tokens Tutorial::  
                    221: * Wordlists and Search Order Tutorial::  
1.29      crook     222: 
1.24      anton     223: An Introduction to ANS Forth
                    224: 
1.67      anton     225: * Introducing the Text Interpreter::  
                    226: * Stacks and Postfix notation::  
                    227: * Your first definition::       
                    228: * How does that work?::         
                    229: * Forth is written in Forth::   
                    230: * Review - elements of a Forth system::  
                    231: * Where to go next::            
                    232: * Exercises::                   
1.24      anton     233: 
1.12      anton     234: Forth Words
                    235: 
                    236: * Notation::                    
1.65      anton     237: * Case insensitivity::          
                    238: * Comments::                    
                    239: * Boolean Flags::               
1.12      anton     240: * Arithmetic::                  
                    241: * Stack Manipulation::          
                    242: * Memory::                      
                    243: * Control Structures::          
                    244: * Defining Words::              
1.65      anton     245: * Interpretation and Compilation Semantics::  
1.47      crook     246: * Tokens for Words::            
1.81      anton     247: * Compiling words::             
1.65      anton     248: * The Text Interpreter::        
                    249: * Word Lists::                  
                    250: * Environmental Queries::       
1.12      anton     251: * Files::                       
                    252: * Blocks::                      
                    253: * Other I/O::                   
1.78      anton     254: * Locals::                      
                    255: * Structures::                  
                    256: * Object-oriented Forth::       
1.12      anton     257: * Programming Tools::           
                    258: * Assembler and Code Words::    
                    259: * Threading Words::             
1.65      anton     260: * Passing Commands to the OS::  
                    261: * Keeping track of Time::       
                    262: * Miscellaneous Words::         
1.12      anton     263: 
                    264: Arithmetic
                    265: 
                    266: * Single precision::            
1.67      anton     267: * Double precision::            Double-cell integer arithmetic
1.12      anton     268: * Bitwise operations::          
1.67      anton     269: * Numeric comparison::          
1.32      anton     270: * Mixed precision::             Operations with single and double-cell integers
1.12      anton     271: * Floating Point::              
                    272: 
                    273: Stack Manipulation
                    274: 
                    275: * Data stack::                  
                    276: * Floating point stack::        
                    277: * Return stack::                
                    278: * Locals stack::                
                    279: * Stack pointer manipulation::  
                    280: 
                    281: Memory
                    282: 
1.32      anton     283: * Memory model::                
                    284: * Dictionary allocation::       
                    285: * Heap Allocation::             
                    286: * Memory Access::               
                    287: * Address arithmetic::          
                    288: * Memory Blocks::               
1.12      anton     289: 
                    290: Control Structures
                    291: 
1.41      anton     292: * Selection::                   IF ... ELSE ... ENDIF
                    293: * Simple Loops::                BEGIN ...
1.32      anton     294: * Counted Loops::               DO
1.67      anton     295: * Arbitrary control structures::  
                    296: * Calls and returns::           
1.12      anton     297: * Exception Handling::          
                    298: 
                    299: Defining Words
                    300: 
1.67      anton     301: * CREATE::                      
1.44      crook     302: * Variables::                   Variables and user variables
1.67      anton     303: * Constants::                   
1.44      crook     304: * Values::                      Initialised variables
1.67      anton     305: * Colon Definitions::           
1.44      crook     306: * Anonymous Definitions::       Definitions without names
1.71      anton     307: * Supplying names::             Passing definition names as strings
1.67      anton     308: * User-defined Defining Words::  
1.44      crook     309: * Deferred words::              Allow forward references
1.67      anton     310: * Aliases::                     
1.47      crook     311: 
1.63      anton     312: User-defined Defining Words
                    313: 
                    314: * CREATE..DOES> applications::  
                    315: * CREATE..DOES> details::       
                    316: * Advanced does> usage example::  
1.91    ! anton     317: * @code{Const-does>}::          
1.63      anton     318: 
1.47      crook     319: Interpretation and Compilation Semantics
                    320: 
1.67      anton     321: * Combined words::              
1.12      anton     322: 
1.71      anton     323: Tokens for Words
                    324: 
                    325: * Execution token::             represents execution/interpretation semantics
                    326: * Compilation token::           represents compilation semantics
                    327: * Name token::                  represents named words
                    328: 
1.82      anton     329: Compiling words
                    330: 
                    331: * Literals::                    Compiling data values
                    332: * Macros::                      Compiling words
                    333: 
1.21      crook     334: The Text Interpreter
                    335: 
1.67      anton     336: * Input Sources::               
                    337: * Number Conversion::           
                    338: * Interpret/Compile states::    
                    339: * Interpreter Directives::      
1.21      crook     340: 
1.26      crook     341: Word Lists
                    342: 
1.75      anton     343: * Vocabularies::                
1.67      anton     344: * Why use word lists?::         
1.75      anton     345: * Word list example::           
1.26      crook     346: 
                    347: Files
                    348: 
1.48      anton     349: * Forth source files::          
                    350: * General files::               
                    351: * Search Paths::                
                    352: 
                    353: Search Paths
                    354: 
1.75      anton     355: * Source Search Paths::         
1.26      crook     356: * General Search Paths::        
                    357: 
                    358: Other I/O
                    359: 
1.32      anton     360: * Simple numeric output::       Predefined formats
                    361: * Formatted numeric output::    Formatted (pictured) output
                    362: * String Formats::              How Forth stores strings in memory
1.67      anton     363: * Displaying characters and strings::  Other stuff
1.32      anton     364: * Input::                       Input
1.26      crook     365: 
                    366: Locals
                    367: 
                    368: * Gforth locals::               
                    369: * ANS Forth locals::            
                    370: 
                    371: Gforth locals
                    372: 
                    373: * Where are locals visible by name?::  
                    374: * How long do locals live?::    
1.78      anton     375: * Locals programming style::    
                    376: * Locals implementation::       
1.26      crook     377: 
1.12      anton     378: Structures
                    379: 
                    380: * Why explicit structure support?::  
                    381: * Structure Usage::             
                    382: * Structure Naming Convention::  
                    383: * Structure Implementation::    
                    384: * Structure Glossary::          
                    385: 
                    386: Object-oriented Forth
                    387: 
1.48      anton     388: * Why object-oriented programming?::  
                    389: * Object-Oriented Terminology::  
                    390: * Objects::                     
                    391: * OOF::                         
                    392: * Mini-OOF::                    
1.23      crook     393: * Comparison with other object models::  
1.12      anton     394: 
1.24      anton     395: The @file{objects.fs} model
1.12      anton     396: 
                    397: * Properties of the Objects model::  
                    398: * Basic Objects Usage::         
1.41      anton     399: * The Objects base class::      
1.12      anton     400: * Creating objects::            
                    401: * Object-Oriented Programming Style::  
                    402: * Class Binding::               
                    403: * Method conveniences::         
                    404: * Classes and Scoping::         
1.41      anton     405: * Dividing classes::            
1.12      anton     406: * Object Interfaces::           
                    407: * Objects Implementation::      
                    408: * Objects Glossary::            
                    409: 
1.24      anton     410: The @file{oof.fs} model
1.12      anton     411: 
1.67      anton     412: * Properties of the OOF model::  
                    413: * Basic OOF Usage::             
                    414: * The OOF base class::          
                    415: * Class Declaration::           
                    416: * Class Implementation::        
1.12      anton     417: 
1.24      anton     418: The @file{mini-oof.fs} model
1.23      crook     419: 
1.48      anton     420: * Basic Mini-OOF Usage::        
                    421: * Mini-OOF Example::            
                    422: * Mini-OOF Implementation::     
1.23      crook     423: 
1.78      anton     424: Programming Tools
                    425: 
                    426: * Examining::                   
                    427: * Forgetting words::            
                    428: * Debugging::                   Simple and quick.
                    429: * Assertions::                  Making your programs self-checking.
                    430: * Singlestep Debugger::         Executing your program word by word.
                    431: 
                    432: Assembler and Code Words
                    433: 
                    434: * Code and ;code::              
                    435: * Common Assembler::            Assembler Syntax
                    436: * Common Disassembler::         
                    437: * 386 Assembler::               Deviations and special cases
                    438: * Alpha Assembler::             Deviations and special cases
                    439: * MIPS assembler::              Deviations and special cases
                    440: * Other assemblers::            How to write them
                    441: 
1.12      anton     442: Tools
                    443: 
                    444: * ANS Report::                  Report the words used, sorted by wordset.
                    445: 
                    446: ANS conformance
                    447: 
                    448: * The Core Words::              
                    449: * The optional Block word set::  
                    450: * The optional Double Number word set::  
                    451: * The optional Exception word set::  
                    452: * The optional Facility word set::  
                    453: * The optional File-Access word set::  
                    454: * The optional Floating-Point word set::  
                    455: * The optional Locals word set::  
                    456: * The optional Memory-Allocation word set::  
                    457: * The optional Programming-Tools word set::  
                    458: * The optional Search-Order word set::  
                    459: 
                    460: The Core Words
                    461: 
                    462: * core-idef::                   Implementation Defined Options                   
                    463: * core-ambcond::                Ambiguous Conditions                
                    464: * core-other::                  Other System Documentation                  
                    465: 
                    466: The optional Block word set
                    467: 
                    468: * block-idef::                  Implementation Defined Options
                    469: * block-ambcond::               Ambiguous Conditions               
                    470: * block-other::                 Other System Documentation                 
                    471: 
                    472: The optional Double Number word set
                    473: 
                    474: * double-ambcond::              Ambiguous Conditions              
                    475: 
                    476: The optional Exception word set
                    477: 
                    478: * exception-idef::              Implementation Defined Options              
                    479: 
                    480: The optional Facility word set
                    481: 
                    482: * facility-idef::               Implementation Defined Options               
                    483: * facility-ambcond::            Ambiguous Conditions            
                    484: 
                    485: The optional File-Access word set
                    486: 
                    487: * file-idef::                   Implementation Defined Options
                    488: * file-ambcond::                Ambiguous Conditions                
                    489: 
                    490: The optional Floating-Point word set
                    491: 
                    492: * floating-idef::               Implementation Defined Options
                    493: * floating-ambcond::            Ambiguous Conditions            
                    494: 
                    495: The optional Locals word set
                    496: 
                    497: * locals-idef::                 Implementation Defined Options                 
                    498: * locals-ambcond::              Ambiguous Conditions              
                    499: 
                    500: The optional Memory-Allocation word set
                    501: 
                    502: * memory-idef::                 Implementation Defined Options                 
                    503: 
                    504: The optional Programming-Tools word set
                    505: 
                    506: * programming-idef::            Implementation Defined Options            
                    507: * programming-ambcond::         Ambiguous Conditions         
                    508: 
                    509: The optional Search-Order word set
                    510: 
                    511: * search-idef::                 Implementation Defined Options                 
                    512: * search-ambcond::              Ambiguous Conditions              
                    513: 
                    514: Image Files
                    515: 
1.24      anton     516: * Image Licensing Issues::      Distribution terms for images.
                    517: * Image File Background::       Why have image files?
1.67      anton     518: * Non-Relocatable Image Files::  don't always work.
1.24      anton     519: * Data-Relocatable Image Files::  are better.
1.67      anton     520: * Fully Relocatable Image Files::  better yet.
1.24      anton     521: * Stack and Dictionary Sizes::  Setting the default sizes for an image.
1.32      anton     522: * Running Image Files::         @code{gforth -i @i{file}} or @i{file}.
1.24      anton     523: * Modifying the Startup Sequence::  and turnkey applications.
1.12      anton     524: 
                    525: Fully Relocatable Image Files
                    526: 
1.27      crook     527: * gforthmi::                    The normal way
1.12      anton     528: * cross.fs::                    The hard way
                    529: 
                    530: Engine
                    531: 
                    532: * Portability::                 
                    533: * Threading::                   
                    534: * Primitives::                  
                    535: * Performance::                 
                    536: 
                    537: Threading
                    538: 
                    539: * Scheduling::                  
                    540: * Direct or Indirect Threaded?::  
                    541: * DOES>::                       
                    542: 
                    543: Primitives
                    544: 
                    545: * Automatic Generation::        
                    546: * TOS Optimization::            
                    547: * Produced code::               
1.13      pazsan    548: 
                    549: Cross Compiler
                    550: 
1.67      anton     551: * Using the Cross Compiler::    
                    552: * How the Cross Compiler Works::  
1.13      pazsan    553: 
1.24      anton     554: @end detailmenu
1.1       anton     555: @end menu
                    556: 
1.26      crook     557: @node License, Goals, Top, Top
1.1       anton     558: @unnumbered GNU GENERAL PUBLIC LICENSE
                    559: @center Version 2, June 1991
                    560: 
                    561: @display
                    562: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
1.88      anton     563: 59 Temple Place, Suite 330, Boston, MA 02111, USA
1.1       anton     564: 
                    565: Everyone is permitted to copy and distribute verbatim copies
                    566: of this license document, but changing it is not allowed.
                    567: @end display
                    568: 
                    569: @unnumberedsec Preamble
                    570: 
                    571:   The licenses for most software are designed to take away your
                    572: freedom to share and change it.  By contrast, the GNU General Public
                    573: License is intended to guarantee your freedom to share and change free
                    574: software---to make sure the software is free for all its users.  This
                    575: General Public License applies to most of the Free Software
                    576: Foundation's software and to any other program whose authors commit to
                    577: using it.  (Some other Free Software Foundation software is covered by
                    578: the GNU Library General Public License instead.)  You can apply it to
                    579: your programs, too.
                    580: 
                    581:   When we speak of free software, we are referring to freedom, not
                    582: price.  Our General Public Licenses are designed to make sure that you
                    583: have the freedom to distribute copies of free software (and charge for
                    584: this service if you wish), that you receive source code or can get it
                    585: if you want it, that you can change the software or use pieces of it
                    586: in new free programs; and that you know you can do these things.
                    587: 
                    588:   To protect your rights, we need to make restrictions that forbid
                    589: anyone to deny you these rights or to ask you to surrender the rights.
                    590: These restrictions translate to certain responsibilities for you if you
                    591: distribute copies of the software, or if you modify it.
                    592: 
                    593:   For example, if you distribute copies of such a program, whether
                    594: gratis or for a fee, you must give the recipients all the rights that
                    595: you have.  You must make sure that they, too, receive or can get the
                    596: source code.  And you must show them these terms so they know their
                    597: rights.
                    598: 
                    599:   We protect your rights with two steps: (1) copyright the software, and
                    600: (2) offer you this license which gives you legal permission to copy,
                    601: distribute and/or modify the software.
                    602: 
                    603:   Also, for each author's protection and ours, we want to make certain
                    604: that everyone understands that there is no warranty for this free
                    605: software.  If the software is modified by someone else and passed on, we
                    606: want its recipients to know that what they have is not the original, so
                    607: that any problems introduced by others will not reflect on the original
                    608: authors' reputations.
                    609: 
                    610:   Finally, any free program is threatened constantly by software
                    611: patents.  We wish to avoid the danger that redistributors of a free
                    612: program will individually obtain patent licenses, in effect making the
                    613: program proprietary.  To prevent this, we have made it clear that any
                    614: patent must be licensed for everyone's free use or not licensed at all.
                    615: 
                    616:   The precise terms and conditions for copying, distribution and
                    617: modification follow.
                    618: 
                    619: @iftex
                    620: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
                    621: @end iftex
1.49      anton     622: @ifnottex
1.1       anton     623: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
1.49      anton     624: @end ifnottex
1.1       anton     625: 
                    626: @enumerate 0
                    627: @item
                    628: This License applies to any program or other work which contains
                    629: a notice placed by the copyright holder saying it may be distributed
                    630: under the terms of this General Public License.  The ``Program'', below,
                    631: refers to any such program or work, and a ``work based on the Program''
                    632: means either the Program or any derivative work under copyright law:
                    633: that is to say, a work containing the Program or a portion of it,
                    634: either verbatim or with modifications and/or translated into another
                    635: language.  (Hereinafter, translation is included without limitation in
                    636: the term ``modification''.)  Each licensee is addressed as ``you''.
                    637: 
                    638: Activities other than copying, distribution and modification are not
                    639: covered by this License; they are outside its scope.  The act of
                    640: running the Program is not restricted, and the output from the Program
                    641: is covered only if its contents constitute a work based on the
                    642: Program (independent of having been made by running the Program).
                    643: Whether that is true depends on what the Program does.
                    644: 
                    645: @item
                    646: You may copy and distribute verbatim copies of the Program's
                    647: source code as you receive it, in any medium, provided that you
                    648: conspicuously and appropriately publish on each copy an appropriate
                    649: copyright notice and disclaimer of warranty; keep intact all the
                    650: notices that refer to this License and to the absence of any warranty;
                    651: and give any other recipients of the Program a copy of this License
                    652: along with the Program.
                    653: 
                    654: You may charge a fee for the physical act of transferring a copy, and
                    655: you may at your option offer warranty protection in exchange for a fee.
                    656: 
                    657: @item
                    658: You may modify your copy or copies of the Program or any portion
                    659: of it, thus forming a work based on the Program, and copy and
                    660: distribute such modifications or work under the terms of Section 1
                    661: above, provided that you also meet all of these conditions:
                    662: 
                    663: @enumerate a
                    664: @item
                    665: You must cause the modified files to carry prominent notices
                    666: stating that you changed the files and the date of any change.
                    667: 
                    668: @item
                    669: You must cause any work that you distribute or publish, that in
                    670: whole or in part contains or is derived from the Program or any
                    671: part thereof, to be licensed as a whole at no charge to all third
                    672: parties under the terms of this License.
                    673: 
                    674: @item
                    675: If the modified program normally reads commands interactively
                    676: when run, you must cause it, when started running for such
                    677: interactive use in the most ordinary way, to print or display an
                    678: announcement including an appropriate copyright notice and a
                    679: notice that there is no warranty (or else, saying that you provide
                    680: a warranty) and that users may redistribute the program under
                    681: these conditions, and telling the user how to view a copy of this
                    682: License.  (Exception: if the Program itself is interactive but
                    683: does not normally print such an announcement, your work based on
                    684: the Program is not required to print an announcement.)
                    685: @end enumerate
                    686: 
                    687: These requirements apply to the modified work as a whole.  If
                    688: identifiable sections of that work are not derived from the Program,
                    689: and can be reasonably considered independent and separate works in
                    690: themselves, then this License, and its terms, do not apply to those
                    691: sections when you distribute them as separate works.  But when you
                    692: distribute the same sections as part of a whole which is a work based
                    693: on the Program, the distribution of the whole must be on the terms of
                    694: this License, whose permissions for other licensees extend to the
                    695: entire whole, and thus to each and every part regardless of who wrote it.
                    696: 
                    697: Thus, it is not the intent of this section to claim rights or contest
                    698: your rights to work written entirely by you; rather, the intent is to
                    699: exercise the right to control the distribution of derivative or
                    700: collective works based on the Program.
                    701: 
                    702: In addition, mere aggregation of another work not based on the Program
                    703: with the Program (or with a work based on the Program) on a volume of
                    704: a storage or distribution medium does not bring the other work under
                    705: the scope of this License.
                    706: 
                    707: @item
                    708: You may copy and distribute the Program (or a work based on it,
                    709: under Section 2) in object code or executable form under the terms of
                    710: Sections 1 and 2 above provided that you also do one of the following:
                    711: 
                    712: @enumerate a
                    713: @item
                    714: Accompany it with the complete corresponding machine-readable
                    715: source code, which must be distributed under the terms of Sections
                    716: 1 and 2 above on a medium customarily used for software interchange; or,
                    717: 
                    718: @item
                    719: Accompany it with a written offer, valid for at least three
                    720: years, to give any third party, for a charge no more than your
                    721: cost of physically performing source distribution, a complete
                    722: machine-readable copy of the corresponding source code, to be
                    723: distributed under the terms of Sections 1 and 2 above on a medium
                    724: customarily used for software interchange; or,
                    725: 
                    726: @item
                    727: Accompany it with the information you received as to the offer
                    728: to distribute corresponding source code.  (This alternative is
                    729: allowed only for noncommercial distribution and only if you
                    730: received the program in object code or executable form with such
                    731: an offer, in accord with Subsection b above.)
                    732: @end enumerate
                    733: 
                    734: The source code for a work means the preferred form of the work for
                    735: making modifications to it.  For an executable work, complete source
                    736: code means all the source code for all modules it contains, plus any
                    737: associated interface definition files, plus the scripts used to
                    738: control compilation and installation of the executable.  However, as a
                    739: special exception, the source code distributed need not include
                    740: anything that is normally distributed (in either source or binary
                    741: form) with the major components (compiler, kernel, and so on) of the
                    742: operating system on which the executable runs, unless that component
                    743: itself accompanies the executable.
                    744: 
                    745: If distribution of executable or object code is made by offering
                    746: access to copy from a designated place, then offering equivalent
                    747: access to copy the source code from the same place counts as
                    748: distribution of the source code, even though third parties are not
                    749: compelled to copy the source along with the object code.
                    750: 
                    751: @item
                    752: You may not copy, modify, sublicense, or distribute the Program
                    753: except as expressly provided under this License.  Any attempt
                    754: otherwise to copy, modify, sublicense or distribute the Program is
                    755: void, and will automatically terminate your rights under this License.
                    756: However, parties who have received copies, or rights, from you under
                    757: this License will not have their licenses terminated so long as such
                    758: parties remain in full compliance.
                    759: 
                    760: @item
                    761: You are not required to accept this License, since you have not
                    762: signed it.  However, nothing else grants you permission to modify or
                    763: distribute the Program or its derivative works.  These actions are
                    764: prohibited by law if you do not accept this License.  Therefore, by
                    765: modifying or distributing the Program (or any work based on the
                    766: Program), you indicate your acceptance of this License to do so, and
                    767: all its terms and conditions for copying, distributing or modifying
                    768: the Program or works based on it.
                    769: 
                    770: @item
                    771: Each time you redistribute the Program (or any work based on the
                    772: Program), the recipient automatically receives a license from the
                    773: original licensor to copy, distribute or modify the Program subject to
                    774: these terms and conditions.  You may not impose any further
                    775: restrictions on the recipients' exercise of the rights granted herein.
                    776: You are not responsible for enforcing compliance by third parties to
                    777: this License.
                    778: 
                    779: @item
                    780: If, as a consequence of a court judgment or allegation of patent
                    781: infringement or for any other reason (not limited to patent issues),
                    782: conditions are imposed on you (whether by court order, agreement or
                    783: otherwise) that contradict the conditions of this License, they do not
                    784: excuse you from the conditions of this License.  If you cannot
                    785: distribute so as to satisfy simultaneously your obligations under this
                    786: License and any other pertinent obligations, then as a consequence you
                    787: may not distribute the Program at all.  For example, if a patent
                    788: license would not permit royalty-free redistribution of the Program by
                    789: all those who receive copies directly or indirectly through you, then
                    790: the only way you could satisfy both it and this License would be to
                    791: refrain entirely from distribution of the Program.
                    792: 
                    793: If any portion of this section is held invalid or unenforceable under
                    794: any particular circumstance, the balance of the section is intended to
                    795: apply and the section as a whole is intended to apply in other
                    796: circumstances.
                    797: 
                    798: It is not the purpose of this section to induce you to infringe any
                    799: patents or other property right claims or to contest validity of any
                    800: such claims; this section has the sole purpose of protecting the
                    801: integrity of the free software distribution system, which is
                    802: implemented by public license practices.  Many people have made
                    803: generous contributions to the wide range of software distributed
                    804: through that system in reliance on consistent application of that
                    805: system; it is up to the author/donor to decide if he or she is willing
                    806: to distribute software through any other system and a licensee cannot
                    807: impose that choice.
                    808: 
                    809: This section is intended to make thoroughly clear what is believed to
                    810: be a consequence of the rest of this License.
                    811: 
                    812: @item
                    813: If the distribution and/or use of the Program is restricted in
                    814: certain countries either by patents or by copyrighted interfaces, the
                    815: original copyright holder who places the Program under this License
                    816: may add an explicit geographical distribution limitation excluding
                    817: those countries, so that distribution is permitted only in or among
                    818: countries not thus excluded.  In such case, this License incorporates
                    819: the limitation as if written in the body of this License.
                    820: 
                    821: @item
                    822: The Free Software Foundation may publish revised and/or new versions
                    823: of the General Public License from time to time.  Such new versions will
                    824: be similar in spirit to the present version, but may differ in detail to
                    825: address new problems or concerns.
                    826: 
                    827: Each version is given a distinguishing version number.  If the Program
                    828: specifies a version number of this License which applies to it and ``any
                    829: later version'', you have the option of following the terms and conditions
                    830: either of that version or of any later version published by the Free
                    831: Software Foundation.  If the Program does not specify a version number of
                    832: this License, you may choose any version ever published by the Free Software
                    833: Foundation.
                    834: 
                    835: @item
                    836: If you wish to incorporate parts of the Program into other free
                    837: programs whose distribution conditions are different, write to the author
                    838: to ask for permission.  For software which is copyrighted by the Free
                    839: Software Foundation, write to the Free Software Foundation; we sometimes
                    840: make exceptions for this.  Our decision will be guided by the two goals
                    841: of preserving the free status of all derivatives of our free software and
                    842: of promoting the sharing and reuse of software generally.
                    843: 
                    844: @iftex
                    845: @heading NO WARRANTY
                    846: @end iftex
1.49      anton     847: @ifnottex
1.1       anton     848: @center NO WARRANTY
1.49      anton     849: @end ifnottex
1.1       anton     850: 
                    851: @item
                    852: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
                    853: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW.  EXCEPT WHEN
                    854: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
                    855: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
                    856: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
                    857: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.  THE ENTIRE RISK AS
                    858: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU.  SHOULD THE
                    859: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
                    860: REPAIR OR CORRECTION.
                    861: 
                    862: @item
                    863: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
                    864: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
                    865: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
                    866: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
                    867: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
                    868: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
                    869: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
                    870: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
                    871: POSSIBILITY OF SUCH DAMAGES.
                    872: @end enumerate
                    873: 
                    874: @iftex
                    875: @heading END OF TERMS AND CONDITIONS
                    876: @end iftex
1.49      anton     877: @ifnottex
1.1       anton     878: @center END OF TERMS AND CONDITIONS
1.49      anton     879: @end ifnottex
1.1       anton     880: 
                    881: @page
                    882: @unnumberedsec How to Apply These Terms to Your New Programs
                    883: 
                    884:   If you develop a new program, and you want it to be of the greatest
                    885: possible use to the public, the best way to achieve this is to make it
                    886: free software which everyone can redistribute and change under these terms.
                    887: 
                    888:   To do so, attach the following notices to the program.  It is safest
                    889: to attach them to the start of each source file to most effectively
                    890: convey the exclusion of warranty; and each file should have at least
                    891: the ``copyright'' line and a pointer to where the full notice is found.
                    892: 
                    893: @smallexample
                    894: @var{one line to give the program's name and a brief idea of what it does.}
                    895: Copyright (C) 19@var{yy}  @var{name of author}
                    896: 
                    897: This program is free software; you can redistribute it and/or modify 
                    898: it under the terms of the GNU General Public License as published by 
                    899: the Free Software Foundation; either version 2 of the License, or 
                    900: (at your option) any later version.
                    901: 
                    902: This program is distributed in the hope that it will be useful,
                    903: but WITHOUT ANY WARRANTY; without even the implied warranty of
                    904: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
                    905: GNU General Public License for more details.
                    906: 
                    907: You should have received a copy of the GNU General Public License
                    908: along with this program; if not, write to the Free Software
1.88      anton     909: Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
1.1       anton     910: @end smallexample
                    911: 
                    912: Also add information on how to contact you by electronic and paper mail.
                    913: 
                    914: If the program is interactive, make it output a short notice like this
                    915: when it starts in an interactive mode:
                    916: 
                    917: @smallexample
                    918: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
                    919: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
                    920: type `show w'.  
                    921: This is free software, and you are welcome to redistribute it 
                    922: under certain conditions; type `show c' for details.
                    923: @end smallexample
                    924: 
                    925: The hypothetical commands @samp{show w} and @samp{show c} should show
                    926: the appropriate parts of the General Public License.  Of course, the
                    927: commands you use may be called something other than @samp{show w} and
                    928: @samp{show c}; they could even be mouse-clicks or menu items---whatever
                    929: suits your program.
                    930: 
                    931: You should also get your employer (if you work as a programmer) or your
                    932: school, if any, to sign a ``copyright disclaimer'' for the program, if
                    933: necessary.  Here is a sample; alter the names:
                    934: 
                    935: @smallexample
                    936: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
                    937: `Gnomovision' (which makes passes at compilers) written by James Hacker.
                    938: 
                    939: @var{signature of Ty Coon}, 1 April 1989
                    940: Ty Coon, President of Vice
                    941: @end smallexample
                    942: 
                    943: This General Public License does not permit incorporating your program into
                    944: proprietary programs.  If your program is a subroutine library, you may
                    945: consider it more useful to permit linking proprietary applications with the
                    946: library.  If this is what you want to do, use the GNU Library General
                    947: Public License instead of this License.
                    948: 
                    949: @iftex
                    950: @unnumbered Preface
                    951: @cindex Preface
1.21      crook     952: This manual documents Gforth. Some introductory material is provided for
                    953: readers who are unfamiliar with Forth or who are migrating to Gforth
                    954: from other Forth compilers. However, this manual is primarily a
                    955: reference manual.
1.1       anton     956: @end iftex
                    957: 
1.28      crook     958: @comment TODO much more blurb here.
1.26      crook     959: 
                    960: @c ******************************************************************
1.29      crook     961: @node Goals, Gforth Environment, License, Top
1.26      crook     962: @comment node-name,     next,           previous, up
                    963: @chapter Goals of Gforth
                    964: @cindex goals of the Gforth project
                    965: The goal of the Gforth Project is to develop a standard model for
                    966: ANS Forth. This can be split into several subgoals:
                    967: 
                    968: @itemize @bullet
                    969: @item
                    970: Gforth should conform to the ANS Forth Standard.
                    971: @item
                    972: It should be a model, i.e. it should define all the
                    973: implementation-dependent things.
                    974: @item
                    975: It should become standard, i.e. widely accepted and used. This goal
                    976: is the most difficult one.
                    977: @end itemize
                    978: 
                    979: To achieve these goals Gforth should be
                    980: @itemize @bullet
                    981: @item
                    982: Similar to previous models (fig-Forth, F83)
                    983: @item
                    984: Powerful. It should provide for all the things that are considered
                    985: necessary today and even some that are not yet considered necessary.
                    986: @item
                    987: Efficient. It should not get the reputation of being exceptionally
                    988: slow.
                    989: @item
                    990: Free.
                    991: @item
                    992: Available on many machines/easy to port.
                    993: @end itemize
                    994: 
                    995: Have we achieved these goals? Gforth conforms to the ANS Forth
                    996: standard. It may be considered a model, but we have not yet documented
                    997: which parts of the model are stable and which parts we are likely to
                    998: change. It certainly has not yet become a de facto standard, but it
                    999: appears to be quite popular. It has some similarities to and some
                   1000: differences from previous models. It has some powerful features, but not
                   1001: yet everything that we envisioned. We certainly have achieved our
1.65      anton    1002: execution speed goals (@pxref{Performance})@footnote{However, in 1998
                   1003: the bar was raised when the major commercial Forth vendors switched to
                   1004: native code compilers.}.  It is free and available on many machines.
1.29      crook    1005: 
1.26      crook    1006: @c ******************************************************************
1.48      anton    1007: @node Gforth Environment, Tutorial, Goals, Top
1.29      crook    1008: @chapter Gforth Environment
                   1009: @cindex Gforth environment
1.21      crook    1010: 
1.45      crook    1011: Note: ultimately, the Gforth man page will be auto-generated from the
1.29      crook    1012: material in this chapter.
1.21      crook    1013: 
                   1014: @menu
1.29      crook    1015: * Invoking Gforth::             Getting in
                   1016: * Leaving Gforth::              Getting out
                   1017: * Command-line editing::        
1.48      anton    1018: * Environment variables::       that affect how Gforth starts up
1.29      crook    1019: * Gforth Files::                What gets installed and where
1.48      anton    1020: * Startup speed::               When 35ms is not fast enough ...
1.21      crook    1021: @end menu
                   1022: 
1.49      anton    1023: For related information about the creation of images see @ref{Image Files}.
1.29      crook    1024: 
1.21      crook    1025: @comment ----------------------------------------------
1.48      anton    1026: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1.29      crook    1027: @section Invoking Gforth
                   1028: @cindex invoking Gforth
                   1029: @cindex running Gforth
                   1030: @cindex command-line options
                   1031: @cindex options on the command line
                   1032: @cindex flags on the command line
1.21      crook    1033: 
1.30      anton    1034: Gforth is made up of two parts; an executable ``engine'' (named
                   1035: @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
                   1036: will usually just say @code{gforth} -- this automatically loads the
                   1037: default image file @file{gforth.fi}. In many other cases the default
                   1038: Gforth image will be invoked like this:
1.21      crook    1039: @example
1.30      anton    1040: gforth [file | -e forth-code] ...
1.21      crook    1041: @end example
1.29      crook    1042: @noindent
                   1043: This interprets the contents of the files and the Forth code in the order they
                   1044: are given.
1.21      crook    1045: 
1.30      anton    1046: In addition to the @file{gforth} engine, there is also an engine called
                   1047: @file{gforth-fast}, which is faster, but gives less informative error
1.89      anton    1048: messages (@pxref{Error messages}).  You should use it for debugged,
                   1049: performance-critical programs.
1.30      anton    1050: 
1.29      crook    1051: In general, the command line looks like this:
1.21      crook    1052: 
                   1053: @example
1.30      anton    1054: gforth[-fast] [engine options] [image options]
1.21      crook    1055: @end example
                   1056: 
1.30      anton    1057: The engine options must come before the rest of the command
1.29      crook    1058: line. They are:
1.26      crook    1059: 
1.29      crook    1060: @table @code
                   1061: @cindex -i, command-line option
                   1062: @cindex --image-file, command-line option
                   1063: @item --image-file @i{file}
                   1064: @itemx -i @i{file}
                   1065: Loads the Forth image @i{file} instead of the default
                   1066: @file{gforth.fi} (@pxref{Image Files}).
1.21      crook    1067: 
1.39      anton    1068: @cindex --appl-image, command-line option
                   1069: @item --appl-image @i{file}
                   1070: Loads the image @i{file} and leaves all further command-line arguments
1.65      anton    1071: to the image (instead of processing them as engine options).  This is
                   1072: useful for building executable application images on Unix, built with
1.39      anton    1073: @code{gforthmi --application ...}.
                   1074: 
1.29      crook    1075: @cindex --path, command-line option
                   1076: @cindex -p, command-line option
                   1077: @item --path @i{path}
                   1078: @itemx -p @i{path}
                   1079: Uses @i{path} for searching the image file and Forth source code files
                   1080: instead of the default in the environment variable @code{GFORTHPATH} or
                   1081: the path specified at installation time (e.g.,
                   1082: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
                   1083: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1.21      crook    1084: 
1.29      crook    1085: @cindex --dictionary-size, command-line option
                   1086: @cindex -m, command-line option
                   1087: @cindex @i{size} parameters for command-line options
                   1088: @cindex size of the dictionary and the stacks
                   1089: @item --dictionary-size @i{size}
                   1090: @itemx -m @i{size}
                   1091: Allocate @i{size} space for the Forth dictionary space instead of
                   1092: using the default specified in the image (typically 256K). The
                   1093: @i{size} specification for this and subsequent options consists of
                   1094: an integer and a unit (e.g.,
                   1095: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
                   1096: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
                   1097: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
                   1098: @code{e} is used.
1.21      crook    1099: 
1.29      crook    1100: @cindex --data-stack-size, command-line option
                   1101: @cindex -d, command-line option
                   1102: @item --data-stack-size @i{size}
                   1103: @itemx -d @i{size}
                   1104: Allocate @i{size} space for the data stack instead of using the
                   1105: default specified in the image (typically 16K).
1.21      crook    1106: 
1.29      crook    1107: @cindex --return-stack-size, command-line option
                   1108: @cindex -r, command-line option
                   1109: @item --return-stack-size @i{size}
                   1110: @itemx -r @i{size}
                   1111: Allocate @i{size} space for the return stack instead of using the
                   1112: default specified in the image (typically 15K).
1.21      crook    1113: 
1.29      crook    1114: @cindex --fp-stack-size, command-line option
                   1115: @cindex -f, command-line option
                   1116: @item --fp-stack-size @i{size}
                   1117: @itemx -f @i{size}
                   1118: Allocate @i{size} space for the floating point stack instead of
                   1119: using the default specified in the image (typically 15.5K). In this case
                   1120: the unit specifier @code{e} refers to floating point numbers.
1.21      crook    1121: 
1.48      anton    1122: @cindex --locals-stack-size, command-line option
                   1123: @cindex -l, command-line option
                   1124: @item --locals-stack-size @i{size}
                   1125: @itemx -l @i{size}
                   1126: Allocate @i{size} space for the locals stack instead of using the
                   1127: default specified in the image (typically 14.5K).
                   1128: 
                   1129: @cindex -h, command-line option
                   1130: @cindex --help, command-line option
                   1131: @item --help
                   1132: @itemx -h
                   1133: Print a message about the command-line options
                   1134: 
                   1135: @cindex -v, command-line option
                   1136: @cindex --version, command-line option
                   1137: @item --version
                   1138: @itemx -v
                   1139: Print version and exit
                   1140: 
                   1141: @cindex --debug, command-line option
                   1142: @item --debug
                   1143: Print some information useful for debugging on startup.
                   1144: 
                   1145: @cindex --offset-image, command-line option
                   1146: @item --offset-image
                   1147: Start the dictionary at a slightly different position than would be used
                   1148: otherwise (useful for creating data-relocatable images,
                   1149: @pxref{Data-Relocatable Image Files}).
                   1150: 
                   1151: @cindex --no-offset-im, command-line option
                   1152: @item --no-offset-im
                   1153: Start the dictionary at the normal position.
                   1154: 
                   1155: @cindex --clear-dictionary, command-line option
                   1156: @item --clear-dictionary
                   1157: Initialize all bytes in the dictionary to 0 before loading the image
                   1158: (@pxref{Data-Relocatable Image Files}).
                   1159: 
                   1160: @cindex --die-on-signal, command-line-option
                   1161: @item --die-on-signal
                   1162: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
                   1163: or the segmentation violation SIGSEGV) by translating it into a Forth
                   1164: @code{THROW}. With this option, Gforth exits if it receives such a
                   1165: signal. This option is useful when the engine and/or the image might be
                   1166: severely broken (such that it causes another signal before recovering
                   1167: from the first); this option avoids endless loops in such cases.
                   1168: @end table
                   1169: 
                   1170: @cindex loading files at startup
                   1171: @cindex executing code on startup
                   1172: @cindex batch processing with Gforth
                   1173: As explained above, the image-specific command-line arguments for the
                   1174: default image @file{gforth.fi} consist of a sequence of filenames and
                   1175: @code{-e @var{forth-code}} options that are interpreted in the sequence
                   1176: in which they are given. The @code{-e @var{forth-code}} or
                   1177: @code{--evaluate @var{forth-code}} option evaluates the Forth
                   1178: code. This option takes only one argument; if you want to evaluate more
                   1179: Forth words, you have to quote them or use @code{-e} several times. To exit
                   1180: after processing the command line (instead of entering interactive mode)
                   1181: append @code{-e bye} to the command line.
                   1182: 
                   1183: @cindex versions, invoking other versions of Gforth
                   1184: If you have several versions of Gforth installed, @code{gforth} will
                   1185: invoke the version that was installed last. @code{gforth-@i{version}}
                   1186: invokes a specific version. If your environment contains the variable
                   1187: @code{GFORTHPATH}, you may want to override it by using the
                   1188: @code{--path} option.
                   1189: 
                   1190: Not yet implemented:
                   1191: On startup the system first executes the system initialization file
                   1192: (unless the option @code{--no-init-file} is given; note that the system
                   1193: resulting from using this option may not be ANS Forth conformant). Then
                   1194: the user initialization file @file{.gforth.fs} is executed, unless the
1.62      crook    1195: option @code{--no-rc} is given; this file is searched for in @file{.},
1.48      anton    1196: then in @file{~}, then in the normal path (see above).
                   1197: 
                   1198: 
                   1199: 
                   1200: @comment ----------------------------------------------
                   1201: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
                   1202: @section Leaving Gforth
                   1203: @cindex Gforth - leaving
                   1204: @cindex leaving Gforth
                   1205: 
                   1206: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
                   1207: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
                   1208: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1.49      anton    1209: data are discarded.  For ways of saving the state of the system before
                   1210: leaving Gforth see @ref{Image Files}.
1.48      anton    1211: 
                   1212: doc-bye
                   1213: 
                   1214: 
                   1215: @comment ----------------------------------------------
1.65      anton    1216: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1.48      anton    1217: @section Command-line editing
                   1218: @cindex command-line editing
                   1219: 
                   1220: Gforth maintains a history file that records every line that you type to
                   1221: the text interpreter. This file is preserved between sessions, and is
                   1222: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
                   1223: repeatedly you can recall successively older commands from this (or
                   1224: previous) session(s). The full list of command-line editing facilities is:
                   1225: 
                   1226: @itemize @bullet
                   1227: @item
                   1228: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
                   1229: commands from the history buffer.
                   1230: @item
                   1231: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
                   1232: from the history buffer.
                   1233: @item
                   1234: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
                   1235: @item
                   1236: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
                   1237: @item
                   1238: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
                   1239: closing up the line.
                   1240: @item
                   1241: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
                   1242: @item
                   1243: @kbd{Ctrl-a} to move the cursor to the start of the line.
                   1244: @item
                   1245: @kbd{Ctrl-e} to move the cursor to the end of the line.
                   1246: @item
                   1247: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
                   1248: line.
                   1249: @item
                   1250: @key{TAB} to step through all possible full-word completions of the word
                   1251: currently being typed.
                   1252: @item
1.65      anton    1253: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
                   1254: using @code{bye}). 
                   1255: @item
                   1256: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
                   1257: character under the cursor.
1.48      anton    1258: @end itemize
                   1259: 
                   1260: When editing, displayable characters are inserted to the left of the
                   1261: cursor position; the line is always in ``insert'' (as opposed to
                   1262: ``overstrike'') mode.
                   1263: 
                   1264: @cindex history file
                   1265: @cindex @file{.gforth-history}
                   1266: On Unix systems, the history file is @file{~/.gforth-history} by
                   1267: default@footnote{i.e. it is stored in the user's home directory.}. You
                   1268: can find out the name and location of your history file using:
                   1269: 
                   1270: @example 
                   1271: history-file type \ Unix-class systems
                   1272: 
                   1273: history-file type \ Other systems
                   1274: history-dir  type
                   1275: @end example
                   1276: 
                   1277: If you enter long definitions by hand, you can use a text editor to
                   1278: paste them out of the history file into a Forth source file for reuse at
                   1279: a later time.
                   1280: 
                   1281: Gforth never trims the size of the history file, so you should do this
                   1282: periodically, if necessary.
                   1283: 
                   1284: @comment this is all defined in history.fs
                   1285: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
                   1286: @comment chosen?
                   1287: 
                   1288: 
                   1289: @comment ----------------------------------------------
1.65      anton    1290: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1.48      anton    1291: @section Environment variables
                   1292: @cindex environment variables
                   1293: 
                   1294: Gforth uses these environment variables:
                   1295: 
                   1296: @itemize @bullet
                   1297: @item
                   1298: @cindex @code{GFORTHHIST} -- environment variable
                   1299: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
                   1300: open/create the history file, @file{.gforth-history}. Default:
                   1301: @code{$HOME}.
                   1302: 
                   1303: @item
                   1304: @cindex @code{GFORTHPATH} -- environment variable
                   1305: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
                   1306: for Forth source-code files.
                   1307: 
                   1308: @item
                   1309: @cindex @code{GFORTH} -- environment variable
1.49      anton    1310: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48      anton    1311: 
                   1312: @item
                   1313: @cindex @code{GFORTHD} -- environment variable
1.62      crook    1314: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48      anton    1315: 
                   1316: @item
                   1317: @cindex @code{TMP}, @code{TEMP} - environment variable
                   1318: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
                   1319: location for the history file.
                   1320: @end itemize
                   1321: 
                   1322: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
                   1323: @comment mentioning these.
                   1324: 
                   1325: All the Gforth environment variables default to sensible values if they
                   1326: are not set.
                   1327: 
                   1328: 
                   1329: @comment ----------------------------------------------
                   1330: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
                   1331: @section Gforth files
                   1332: @cindex Gforth files
                   1333: 
                   1334: When you install Gforth on a Unix system, it installs files in these
                   1335: locations by default:
                   1336: 
                   1337: @itemize @bullet
                   1338: @item
                   1339: @file{/usr/local/bin/gforth}
                   1340: @item
                   1341: @file{/usr/local/bin/gforthmi}
                   1342: @item
                   1343: @file{/usr/local/man/man1/gforth.1} - man page.
                   1344: @item
                   1345: @file{/usr/local/info} - the Info version of this manual.
                   1346: @item
                   1347: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
                   1348: @item
                   1349: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
                   1350: @item
                   1351: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
                   1352: @item
                   1353: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
                   1354: @end itemize
                   1355: 
                   1356: You can select different places for installation by using
                   1357: @code{configure} options (listed with @code{configure --help}).
                   1358: 
                   1359: @comment ----------------------------------------------
                   1360: @node Startup speed,  , Gforth Files, Gforth Environment
                   1361: @section Startup speed
                   1362: @cindex Startup speed
                   1363: @cindex speed, startup
                   1364: 
                   1365: If Gforth is used for CGI scripts or in shell scripts, its startup
                   1366: speed may become a problem.  On a 300MHz 21064a under Linux-2.2.13 with
                   1367: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
                   1368: system time.
                   1369: 
                   1370: If startup speed is a problem, you may consider the following ways to
                   1371: improve it; or you may consider ways to reduce the number of startups
1.62      crook    1372: (for example, by using Fast-CGI).
1.48      anton    1373: 
                   1374: The first step to improve startup speed is to statically link Gforth, by
                   1375: building it with @code{XLDFLAGS=-static}.  This requires more memory for
                   1376: the code and will therefore slow down the first invocation, but
                   1377: subsequent invocations avoid the dynamic linking overhead.  Another
                   1378: disadvantage is that Gforth won't profit from library upgrades.  As a
                   1379: result, @code{gforth-static -e bye} takes about 17.1ms user and
                   1380: 8.2ms system time.
                   1381: 
                   1382: The next step to improve startup speed is to use a non-relocatable image
1.65      anton    1383: (@pxref{Non-Relocatable Image Files}).  You can create this image with
1.48      anton    1384: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
                   1385: @code{gforth -i gforthnr.fi ...}.  This avoids the relocation overhead
                   1386: and a part of the copy-on-write overhead.  The disadvantage is that the
1.62      crook    1387: non-relocatable image does not work if the OS gives Gforth a different
1.48      anton    1388: address for the dictionary, for whatever reason; so you better provide a
                   1389: fallback on a relocatable image.  @code{gforth-static -i gforthnr.fi -e
                   1390: bye} takes about 15.3ms user and 7.5ms system time.
                   1391: 
                   1392: The final step is to disable dictionary hashing in Gforth.  Gforth
                   1393: builds the hash table on startup, which takes much of the startup
                   1394: overhead. You can do this by commenting out the @code{include hash.fs}
                   1395: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
                   1396: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
                   1397: The disadvantages are that functionality like @code{table} and
                   1398: @code{ekey} is missing and that text interpretation (e.g., compiling)
                   1399: now takes much longer. So, you should only use this method if there is
                   1400: no significant text interpretation to perform (the script should be
1.62      crook    1401: compiled into the image, amongst other things).  @code{gforth-static -i
1.48      anton    1402: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
                   1403: 
                   1404: @c ******************************************************************
                   1405: @node Tutorial, Introduction, Gforth Environment, Top
                   1406: @chapter Forth Tutorial
                   1407: @cindex Tutorial
                   1408: @cindex Forth Tutorial
                   1409: 
1.67      anton    1410: @c Topics from nac's Introduction that could be mentioned:
                   1411: @c press <ret> after each line
                   1412: @c Prompt
                   1413: @c numbers vs. words in dictionary on text interpretation
                   1414: @c what happens on redefinition
                   1415: @c parsing words (in particular, defining words)
                   1416: 
1.83      anton    1417: The difference of this chapter from the Introduction
                   1418: (@pxref{Introduction}) is that this tutorial is more fast-paced, should
                   1419: be used while sitting in front of a computer, and covers much more
                   1420: material, but does not explain how the Forth system works.
                   1421: 
1.62      crook    1422: This tutorial can be used with any ANS-compliant Forth; any
                   1423: Gforth-specific features are marked as such and you can skip them if you
                   1424: work with another Forth.  This tutorial does not explain all features of
                   1425: Forth, just enough to get you started and give you some ideas about the
                   1426: facilities available in Forth.  Read the rest of the manual and the
                   1427: standard when you are through this.
1.48      anton    1428: 
                   1429: The intended way to use this tutorial is that you work through it while
                   1430: sitting in front of the console, take a look at the examples and predict
                   1431: what they will do, then try them out; if the outcome is not as expected,
                   1432: find out why (e.g., by trying out variations of the example), so you
                   1433: understand what's going on.  There are also some assignments that you
                   1434: should solve.
                   1435: 
                   1436: This tutorial assumes that you have programmed before and know what,
                   1437: e.g., a loop is.
                   1438: 
                   1439: @c !! explain compat library
                   1440: 
                   1441: @menu
                   1442: * Starting Gforth Tutorial::    
                   1443: * Syntax Tutorial::             
                   1444: * Crash Course Tutorial::       
                   1445: * Stack Tutorial::              
                   1446: * Arithmetics Tutorial::        
                   1447: * Stack Manipulation Tutorial::  
                   1448: * Using files for Forth code Tutorial::  
                   1449: * Comments Tutorial::           
                   1450: * Colon Definitions Tutorial::  
                   1451: * Decompilation Tutorial::      
                   1452: * Stack-Effect Comments Tutorial::  
                   1453: * Types Tutorial::              
                   1454: * Factoring Tutorial::          
                   1455: * Designing the stack effect Tutorial::  
                   1456: * Local Variables Tutorial::    
                   1457: * Conditional execution Tutorial::  
                   1458: * Flags and Comparisons Tutorial::  
                   1459: * General Loops Tutorial::      
                   1460: * Counted loops Tutorial::      
                   1461: * Recursion Tutorial::          
                   1462: * Leaving definitions or loops Tutorial::  
                   1463: * Return Stack Tutorial::       
                   1464: * Memory Tutorial::             
                   1465: * Characters and Strings Tutorial::  
                   1466: * Alignment Tutorial::          
1.87      anton    1467: * Files Tutorial::              
1.48      anton    1468: * Interpretation and Compilation Semantics and Immediacy Tutorial::  
                   1469: * Execution Tokens Tutorial::   
                   1470: * Exceptions Tutorial::         
                   1471: * Defining Words Tutorial::     
                   1472: * Arrays and Records Tutorial::  
                   1473: * POSTPONE Tutorial::           
                   1474: * Literal Tutorial::            
                   1475: * Advanced macros Tutorial::    
                   1476: * Compilation Tokens Tutorial::  
                   1477: * Wordlists and Search Order Tutorial::  
                   1478: @end menu
                   1479: 
                   1480: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
                   1481: @section Starting Gforth
1.66      anton    1482: @cindex starting Gforth tutorial
1.48      anton    1483: You can start Gforth by typing its name:
                   1484: 
                   1485: @example
                   1486: gforth
                   1487: @end example
                   1488: 
                   1489: That puts you into interactive mode; you can leave Gforth by typing
                   1490: @code{bye}.  While in Gforth, you can edit the command line and access
                   1491: the command line history with cursor keys, similar to bash.
                   1492: 
                   1493: 
                   1494: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
                   1495: @section Syntax
1.66      anton    1496: @cindex syntax tutorial
1.48      anton    1497: 
                   1498: A @dfn{word} is a sequence of arbitrary characters (expcept white
                   1499: space).  Words are separated by white space.  E.g., each of the
                   1500: following lines contains exactly one word:
                   1501: 
                   1502: @example
                   1503: word
                   1504: !@@#$%^&*()
                   1505: 1234567890
                   1506: 5!a
                   1507: @end example
                   1508: 
                   1509: A frequent beginner's error is to leave away necessary white space,
                   1510: resulting in an error like @samp{Undefined word}; so if you see such an
                   1511: error, check if you have put spaces wherever necessary.
                   1512: 
                   1513: @example
                   1514: ." hello, world" \ correct
                   1515: ."hello, world"  \ gives an "Undefined word" error
                   1516: @end example
                   1517: 
1.65      anton    1518: Gforth and most other Forth systems ignore differences in case (they are
1.48      anton    1519: case-insensitive), i.e., @samp{word} is the same as @samp{Word}.  If
                   1520: your system is case-sensitive, you may have to type all the examples
                   1521: given here in upper case.
                   1522: 
                   1523: 
                   1524: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
                   1525: @section Crash Course
                   1526: 
                   1527: Type
                   1528: 
                   1529: @example
                   1530: 0 0 !
                   1531: here execute
                   1532: ' catch >body 20 erase abort
                   1533: ' (quit) >body 20 erase
                   1534: @end example
                   1535: 
                   1536: The last two examples are guaranteed to destroy parts of Gforth (and
                   1537: most other systems), so you better leave Gforth afterwards (if it has
                   1538: not finished by itself).  On some systems you may have to kill gforth
                   1539: from outside (e.g., in Unix with @code{kill}).
                   1540: 
                   1541: Now that you know how to produce crashes (and that there's not much to
                   1542: them), let's learn how to produce meaningful programs.
                   1543: 
                   1544: 
                   1545: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
                   1546: @section Stack
1.66      anton    1547: @cindex stack tutorial
1.48      anton    1548: 
                   1549: The most obvious feature of Forth is the stack.  When you type in a
                   1550: number, it is pushed on the stack.  You can display the content of the
                   1551: stack with @code{.s}.
                   1552: 
                   1553: @example
                   1554: 1 2 .s
                   1555: 3 .s
                   1556: @end example
                   1557: 
                   1558: @code{.s} displays the top-of-stack to the right, i.e., the numbers
                   1559: appear in @code{.s} output as they appeared in the input.
                   1560: 
                   1561: You can print the top of stack element with @code{.}.
                   1562: 
                   1563: @example
                   1564: 1 2 3 . . .
                   1565: @end example
                   1566: 
                   1567: In general, words consume their stack arguments (@code{.s} is an
                   1568: exception).
                   1569: 
                   1570: @assignment
                   1571: What does the stack contain after @code{5 6 7 .}?
                   1572: @endassignment
                   1573: 
                   1574: 
                   1575: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
                   1576: @section Arithmetics
1.66      anton    1577: @cindex arithmetics tutorial
1.48      anton    1578: 
                   1579: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
                   1580: operate on the top two stack items:
                   1581: 
                   1582: @example
1.67      anton    1583: 2 2 .s
                   1584: + .s
                   1585: .
1.48      anton    1586: 2 1 - .
                   1587: 7 3 mod .
                   1588: @end example
                   1589: 
                   1590: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
                   1591: as in the corresponding infix expression (this is generally the case in
                   1592: Forth).
                   1593: 
                   1594: Parentheses are superfluous (and not available), because the order of
                   1595: the words unambiguously determines the order of evaluation and the
                   1596: operands:
                   1597: 
                   1598: @example
                   1599: 3 4 + 5 * .
                   1600: 3 4 5 * + .
                   1601: @end example
                   1602: 
                   1603: @assignment
                   1604: What are the infix expressions corresponding to the Forth code above?
                   1605: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
                   1606: known as Postfix or RPN (Reverse Polish Notation).}.
                   1607: @endassignment
                   1608: 
                   1609: To change the sign, use @code{negate}:
                   1610: 
                   1611: @example
                   1612: 2 negate .
                   1613: @end example
                   1614: 
                   1615: @assignment
                   1616: Convert -(-3)*4-5 to Forth.
                   1617: @endassignment
                   1618: 
                   1619: @code{/mod} performs both @code{/} and @code{mod}.
                   1620: 
                   1621: @example
                   1622: 7 3 /mod . .
                   1623: @end example
                   1624: 
1.66      anton    1625: Reference: @ref{Arithmetic}.
                   1626: 
                   1627: 
1.48      anton    1628: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
                   1629: @section Stack Manipulation
1.66      anton    1630: @cindex stack manipulation tutorial
1.48      anton    1631: 
                   1632: Stack manipulation words rearrange the data on the stack.
                   1633: 
                   1634: @example
                   1635: 1 .s drop .s
                   1636: 1 .s dup .s drop drop .s
                   1637: 1 2 .s over .s drop drop drop
                   1638: 1 2 .s swap .s drop drop
                   1639: 1 2 3 .s rot .s drop drop drop
                   1640: @end example
                   1641: 
                   1642: These are the most important stack manipulation words.  There are also
                   1643: variants that manipulate twice as many stack items:
                   1644: 
                   1645: @example
                   1646: 1 2 3 4 .s 2swap .s 2drop 2drop
                   1647: @end example
                   1648: 
                   1649: Two more stack manipulation words are:
                   1650: 
                   1651: @example
                   1652: 1 2 .s nip .s drop
                   1653: 1 2 .s tuck .s 2drop drop
                   1654: @end example
                   1655: 
                   1656: @assignment
                   1657: Replace @code{nip} and @code{tuck} with combinations of other stack
                   1658: manipulation words.
                   1659: 
                   1660: @example
                   1661: Given:          How do you get:
                   1662: 1 2 3           3 2 1           
                   1663: 1 2 3           1 2 3 2                 
                   1664: 1 2 3           1 2 3 3                 
                   1665: 1 2 3           1 3 3           
                   1666: 1 2 3           2 1 3           
                   1667: 1 2 3 4         4 3 2 1         
                   1668: 1 2 3           1 2 3 1 2 3             
                   1669: 1 2 3 4         1 2 3 4 1 2             
                   1670: 1 2 3
                   1671: 1 2 3           1 2 3 4                 
                   1672: 1 2 3           1 3             
                   1673: @end example
                   1674: @endassignment
                   1675: 
                   1676: @example
                   1677: 5 dup * .
                   1678: @end example
                   1679: 
                   1680: @assignment
                   1681: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
                   1682: Write a piece of Forth code that expects two numbers on the stack
                   1683: (@var{a} and @var{b}, with @var{b} on top) and computes
                   1684: @code{(a-b)(a+1)}.
                   1685: @endassignment
                   1686: 
1.66      anton    1687: Reference: @ref{Stack Manipulation}.
                   1688: 
                   1689: 
1.48      anton    1690: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
                   1691: @section Using files for Forth code
1.66      anton    1692: @cindex loading Forth code, tutorial
                   1693: @cindex files containing Forth code, tutorial
1.48      anton    1694: 
                   1695: While working at the Forth command line is convenient for one-line
                   1696: examples and short one-off code, you probably want to store your source
                   1697: code in files for convenient editing and persistence.  You can use your
                   1698: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
                   1699: Gforth}) to create @var{file} and use
                   1700: 
                   1701: @example
                   1702: s" @var{file}" included
                   1703: @end example
                   1704: 
                   1705: to load it into your Forth system.  The file name extension I use for
                   1706: Forth files is @samp{.fs}.
                   1707: 
                   1708: You can easily start Gforth with some files loaded like this:
                   1709: 
                   1710: @example
                   1711: gforth @var{file1} @var{file2}
                   1712: @end example
                   1713: 
                   1714: If an error occurs during loading these files, Gforth terminates,
                   1715: whereas an error during @code{INCLUDED} within Gforth usually gives you
                   1716: a Gforth command line.  Starting the Forth system every time gives you a
                   1717: clean start every time, without interference from the results of earlier
                   1718: tries.
                   1719: 
                   1720: I often put all the tests in a file, then load the code and run the
                   1721: tests with
                   1722: 
                   1723: @example
                   1724: gforth @var{code} @var{tests} -e bye
                   1725: @end example
                   1726: 
                   1727: (often by performing this command with @kbd{C-x C-e} in Emacs).  The
                   1728: @code{-e bye} ensures that Gforth terminates afterwards so that I can
                   1729: restart this command without ado.
                   1730: 
                   1731: The advantage of this approach is that the tests can be repeated easily
                   1732: every time the program ist changed, making it easy to catch bugs
                   1733: introduced by the change.
                   1734: 
1.66      anton    1735: Reference: @ref{Forth source files}.
                   1736: 
1.48      anton    1737: 
                   1738: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
                   1739: @section Comments
1.66      anton    1740: @cindex comments tutorial
1.48      anton    1741: 
                   1742: @example
                   1743: \ That's a comment; it ends at the end of the line
                   1744: ( Another comment; it ends here: )  .s
                   1745: @end example
                   1746: 
                   1747: @code{\} and @code{(} are ordinary Forth words and therefore have to be
                   1748: separated with white space from the following text.
                   1749: 
                   1750: @example
                   1751: \This gives an "Undefined word" error
                   1752: @end example
                   1753: 
                   1754: The first @code{)} ends a comment started with @code{(}, so you cannot
                   1755: nest @code{(}-comments; and you cannot comment out text containing a
                   1756: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
                   1757: avoid @code{)} in word names.}.
                   1758: 
                   1759: I use @code{\}-comments for descriptive text and for commenting out code
                   1760: of one or more line; I use @code{(}-comments for describing the stack
                   1761: effect, the stack contents, or for commenting out sub-line pieces of
                   1762: code.
                   1763: 
                   1764: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
                   1765: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
                   1766: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
                   1767: with @kbd{M-q}.
                   1768: 
1.66      anton    1769: Reference: @ref{Comments}.
                   1770: 
1.48      anton    1771: 
                   1772: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
                   1773: @section Colon Definitions
1.66      anton    1774: @cindex colon definitions, tutorial
                   1775: @cindex definitions, tutorial
                   1776: @cindex procedures, tutorial
                   1777: @cindex functions, tutorial
1.48      anton    1778: 
                   1779: are similar to procedures and functions in other programming languages.
                   1780: 
                   1781: @example
                   1782: : squared ( n -- n^2 )
                   1783:    dup * ;
                   1784: 5 squared .
                   1785: 7 squared .
                   1786: @end example
                   1787: 
                   1788: @code{:} starts the colon definition; its name is @code{squared}.  The
                   1789: following comment describes its stack effect.  The words @code{dup *}
                   1790: are not executed, but compiled into the definition.  @code{;} ends the
                   1791: colon definition.
                   1792: 
                   1793: The newly-defined word can be used like any other word, including using
                   1794: it in other definitions:
                   1795: 
                   1796: @example
                   1797: : cubed ( n -- n^3 )
                   1798:    dup squared * ;
                   1799: -5 cubed .
                   1800: : fourth-power ( n -- n^4 )
                   1801:    squared squared ;
                   1802: 3 fourth-power .
                   1803: @end example
                   1804: 
                   1805: @assignment
                   1806: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
                   1807: @code{/mod} in terms of other Forth words, and check if they work (hint:
                   1808: test your tests on the originals first).  Don't let the
                   1809: @samp{redefined}-Messages spook you, they are just warnings.
                   1810: @endassignment
                   1811: 
1.66      anton    1812: Reference: @ref{Colon Definitions}.
                   1813: 
1.48      anton    1814: 
                   1815: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
                   1816: @section Decompilation
1.66      anton    1817: @cindex decompilation tutorial
                   1818: @cindex see tutorial
1.48      anton    1819: 
                   1820: You can decompile colon definitions with @code{see}:
                   1821: 
                   1822: @example
                   1823: see squared
                   1824: see cubed
                   1825: @end example
                   1826: 
                   1827: In Gforth @code{see} shows you a reconstruction of the source code from
                   1828: the executable code.  Informations that were present in the source, but
                   1829: not in the executable code, are lost (e.g., comments).
                   1830: 
1.65      anton    1831: You can also decompile the predefined words:
                   1832: 
                   1833: @example
                   1834: see .
                   1835: see +
                   1836: @end example
                   1837: 
                   1838: 
1.48      anton    1839: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
                   1840: @section Stack-Effect Comments
1.66      anton    1841: @cindex stack-effect comments, tutorial
                   1842: @cindex --, tutorial
1.48      anton    1843: By convention the comment after the name of a definition describes the
                   1844: stack effect: The part in from of the @samp{--} describes the state of
                   1845: the stack before the execution of the definition, i.e., the parameters
                   1846: that are passed into the colon definition; the part behind the @samp{--}
                   1847: is the state of the stack after the execution of the definition, i.e.,
                   1848: the results of the definition.  The stack comment only shows the top
                   1849: stack items that the definition accesses and/or changes.
                   1850: 
                   1851: You should put a correct stack effect on every definition, even if it is
                   1852: just @code{( -- )}.  You should also add some descriptive comment to
                   1853: more complicated words (I usually do this in the lines following
                   1854: @code{:}).  If you don't do this, your code becomes unreadable (because
                   1855: you have to work through every definition before you can undertsand
                   1856: any).
                   1857: 
                   1858: @assignment
                   1859: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
                   1860: x2 x1}.  Describe the stack effect of @code{-}, @code{drop}, @code{dup},
                   1861: @code{over}, @code{rot}, @code{nip}, and @code{tuck}.  Hint: When you
1.65      anton    1862: are done, you can compare your stack effects to those in this manual
1.48      anton    1863: (@pxref{Word Index}).
                   1864: @endassignment
                   1865: 
                   1866: Sometimes programmers put comments at various places in colon
                   1867: definitions that describe the contents of the stack at that place (stack
                   1868: comments); i.e., they are like the first part of a stack-effect
                   1869: comment. E.g.,
                   1870: 
                   1871: @example
                   1872: : cubed ( n -- n^3 )
                   1873:    dup squared  ( n n^2 ) * ;
                   1874: @end example
                   1875: 
                   1876: In this case the stack comment is pretty superfluous, because the word
                   1877: is simple enough.  If you think it would be a good idea to add such a
                   1878: comment to increase readability, you should also consider factoring the
                   1879: word into several simpler words (@pxref{Factoring Tutorial,,
1.60      anton    1880: Factoring}), which typically eliminates the need for the stack comment;
1.48      anton    1881: however, if you decide not to refactor it, then having such a comment is
                   1882: better than not having it.
                   1883: 
                   1884: The names of the stack items in stack-effect and stack comments in the
                   1885: standard, in this manual, and in many programs specify the type through
                   1886: a type prefix, similar to Fortran and Hungarian notation.  The most
                   1887: frequent prefixes are:
                   1888: 
                   1889: @table @code
                   1890: @item n
                   1891: signed integer
                   1892: @item u
                   1893: unsigned integer
                   1894: @item c
                   1895: character
                   1896: @item f
                   1897: Boolean flags, i.e. @code{false} or @code{true}.
                   1898: @item a-addr,a-
                   1899: Cell-aligned address
                   1900: @item c-addr,c-
                   1901: Char-aligned address (note that a Char may have two bytes in Windows NT)
                   1902: @item xt
                   1903: Execution token, same size as Cell
                   1904: @item w,x
                   1905: Cell, can contain an integer or an address.  It usually takes 32, 64 or
                   1906: 16 bits (depending on your platform and Forth system). A cell is more
                   1907: commonly known as machine word, but the term @emph{word} already means
                   1908: something different in Forth.
                   1909: @item d
                   1910: signed double-cell integer
                   1911: @item ud
                   1912: unsigned double-cell integer
                   1913: @item r
                   1914: Float (on the FP stack)
                   1915: @end table
                   1916: 
                   1917: You can find a more complete list in @ref{Notation}.
                   1918: 
                   1919: @assignment
                   1920: Write stack-effect comments for all definitions you have written up to
                   1921: now.
                   1922: @endassignment
                   1923: 
                   1924: 
                   1925: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
                   1926: @section Types
1.66      anton    1927: @cindex types tutorial
1.48      anton    1928: 
                   1929: In Forth the names of the operations are not overloaded; so similar
                   1930: operations on different types need different names; e.g., @code{+} adds
                   1931: integers, and you have to use @code{f+} to add floating-point numbers.
                   1932: The following prefixes are often used for related operations on
                   1933: different types:
                   1934: 
                   1935: @table @code
                   1936: @item (none)
                   1937: signed integer
                   1938: @item u
                   1939: unsigned integer
                   1940: @item c
                   1941: character
                   1942: @item d
                   1943: signed double-cell integer
                   1944: @item ud, du
                   1945: unsigned double-cell integer
                   1946: @item 2
                   1947: two cells (not-necessarily double-cell numbers)
                   1948: @item m, um
                   1949: mixed single-cell and double-cell operations
                   1950: @item f
                   1951: floating-point (note that in stack comments @samp{f} represents flags,
1.66      anton    1952: and @samp{r} represents FP numbers).
1.48      anton    1953: @end table
                   1954: 
                   1955: If there are no differences between the signed and the unsigned variant
                   1956: (e.g., for @code{+}), there is only the prefix-less variant.
                   1957: 
                   1958: Forth does not perform type checking, neither at compile time, nor at
                   1959: run time.  If you use the wrong oeration, the data are interpreted
                   1960: incorrectly:
                   1961: 
                   1962: @example
                   1963: -1 u.
                   1964: @end example
                   1965: 
                   1966: If you have only experience with type-checked languages until now, and
                   1967: have heard how important type-checking is, don't panic!  In my
                   1968: experience (and that of other Forthers), type errors in Forth code are
                   1969: usually easy to find (once you get used to it), the increased vigilance
                   1970: of the programmer tends to catch some harder errors in addition to most
                   1971: type errors, and you never have to work around the type system, so in
                   1972: most situations the lack of type-checking seems to be a win (projects to
                   1973: add type checking to Forth have not caught on).
                   1974: 
                   1975: 
                   1976: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
                   1977: @section Factoring
1.66      anton    1978: @cindex factoring tutorial
1.48      anton    1979: 
                   1980: If you try to write longer definitions, you will soon find it hard to
                   1981: keep track of the stack contents.  Therefore, good Forth programmers
                   1982: tend to write only short definitions (e.g., three lines).  The art of
                   1983: finding meaningful short definitions is known as factoring (as in
                   1984: factoring polynomials).
                   1985: 
                   1986: Well-factored programs offer additional advantages: smaller, more
                   1987: general words, are easier to test and debug and can be reused more and
                   1988: better than larger, specialized words.
                   1989: 
                   1990: So, if you run into difficulties with stack management, when writing
                   1991: code, try to define meaningful factors for the word, and define the word
                   1992: in terms of those.  Even if a factor contains only two words, it is
                   1993: often helpful.
                   1994: 
1.65      anton    1995: Good factoring is not easy, and it takes some practice to get the knack
                   1996: for it; but even experienced Forth programmers often don't find the
                   1997: right solution right away, but only when rewriting the program.  So, if
                   1998: you don't come up with a good solution immediately, keep trying, don't
                   1999: despair.
1.48      anton    2000: 
                   2001: @c example !!
                   2002: 
                   2003: 
                   2004: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
                   2005: @section Designing the stack effect
1.66      anton    2006: @cindex Stack effect design, tutorial
                   2007: @cindex design of stack effects, tutorial
1.48      anton    2008: 
                   2009: In other languages you can use an arbitrary order of parameters for a
1.65      anton    2010: function; and since there is only one result, you don't have to deal with
1.48      anton    2011: the order of results, either.
                   2012: 
                   2013: In Forth (and other stack-based languages, e.g., Postscript) the
                   2014: parameter and result order of a definition is important and should be
                   2015: designed well.  The general guideline is to design the stack effect such
                   2016: that the word is simple to use in most cases, even if that complicates
                   2017: the implementation of the word.  Some concrete rules are:
                   2018: 
                   2019: @itemize @bullet
                   2020: 
                   2021: @item
                   2022: Words consume all of their parameters (e.g., @code{.}).
                   2023: 
                   2024: @item
                   2025: If there is a convention on the order of parameters (e.g., from
                   2026: mathematics or another programming language), stick with it (e.g.,
                   2027: @code{-}).
                   2028: 
                   2029: @item
                   2030: If one parameter usually requires only a short computation (e.g., it is
                   2031: a constant), pass it on the top of the stack.  Conversely, parameters
                   2032: that usually require a long sequence of code to compute should be passed
                   2033: as the bottom (i.e., first) parameter.  This makes the code easier to
                   2034: read, because reader does not need to keep track of the bottom item
                   2035: through a long sequence of code (or, alternatively, through stack
1.49      anton    2036: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1.48      anton    2037: address on top of the stack because it is usually simpler to compute
                   2038: than the stored value (often the address is just a variable).
                   2039: 
                   2040: @item
                   2041: Similarly, results that are usually consumed quickly should be returned
                   2042: on the top of stack, whereas a result that is often used in long
                   2043: computations should be passed as bottom result.  E.g., the file words
                   2044: like @code{open-file} return the error code on the top of stack, because
                   2045: it is usually consumed quickly by @code{throw}; moreover, the error code
                   2046: has to be checked before doing anything with the other results.
                   2047: 
                   2048: @end itemize
                   2049: 
                   2050: These rules are just general guidelines, don't lose sight of the overall
                   2051: goal to make the words easy to use.  E.g., if the convention rule
                   2052: conflicts with the computation-length rule, you might decide in favour
                   2053: of the convention if the word will be used rarely, and in favour of the
                   2054: computation-length rule if the word will be used frequently (because
                   2055: with frequent use the cost of breaking the computation-length rule would
                   2056: be quite high, and frequent use makes it easier to remember an
                   2057: unconventional order).
                   2058: 
                   2059: @c example !! structure package
                   2060: 
1.65      anton    2061: 
1.48      anton    2062: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
                   2063: @section Local Variables
1.66      anton    2064: @cindex local variables, tutorial
1.48      anton    2065: 
                   2066: You can define local variables (@emph{locals}) in a colon definition:
                   2067: 
                   2068: @example
                   2069: : swap @{ a b -- b a @}
                   2070:   b a ;
                   2071: 1 2 swap .s 2drop
                   2072: @end example
                   2073: 
                   2074: (If your Forth system does not support this syntax, include
                   2075: @file{compat/anslocals.fs} first).
                   2076: 
                   2077: In this example @code{@{ a b -- b a @}} is the locals definition; it
                   2078: takes two cells from the stack, puts the top of stack in @code{b} and
                   2079: the next stack element in @code{a}.  @code{--} starts a comment ending
                   2080: with @code{@}}.  After the locals definition, using the name of the
                   2081: local will push its value on the stack.  You can leave the comment
                   2082: part (@code{-- b a}) away:
                   2083: 
                   2084: @example
                   2085: : swap ( x1 x2 -- x2 x1 )
                   2086:   @{ a b @} b a ;
                   2087: @end example
                   2088: 
                   2089: In Gforth you can have several locals definitions, anywhere in a colon
                   2090: definition; in contrast, in a standard program you can have only one
                   2091: locals definition per colon definition, and that locals definition must
                   2092: be outside any controll structure.
                   2093: 
                   2094: With locals you can write slightly longer definitions without running
                   2095: into stack trouble.  However, I recommend trying to write colon
                   2096: definitions without locals for exercise purposes to help you gain the
                   2097: essential factoring skills.
                   2098: 
                   2099: @assignment
                   2100: Rewrite your definitions until now with locals
                   2101: @endassignment
                   2102: 
1.66      anton    2103: Reference: @ref{Locals}.
                   2104: 
1.48      anton    2105: 
                   2106: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
                   2107: @section Conditional execution
1.66      anton    2108: @cindex conditionals, tutorial
                   2109: @cindex if, tutorial
1.48      anton    2110: 
                   2111: In Forth you can use control structures only inside colon definitions.
                   2112: An @code{if}-structure looks like this:
                   2113: 
                   2114: @example
                   2115: : abs ( n1 -- +n2 )
                   2116:     dup 0 < if
                   2117:         negate
                   2118:     endif ;
                   2119: 5 abs .
                   2120: -5 abs .
                   2121: @end example
                   2122: 
                   2123: @code{if} takes a flag from the stack.  If the flag is non-zero (true),
                   2124: the following code is performed, otherwise execution continues after the
1.51      pazsan   2125: @code{endif} (or @code{else}).  @code{<} compares the top two stack
1.48      anton    2126: elements and prioduces a flag:
                   2127: 
                   2128: @example
                   2129: 1 2 < .
                   2130: 2 1 < .
                   2131: 1 1 < .
                   2132: @end example
                   2133: 
                   2134: Actually the standard name for @code{endif} is @code{then}.  This
                   2135: tutorial presents the examples using @code{endif}, because this is often
                   2136: less confusing for people familiar with other programming languages
                   2137: where @code{then} has a different meaning.  If your system does not have
                   2138: @code{endif}, define it with
                   2139: 
                   2140: @example
                   2141: : endif postpone then ; immediate
                   2142: @end example
                   2143: 
                   2144: You can optionally use an @code{else}-part:
                   2145: 
                   2146: @example
                   2147: : min ( n1 n2 -- n )
                   2148:   2dup < if
                   2149:     drop
                   2150:   else
                   2151:     nip
                   2152:   endif ;
                   2153: 2 3 min .
                   2154: 3 2 min .
                   2155: @end example
                   2156: 
                   2157: @assignment
                   2158: Write @code{min} without @code{else}-part (hint: what's the definition
                   2159: of @code{nip}?).
                   2160: @endassignment
                   2161: 
1.66      anton    2162: Reference: @ref{Selection}.
                   2163: 
1.48      anton    2164: 
                   2165: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
                   2166: @section Flags and Comparisons
1.66      anton    2167: @cindex flags tutorial
                   2168: @cindex comparison tutorial
1.48      anton    2169: 
                   2170: In a false-flag all bits are clear (0 when interpreted as integer).  In
                   2171: a canonical true-flag all bits are set (-1 as a twos-complement signed
                   2172: integer); in many contexts (e.g., @code{if}) any non-zero value is
                   2173: treated as true flag.
                   2174: 
                   2175: @example
                   2176: false .
                   2177: true .
                   2178: true hex u. decimal
                   2179: @end example
                   2180: 
                   2181: Comparison words produce canonical flags:
                   2182: 
                   2183: @example
                   2184: 1 1 = .
                   2185: 1 0= .
                   2186: 0 1 < .
                   2187: 0 0 < .
                   2188: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
                   2189: -1 1 < .
                   2190: @end example
                   2191: 
1.66      anton    2192: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
                   2193: (or none) and the comparisons @code{= <> < > <= >=}.  Only a part of
                   2194: these combinations are standard (for details see the standard,
                   2195: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
1.48      anton    2196: 
                   2197: You can use @code{and or xor invert} can be used as operations on
                   2198: canonical flags.  Actually they are bitwise operations:
                   2199: 
                   2200: @example
                   2201: 1 2 and .
                   2202: 1 2 or .
                   2203: 1 3 xor .
                   2204: 1 invert .
                   2205: @end example
                   2206: 
                   2207: You can convert a zero/non-zero flag into a canonical flag with
                   2208: @code{0<>} (and complement it on the way with @code{0=}).
                   2209: 
                   2210: @example
                   2211: 1 0= .
                   2212: 1 0<> .
                   2213: @end example
                   2214: 
1.65      anton    2215: You can use the all-bits-set feature of canonical flags and the bitwise
1.48      anton    2216: operation of the Boolean operations to avoid @code{if}s:
                   2217: 
                   2218: @example
                   2219: : foo ( n1 -- n2 )
                   2220:   0= if
                   2221:     14
                   2222:   else
                   2223:     0
                   2224:   endif ;
                   2225: 0 foo .
                   2226: 1 foo .
                   2227: 
                   2228: : foo ( n1 -- n2 )
                   2229:   0= 14 and ;
                   2230: 0 foo .
                   2231: 1 foo .
                   2232: @end example
                   2233: 
                   2234: @assignment
                   2235: Write @code{min} without @code{if}.
                   2236: @endassignment
                   2237: 
1.66      anton    2238: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
                   2239: @ref{Bitwise operations}.
                   2240: 
1.48      anton    2241: 
                   2242: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
                   2243: @section General Loops
1.66      anton    2244: @cindex loops, indefinite, tutorial
1.48      anton    2245: 
                   2246: The endless loop is the most simple one:
                   2247: 
                   2248: @example
                   2249: : endless ( -- )
                   2250:   0 begin
                   2251:     dup . 1+
                   2252:   again ;
                   2253: endless
                   2254: @end example
                   2255: 
                   2256: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth).  @code{begin}
                   2257: does nothing at run-time, @code{again} jumps back to @code{begin}.
                   2258: 
                   2259: A loop with one exit at any place looks like this:
                   2260: 
                   2261: @example
                   2262: : log2 ( +n1 -- n2 )
                   2263: \ logarithmus dualis of n1>0, rounded down to the next integer
                   2264:   assert( dup 0> )
                   2265:   2/ 0 begin
                   2266:     over 0> while
                   2267:       1+ swap 2/ swap
                   2268:   repeat
                   2269:   nip ;
                   2270: 7 log2 .
                   2271: 8 log2 .
                   2272: @end example
                   2273: 
                   2274: At run-time @code{while} consumes a flag; if it is 0, execution
1.51      pazsan   2275: continues behind the @code{repeat}; if the flag is non-zero, execution
1.48      anton    2276: continues behind the @code{while}.  @code{Repeat} jumps back to
                   2277: @code{begin}, just like @code{again}.
                   2278: 
                   2279: In Forth there are many combinations/abbreviations, like @code{1+}.
1.90      anton    2280: However, @code{2/} is not one of them; it shifts its argument right by
1.48      anton    2281: one bit (arithmetic shift right):
                   2282: 
                   2283: @example
                   2284: -5 2 / .
                   2285: -5 2/ .
                   2286: @end example
                   2287: 
                   2288: @code{assert(} is no standard word, but you can get it on systems other
                   2289: then Gforth by including @file{compat/assert.fs}.  You can see what it
                   2290: does by trying
                   2291: 
                   2292: @example
                   2293: 0 log2 .
                   2294: @end example
                   2295: 
                   2296: Here's a loop with an exit at the end:
                   2297: 
                   2298: @example
                   2299: : log2 ( +n1 -- n2 )
                   2300: \ logarithmus dualis of n1>0, rounded down to the next integer
                   2301:   assert( dup 0 > )
                   2302:   -1 begin
                   2303:     1+ swap 2/ swap
                   2304:     over 0 <=
                   2305:   until
                   2306:   nip ;
                   2307: @end example
                   2308: 
                   2309: @code{Until} consumes a flag; if it is non-zero, execution continues at
                   2310: the @code{begin}, otherwise after the @code{until}.
                   2311: 
                   2312: @assignment
                   2313: Write a definition for computing the greatest common divisor.
                   2314: @endassignment
                   2315: 
1.66      anton    2316: Reference: @ref{Simple Loops}.
                   2317: 
1.48      anton    2318: 
                   2319: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
                   2320: @section Counted loops
1.66      anton    2321: @cindex loops, counted, tutorial
1.48      anton    2322: 
                   2323: @example
                   2324: : ^ ( n1 u -- n )
                   2325: \ n = the uth power of u1
                   2326:   1 swap 0 u+do
                   2327:     over *
                   2328:   loop
                   2329:   nip ;
                   2330: 3 2 ^ .
                   2331: 4 3 ^ .
                   2332: @end example
                   2333: 
                   2334: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
                   2335: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
                   2336: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
                   2337: times (or not at all, if @code{u3-u4<0}).
                   2338: 
                   2339: You can see the stack effect design rules at work in the stack effect of
                   2340: the loop start words: Since the start value of the loop is more
                   2341: frequently constant than the end value, the start value is passed on
                   2342: the top-of-stack.
                   2343: 
                   2344: You can access the counter of a counted loop with @code{i}:
                   2345: 
                   2346: @example
                   2347: : fac ( u -- u! )
                   2348:   1 swap 1+ 1 u+do
                   2349:     i *
                   2350:   loop ;
                   2351: 5 fac .
                   2352: 7 fac .
                   2353: @end example
                   2354: 
                   2355: There is also @code{+do}, which expects signed numbers (important for
                   2356: deciding whether to enter the loop).
                   2357: 
                   2358: @assignment
                   2359: Write a definition for computing the nth Fibonacci number.
                   2360: @endassignment
                   2361: 
1.65      anton    2362: You can also use increments other than 1:
                   2363: 
                   2364: @example
                   2365: : up2 ( n1 n2 -- )
                   2366:   +do
                   2367:     i .
                   2368:   2 +loop ;
                   2369: 10 0 up2
                   2370: 
                   2371: : down2 ( n1 n2 -- )
                   2372:   -do
                   2373:     i .
                   2374:   2 -loop ;
                   2375: 0 10 down2
                   2376: @end example
1.48      anton    2377: 
1.66      anton    2378: Reference: @ref{Counted Loops}.
                   2379: 
1.48      anton    2380: 
                   2381: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
                   2382: @section Recursion
1.66      anton    2383: @cindex recursion tutorial
1.48      anton    2384: 
                   2385: Usually the name of a definition is not visible in the definition; but
                   2386: earlier definitions are usually visible:
                   2387: 
                   2388: @example
                   2389: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
                   2390: : / ( n1 n2 -- n )
                   2391:   dup 0= if
                   2392:     -10 throw \ report division by zero
                   2393:   endif
                   2394:   /           \ old version
                   2395: ;
                   2396: 1 0 /
                   2397: @end example
                   2398: 
                   2399: For recursive definitions you can use @code{recursive} (non-standard) or
                   2400: @code{recurse}:
                   2401: 
                   2402: @example
                   2403: : fac1 ( n -- n! ) recursive
                   2404:  dup 0> if
                   2405:    dup 1- fac1 *
                   2406:  else
                   2407:    drop 1
                   2408:  endif ;
                   2409: 7 fac1 .
                   2410: 
                   2411: : fac2 ( n -- n! )
                   2412:  dup 0> if
                   2413:    dup 1- recurse *
                   2414:  else
                   2415:    drop 1
                   2416:  endif ;
                   2417: 8 fac2 .
                   2418: @end example
                   2419: 
                   2420: @assignment
                   2421: Write a recursive definition for computing the nth Fibonacci number.
                   2422: @endassignment
                   2423: 
1.66      anton    2424: Reference (including indirect recursion): @xref{Calls and returns}.
                   2425: 
1.48      anton    2426: 
                   2427: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
                   2428: @section Leaving definitions or loops
1.66      anton    2429: @cindex leaving definitions, tutorial
                   2430: @cindex leaving loops, tutorial
1.48      anton    2431: 
                   2432: @code{EXIT} exits the current definition right away.  For every counted
                   2433: loop that is left in this way, an @code{UNLOOP} has to be performed
                   2434: before the @code{EXIT}:
                   2435: 
                   2436: @c !! real examples
                   2437: @example
                   2438: : ...
                   2439:  ... u+do
                   2440:    ... if
                   2441:      ... unloop exit
                   2442:    endif
                   2443:    ...
                   2444:  loop
                   2445:  ... ;
                   2446: @end example
                   2447: 
                   2448: @code{LEAVE} leaves the innermost counted loop right away:
                   2449: 
                   2450: @example
                   2451: : ...
                   2452:  ... u+do
                   2453:    ... if
                   2454:      ... leave
                   2455:    endif
                   2456:    ...
                   2457:  loop
                   2458:  ... ;
                   2459: @end example
                   2460: 
1.65      anton    2461: @c !! example
1.48      anton    2462: 
1.66      anton    2463: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
                   2464: 
                   2465: 
1.48      anton    2466: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
                   2467: @section Return Stack
1.66      anton    2468: @cindex return stack tutorial
1.48      anton    2469: 
                   2470: In addition to the data stack Forth also has a second stack, the return
                   2471: stack; most Forth systems store the return addresses of procedure calls
                   2472: there (thus its name).  Programmers can also use this stack:
                   2473: 
                   2474: @example
                   2475: : foo ( n1 n2 -- )
                   2476:  .s
                   2477:  >r .s
1.50      anton    2478:  r@@ .
1.48      anton    2479:  >r .s
1.50      anton    2480:  r@@ .
1.48      anton    2481:  r> .
1.50      anton    2482:  r@@ .
1.48      anton    2483:  r> . ;
                   2484: 1 2 foo
                   2485: @end example
                   2486: 
                   2487: @code{>r} takes an element from the data stack and pushes it onto the
                   2488: return stack; conversely, @code{r>} moves an elementm from the return to
                   2489: the data stack; @code{r@@} pushes a copy of the top of the return stack
                   2490: on the return stack.
                   2491: 
                   2492: Forth programmers usually use the return stack for storing data
                   2493: temporarily, if using the data stack alone would be too complex, and
                   2494: factoring and locals are not an option:
                   2495: 
                   2496: @example
                   2497: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
                   2498:  rot >r rot r> ;
                   2499: @end example
                   2500: 
                   2501: The return address of the definition and the loop control parameters of
                   2502: counted loops usually reside on the return stack, so you have to take
                   2503: all items, that you have pushed on the return stack in a colon
                   2504: definition or counted loop, from the return stack before the definition
                   2505: or loop ends.  You cannot access items that you pushed on the return
                   2506: stack outside some definition or loop within the definition of loop.
                   2507: 
                   2508: If you miscount the return stack items, this usually ends in a crash:
                   2509: 
                   2510: @example
                   2511: : crash ( n -- )
                   2512:   >r ;
                   2513: 5 crash
                   2514: @end example
                   2515: 
                   2516: You cannot mix using locals and using the return stack (according to the
                   2517: standard; Gforth has no problem).  However, they solve the same
                   2518: problems, so this shouldn't be an issue.
                   2519: 
                   2520: @assignment
                   2521: Can you rewrite any of the definitions you wrote until now in a better
                   2522: way using the return stack?
                   2523: @endassignment
                   2524: 
1.66      anton    2525: Reference: @ref{Return stack}.
                   2526: 
1.48      anton    2527: 
                   2528: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
                   2529: @section Memory
1.66      anton    2530: @cindex memory access/allocation tutorial
1.48      anton    2531: 
                   2532: You can create a global variable @code{v} with
                   2533: 
                   2534: @example
                   2535: variable v ( -- addr )
                   2536: @end example
                   2537: 
                   2538: @code{v} pushes the address of a cell in memory on the stack.  This cell
                   2539: was reserved by @code{variable}.  You can use @code{!} (store) to store
                   2540: values into this cell and @code{@@} (fetch) to load the value from the
                   2541: stack into memory:
                   2542: 
                   2543: @example
                   2544: v .
                   2545: 5 v ! .s
1.50      anton    2546: v @@ .
1.48      anton    2547: @end example
                   2548: 
1.65      anton    2549: You can see a raw dump of memory with @code{dump}:
                   2550: 
                   2551: @example
                   2552: v 1 cells .s dump
                   2553: @end example
                   2554: 
                   2555: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
                   2556: generally, address units (aus)) that @code{n1 cells} occupy.  You can
                   2557: also reserve more memory:
1.48      anton    2558: 
                   2559: @example
                   2560: create v2 20 cells allot
1.65      anton    2561: v2 20 cells dump
1.48      anton    2562: @end example
                   2563: 
1.65      anton    2564: creates a word @code{v2} and reserves 20 uninitialized cells; the
                   2565: address pushed by @code{v2} points to the start of these 20 cells.  You
                   2566: can use address arithmetic to access these cells:
1.48      anton    2567: 
                   2568: @example
                   2569: 3 v2 5 cells + !
1.65      anton    2570: v2 20 cells dump
1.48      anton    2571: @end example
                   2572: 
                   2573: You can reserve and initialize memory with @code{,}:
                   2574: 
                   2575: @example
                   2576: create v3
                   2577:   5 , 4 , 3 , 2 , 1 ,
1.50      anton    2578: v3 @@ .
                   2579: v3 cell+ @@ .
                   2580: v3 2 cells + @@ .
1.65      anton    2581: v3 5 cells dump
1.48      anton    2582: @end example
                   2583: 
                   2584: @assignment
                   2585: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
                   2586: @code{u} cells, with the first of these cells at @code{addr}, the next
                   2587: one at @code{addr cell+} etc.
                   2588: @endassignment
                   2589: 
                   2590: You can also reserve memory without creating a new word:
                   2591: 
                   2592: @example
1.60      anton    2593: here 10 cells allot .
                   2594: here .
1.48      anton    2595: @end example
                   2596: 
                   2597: @code{Here} pushes the start address of the memory area.  You should
                   2598: store it somewhere, or you will have a hard time finding the memory area
                   2599: again.
                   2600: 
                   2601: @code{Allot} manages dictionary memory.  The dictionary memory contains
                   2602: the system's data structures for words etc. on Gforth and most other
                   2603: Forth systems.  It is managed like a stack: You can free the memory that
                   2604: you have just @code{allot}ed with
                   2605: 
                   2606: @example
                   2607: -10 cells allot
1.60      anton    2608: here .
1.48      anton    2609: @end example
                   2610: 
                   2611: Note that you cannot do this if you have created a new word in the
                   2612: meantime (because then your @code{allot}ed memory is no longer on the
                   2613: top of the dictionary ``stack'').
                   2614: 
                   2615: Alternatively, you can use @code{allocate} and @code{free} which allow
                   2616: freeing memory in any order:
                   2617: 
                   2618: @example
                   2619: 10 cells allocate throw .s
                   2620: 20 cells allocate throw .s
                   2621: swap
                   2622: free throw
                   2623: free throw
                   2624: @end example
                   2625: 
                   2626: The @code{throw}s deal with errors (e.g., out of memory).
                   2627: 
1.65      anton    2628: And there is also a
                   2629: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
                   2630: garbage collector}, which eliminates the need to @code{free} memory
                   2631: explicitly.
1.48      anton    2632: 
1.66      anton    2633: Reference: @ref{Memory}.
                   2634: 
1.48      anton    2635: 
                   2636: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
                   2637: @section Characters and Strings
1.66      anton    2638: @cindex strings tutorial
                   2639: @cindex characters tutorial
1.48      anton    2640: 
                   2641: On the stack characters take up a cell, like numbers.  In memory they
                   2642: have their own size (one 8-bit byte on most systems), and therefore
                   2643: require their own words for memory access:
                   2644: 
                   2645: @example
                   2646: create v4 
                   2647:   104 c, 97 c, 108 c, 108 c, 111 c,
1.50      anton    2648: v4 4 chars + c@@ .
1.65      anton    2649: v4 5 chars dump
1.48      anton    2650: @end example
                   2651: 
                   2652: The preferred representation of strings on the stack is @code{addr
                   2653: u-count}, where @code{addr} is the address of the first character and
                   2654: @code{u-count} is the number of characters in the string.
                   2655: 
                   2656: @example
                   2657: v4 5 type
                   2658: @end example
                   2659: 
                   2660: You get a string constant with
                   2661: 
                   2662: @example
                   2663: s" hello, world" .s
                   2664: type
                   2665: @end example
                   2666: 
                   2667: Make sure you have a space between @code{s"} and the string; @code{s"}
                   2668: is a normal Forth word and must be delimited with white space (try what
                   2669: happens when you remove the space).
                   2670: 
                   2671: However, this interpretive use of @code{s"} is quite restricted: the
                   2672: string exists only until the next call of @code{s"} (some Forth systems
                   2673: keep more than one of these strings, but usually they still have a
1.62      crook    2674: limited lifetime).
1.48      anton    2675: 
                   2676: @example
                   2677: s" hello," s" world" .s
                   2678: type
                   2679: type
                   2680: @end example
                   2681: 
1.62      crook    2682: You can also use @code{s"} in a definition, and the resulting
                   2683: strings then live forever (well, for as long as the definition):
1.48      anton    2684: 
                   2685: @example
                   2686: : foo s" hello," s" world" ;
                   2687: foo .s
                   2688: type
                   2689: type
                   2690: @end example
                   2691: 
                   2692: @assignment
                   2693: @code{Emit ( c -- )} types @code{c} as character (not a number).
                   2694: Implement @code{type ( addr u -- )}.
                   2695: @endassignment
                   2696: 
1.66      anton    2697: Reference: @ref{Memory Blocks}.
                   2698: 
                   2699: 
1.84      pazsan   2700: @node Alignment Tutorial, Files Tutorial, Characters and Strings Tutorial, Tutorial
1.48      anton    2701: @section Alignment
1.66      anton    2702: @cindex alignment tutorial
                   2703: @cindex memory alignment tutorial
1.48      anton    2704: 
                   2705: On many processors cells have to be aligned in memory, if you want to
                   2706: access them with @code{@@} and @code{!} (and even if the processor does
1.62      crook    2707: not require alignment, access to aligned cells is faster).
1.48      anton    2708: 
                   2709: @code{Create} aligns @code{here} (i.e., the place where the next
                   2710: allocation will occur, and that the @code{create}d word points to).
                   2711: Likewise, the memory produced by @code{allocate} starts at an aligned
                   2712: address.  Adding a number of @code{cells} to an aligned address produces
                   2713: another aligned address.
                   2714: 
                   2715: However, address arithmetic involving @code{char+} and @code{chars} can
                   2716: create an address that is not cell-aligned.  @code{Aligned ( addr --
                   2717: a-addr )} produces the next aligned address:
                   2718: 
                   2719: @example
1.50      anton    2720: v3 char+ aligned .s @@ .
                   2721: v3 char+ .s @@ .
1.48      anton    2722: @end example
                   2723: 
                   2724: Similarly, @code{align} advances @code{here} to the next aligned
                   2725: address:
                   2726: 
                   2727: @example
                   2728: create v5 97 c,
                   2729: here .
                   2730: align here .
                   2731: 1000 ,
                   2732: @end example
                   2733: 
                   2734: Note that you should use aligned addresses even if your processor does
                   2735: not require them, if you want your program to be portable.
                   2736: 
1.66      anton    2737: Reference: @ref{Address arithmetic}.
                   2738: 
1.48      anton    2739: 
1.84      pazsan   2740: @node Files Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Alignment Tutorial, Tutorial
                   2741: @section Files
                   2742: @cindex files tutorial
                   2743: 
                   2744: This section gives a short introduction into how to use files inside
                   2745: Forth. It's broken up into five easy steps:
                   2746: 
                   2747: @enumerate 1
                   2748: @item Opened an ASCII text file for input
                   2749: @item Opened a file for output
                   2750: @item Read input file until string matched (or some other condition matched)
                   2751: @item Wrote some lines from input ( modified or not) to output
                   2752: @item Closed the files.
                   2753: @end enumerate
                   2754: 
                   2755: @subsection Open file for input
                   2756: 
                   2757: @example
                   2758: s" foo.in"  r/o open-file throw Value fd-in
                   2759: @end example
                   2760: 
                   2761: @subsection Create file for output
                   2762: 
                   2763: @example
                   2764: s" foo.out" w/o create-file throw Value fd-out
                   2765: @end example
                   2766: 
                   2767: The available file modes are r/o for read-only access, r/w for
                   2768: read-write access, and w/o for write-only access. You could open both
                   2769: files with r/w, too, if you like. All file words return error codes; for
                   2770: most applications, it's best to pass there error codes with @code{throw}
                   2771: to the outer error handler.
                   2772: 
                   2773: If you want words for opening and assigning, define them as follows:
                   2774: 
                   2775: @example
                   2776: 0 Value fd-in
                   2777: 0 Value fd-out
                   2778: : open-input ( addr u -- )  r/o open-file throw to fd-in ;
                   2779: : open-output ( addr u -- )  w/o create-file throw to fd-out ;
                   2780: @end example
                   2781: 
                   2782: Usage example:
                   2783: 
                   2784: @example
                   2785: s" foo.in" open-input
                   2786: s" foo.out" open-output
                   2787: @end example
                   2788: 
                   2789: @subsection Scan file for a particular line
                   2790: 
                   2791: @example
                   2792: 256 Constant max-line
                   2793: Create line-buffer  max-line 2 + allot
                   2794: 
                   2795: : scan-file ( addr u -- )
                   2796:   begin
                   2797:       line-buffer max-line fd-in read-line throw
                   2798:   while
                   2799:          >r 2dup line-buffer r> compare 0=
                   2800:      until
                   2801:   else
                   2802:      drop
                   2803:   then
                   2804:   2drop ;
                   2805: @end example
                   2806: 
                   2807: @code{read-line ( addr u1 fd -- u2 flag ior )} reads up to u1 bytes into
                   2808: the buffer at addr, and returns the number of bytes read, a flag that's
                   2809: true when the end of file is reached, and an error code.
                   2810: 
                   2811: @code{compare ( addr1 u1 addr2 u2 -- n )} compares two strings and
                   2812: returns zero if both strings are equal. It returns a positive number if
                   2813: the first string is lexically greater, a negative if the second string
                   2814: is lexically greater.
                   2815: 
                   2816: We haven't seen this loop here; it has two exits. Since the @code{while}
                   2817: exits with the number of bytes read on the stack, we have to clean up
                   2818: that separately; that's after the @code{else}.
                   2819: 
                   2820: Usage example:
                   2821: 
                   2822: @example
                   2823: s" The text I search is here" scan-file
                   2824: @end example
                   2825: 
                   2826: @subsection Copy input to output
                   2827: 
                   2828: @example
                   2829: : copy-file ( -- )
                   2830:   begin
                   2831:       line-buffer max-line fd-in read-line throw
                   2832:   while
                   2833:       line-buffer swap fd-out write-file throw
                   2834:   repeat ;
                   2835: @end example
                   2836: 
                   2837: @subsection Close files
                   2838: 
                   2839: @example
                   2840: fd-in close-file throw
                   2841: fd-out close-file throw
                   2842: @end example
                   2843: 
                   2844: Likewise, you can put that into definitions, too:
                   2845: 
                   2846: @example
                   2847: : close-input ( -- )  fd-in close-file throw ;
                   2848: : close-output ( -- )  fd-out close-file throw ;
                   2849: @end example
                   2850: 
                   2851: @assignment
                   2852: How could you modify @code{copy-file} so that it copies until a second line is
                   2853: matched? Can you write a program that extracts a section of a text file,
                   2854: given the line that starts and the line that terminates that section?
                   2855: @endassignment
                   2856: 
                   2857: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Files Tutorial, Tutorial
1.48      anton    2858: @section Interpretation and Compilation Semantics and Immediacy
1.66      anton    2859: @cindex semantics tutorial
                   2860: @cindex interpretation semantics tutorial
                   2861: @cindex compilation semantics tutorial
                   2862: @cindex immediate, tutorial
1.48      anton    2863: 
                   2864: When a word is compiled, it behaves differently from being interpreted.
                   2865: E.g., consider @code{+}:
                   2866: 
                   2867: @example
                   2868: 1 2 + .
                   2869: : foo + ;
                   2870: @end example
                   2871: 
                   2872: These two behaviours are known as compilation and interpretation
                   2873: semantics.  For normal words (e.g., @code{+}), the compilation semantics
                   2874: is to append the interpretation semantics to the currently defined word
                   2875: (@code{foo} in the example above).  I.e., when @code{foo} is executed
                   2876: later, the interpretation semantics of @code{+} (i.e., adding two
                   2877: numbers) will be performed.
                   2878: 
                   2879: However, there are words with non-default compilation semantics, e.g.,
                   2880: the control-flow words like @code{if}.  You can use @code{immediate} to
                   2881: change the compilation semantics of the last defined word to be equal to
                   2882: the interpretation semantics:
                   2883: 
                   2884: @example
                   2885: : [FOO] ( -- )
                   2886:  5 . ; immediate
                   2887: 
                   2888: [FOO]
                   2889: : bar ( -- )
                   2890:   [FOO] ;
                   2891: bar
                   2892: see bar
                   2893: @end example
                   2894: 
                   2895: Two conventions to mark words with non-default compilation semnatics are
                   2896: names with brackets (more frequently used) and to write them all in
                   2897: upper case (less frequently used).
                   2898: 
                   2899: In Gforth (and many other systems) you can also remove the
                   2900: interpretation semantics with @code{compile-only} (the compilation
                   2901: semantics is derived from the original interpretation semantics):
                   2902: 
                   2903: @example
                   2904: : flip ( -- )
                   2905:  6 . ; compile-only \ but not immediate
                   2906: flip
                   2907: 
                   2908: : flop ( -- )
                   2909:  flip ;
                   2910: flop
                   2911: @end example
                   2912: 
                   2913: In this example the interpretation semantics of @code{flop} is equal to
                   2914: the original interpretation semantics of @code{flip}.
                   2915: 
                   2916: The text interpreter has two states: in interpret state, it performs the
                   2917: interpretation semantics of words it encounters; in compile state, it
                   2918: performs the compilation semantics of these words.
                   2919: 
                   2920: Among other things, @code{:} switches into compile state, and @code{;}
                   2921: switches back to interpret state.  They contain the factors @code{]}
                   2922: (switch to compile state) and @code{[} (switch to interpret state), that
                   2923: do nothing but switch the state.
                   2924: 
                   2925: @example
                   2926: : xxx ( -- )
                   2927:   [ 5 . ]
                   2928: ;
                   2929: 
                   2930: xxx
                   2931: see xxx
                   2932: @end example
                   2933: 
                   2934: These brackets are also the source of the naming convention mentioned
                   2935: above.
                   2936: 
1.66      anton    2937: Reference: @ref{Interpretation and Compilation Semantics}.
                   2938: 
1.48      anton    2939: 
                   2940: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
                   2941: @section Execution Tokens
1.66      anton    2942: @cindex execution tokens tutorial
                   2943: @cindex XT tutorial
1.48      anton    2944: 
                   2945: @code{' word} gives you the execution token (XT) of a word.  The XT is a
                   2946: cell representing the interpretation semantics of a word.  You can
                   2947: execute this semantics with @code{execute}:
                   2948: 
                   2949: @example
                   2950: ' + .s
                   2951: 1 2 rot execute .
                   2952: @end example
                   2953: 
                   2954: The XT is similar to a function pointer in C.  However, parameter
                   2955: passing through the stack makes it a little more flexible:
                   2956: 
                   2957: @example
                   2958: : map-array ( ... addr u xt -- ... )
1.50      anton    2959: \ executes xt ( ... x -- ... ) for every element of the array starting
                   2960: \ at addr and containing u elements
1.48      anton    2961:   @{ xt @}
                   2962:   cells over + swap ?do
1.50      anton    2963:     i @@ xt execute
1.48      anton    2964:   1 cells +loop ;
                   2965: 
                   2966: create a 3 , 4 , 2 , -1 , 4 ,
                   2967: a 5 ' . map-array .s
                   2968: 0 a 5 ' + map-array .
                   2969: s" max-n" environment? drop .s
                   2970: a 5 ' min map-array .
                   2971: @end example
                   2972: 
                   2973: You can use map-array with the XTs of words that consume one element
                   2974: more than they produce.  In theory you can also use it with other XTs,
                   2975: but the stack effect then depends on the size of the array, which is
                   2976: hard to understand.
                   2977: 
1.51      pazsan   2978: Since XTs are cell-sized, you can store them in memory and manipulate
                   2979: them on the stack like other cells.  You can also compile the XT into a
1.48      anton    2980: word with @code{compile,}:
                   2981: 
                   2982: @example
                   2983: : foo1 ( n1 n2 -- n )
                   2984:    [ ' + compile, ] ;
                   2985: see foo
                   2986: @end example
                   2987: 
                   2988: This is non-standard, because @code{compile,} has no compilation
                   2989: semantics in the standard, but it works in good Forth systems.  For the
                   2990: broken ones, use
                   2991: 
                   2992: @example
                   2993: : [compile,] compile, ; immediate
                   2994: 
                   2995: : foo1 ( n1 n2 -- n )
                   2996:    [ ' + ] [compile,] ;
                   2997: see foo
                   2998: @end example
                   2999: 
                   3000: @code{'} is a word with default compilation semantics; it parses the
                   3001: next word when its interpretation semantics are executed, not during
                   3002: compilation:
                   3003: 
                   3004: @example
                   3005: : foo ( -- xt )
                   3006:   ' ;
                   3007: see foo
                   3008: : bar ( ... "word" -- ... )
                   3009:   ' execute ;
                   3010: see bar
1.60      anton    3011: 1 2 bar + .
1.48      anton    3012: @end example
                   3013: 
                   3014: You often want to parse a word during compilation and compile its XT so
                   3015: it will be pushed on the stack at run-time.  @code{[']} does this:
                   3016: 
                   3017: @example
                   3018: : xt-+ ( -- xt )
                   3019:   ['] + ;
                   3020: see xt-+
                   3021: 1 2 xt-+ execute .
                   3022: @end example
                   3023: 
                   3024: Many programmers tend to see @code{'} and the word it parses as one
                   3025: unit, and expect it to behave like @code{[']} when compiled, and are
                   3026: confused by the actual behaviour.  If you are, just remember that the
                   3027: Forth system just takes @code{'} as one unit and has no idea that it is
                   3028: a parsing word (attempts to convenience programmers in this issue have
                   3029: usually resulted in even worse pitfalls, see
1.66      anton    3030: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
                   3031: @code{State}-smartness---Why it is evil and How to Exorcise it}).
1.48      anton    3032: 
                   3033: Note that the state of the interpreter does not come into play when
1.51      pazsan   3034: creating and executing XTs.  I.e., even when you execute @code{'} in
1.48      anton    3035: compile state, it still gives you the interpretation semantics.  And
                   3036: whatever that state is, @code{execute} performs the semantics
1.66      anton    3037: represented by the XT (i.e., for XTs produced with @code{'} the
                   3038: interpretation semantics).
                   3039: 
                   3040: Reference: @ref{Tokens for Words}.
1.48      anton    3041: 
                   3042: 
                   3043: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
                   3044: @section Exceptions
1.66      anton    3045: @cindex exceptions tutorial
1.48      anton    3046: 
                   3047: @code{throw ( n -- )} causes an exception unless n is zero.
                   3048: 
                   3049: @example
                   3050: 100 throw .s
                   3051: 0 throw .s
                   3052: @end example
                   3053: 
                   3054: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
                   3055: it catches exceptions and pushes the number of the exception on the
                   3056: stack (or 0, if the xt executed without exception).  If there was an
                   3057: exception, the stacks have the same depth as when entering @code{catch}:
                   3058: 
                   3059: @example
                   3060: .s
                   3061: 3 0 ' / catch .s
                   3062: 3 2 ' / catch .s
                   3063: @end example
                   3064: 
                   3065: @assignment
                   3066: Try the same with @code{execute} instead of @code{catch}.
                   3067: @endassignment
                   3068: 
                   3069: @code{Throw} always jumps to the dynamically next enclosing
                   3070: @code{catch}, even if it has to leave several call levels to achieve
                   3071: this:
                   3072: 
                   3073: @example
                   3074: : foo 100 throw ;
                   3075: : foo1 foo ." after foo" ;
1.51      pazsan   3076: : bar ['] foo1 catch ;
1.60      anton    3077: bar .
1.48      anton    3078: @end example
                   3079: 
                   3080: It is often important to restore a value upon leaving a definition, even
                   3081: if the definition is left through an exception.  You can ensure this
                   3082: like this:
                   3083: 
                   3084: @example
                   3085: : ...
                   3086:    save-x
1.51      pazsan   3087:    ['] word-changing-x catch ( ... n )
1.48      anton    3088:    restore-x
                   3089:    ( ... n ) throw ;
                   3090: @end example
                   3091: 
1.55      anton    3092: Gforth provides an alternative syntax in addition to @code{catch}:
1.48      anton    3093: @code{try ... recover ... endtry}.  If the code between @code{try} and
                   3094: @code{recover} has an exception, the stack depths are restored, the
                   3095: exception number is pushed on the stack, and the code between
                   3096: @code{recover} and @code{endtry} is performed.  E.g., the definition for
                   3097: @code{catch} is
                   3098: 
                   3099: @example
                   3100: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
                   3101:   try
                   3102:     execute 0
                   3103:   recover
                   3104:     nip
                   3105:   endtry ;
                   3106: @end example
                   3107: 
                   3108: The equivalent to the restoration code above is
                   3109: 
                   3110: @example
                   3111: : ...
                   3112:   save-x
                   3113:   try
                   3114:     word-changing-x
                   3115:   end-try
                   3116:   restore-x
                   3117:   throw ;
                   3118: @end example
                   3119: 
                   3120: As you can see, the @code{recover} part is optional.
                   3121: 
1.66      anton    3122: Reference: @ref{Exception Handling}.
                   3123: 
1.48      anton    3124: 
                   3125: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
                   3126: @section Defining Words
1.66      anton    3127: @cindex defining words tutorial
                   3128: @cindex does> tutorial
                   3129: @cindex create...does> tutorial
                   3130: 
                   3131: @c before semantics?
1.48      anton    3132: 
                   3133: @code{:}, @code{create}, and @code{variable} are definition words: They
                   3134: define other words.  @code{Constant} is another definition word:
                   3135: 
                   3136: @example
                   3137: 5 constant foo
                   3138: foo .
                   3139: @end example
                   3140: 
                   3141: You can also use the prefixes @code{2} (double-cell) and @code{f}
                   3142: (floating point) with @code{variable} and @code{constant}.
                   3143: 
                   3144: You can also define your own defining words.  E.g.:
                   3145: 
                   3146: @example
                   3147: : variable ( "name" -- )
                   3148:   create 0 , ;
                   3149: @end example
                   3150: 
                   3151: You can also define defining words that create words that do something
                   3152: other than just producing their address:
                   3153: 
                   3154: @example
                   3155: : constant ( n "name" -- )
                   3156:   create ,
                   3157: does> ( -- n )
1.50      anton    3158:   ( addr ) @@ ;
1.48      anton    3159: 
                   3160: 5 constant foo
                   3161: foo .
                   3162: @end example
                   3163: 
                   3164: The definition of @code{constant} above ends at the @code{does>}; i.e.,
                   3165: @code{does>} replaces @code{;}, but it also does something else: It
                   3166: changes the last defined word such that it pushes the address of the
                   3167: body of the word and then performs the code after the @code{does>}
                   3168: whenever it is called.
                   3169: 
                   3170: In the example above, @code{constant} uses @code{,} to store 5 into the
                   3171: body of @code{foo}.  When @code{foo} executes, it pushes the address of
                   3172: the body onto the stack, then (in the code after the @code{does>})
                   3173: fetches the 5 from there.
                   3174: 
                   3175: The stack comment near the @code{does>} reflects the stack effect of the
                   3176: defined word, not the stack effect of the code after the @code{does>}
                   3177: (the difference is that the code expects the address of the body that
                   3178: the stack comment does not show).
                   3179: 
                   3180: You can use these definition words to do factoring in cases that involve
                   3181: (other) definition words.  E.g., a field offset is always added to an
                   3182: address.  Instead of defining
                   3183: 
                   3184: @example
                   3185: 2 cells constant offset-field1
                   3186: @end example
                   3187: 
                   3188: and using this like
                   3189: 
                   3190: @example
                   3191: ( addr ) offset-field1 +
                   3192: @end example
                   3193: 
                   3194: you can define a definition word
                   3195: 
                   3196: @example
                   3197: : simple-field ( n "name" -- )
                   3198:   create ,
                   3199: does> ( n1 -- n1+n )
1.50      anton    3200:   ( addr ) @@ + ;
1.48      anton    3201: @end example
1.21      crook    3202: 
1.48      anton    3203: Definition and use of field offsets now look like this:
1.21      crook    3204: 
1.48      anton    3205: @example
                   3206: 2 cells simple-field field1
1.60      anton    3207: create mystruct 4 cells allot
                   3208: mystruct .s field1 .s drop
1.48      anton    3209: @end example
1.21      crook    3210: 
1.48      anton    3211: If you want to do something with the word without performing the code
                   3212: after the @code{does>}, you can access the body of a @code{create}d word
                   3213: with @code{>body ( xt -- addr )}:
1.21      crook    3214: 
1.48      anton    3215: @example
                   3216: : value ( n "name" -- )
                   3217:   create ,
                   3218: does> ( -- n1 )
1.50      anton    3219:   @@ ;
1.48      anton    3220: : to ( n "name" -- )
                   3221:   ' >body ! ;
1.21      crook    3222: 
1.48      anton    3223: 5 value foo
                   3224: foo .
                   3225: 7 to foo
                   3226: foo .
                   3227: @end example
1.21      crook    3228: 
1.48      anton    3229: @assignment
                   3230: Define @code{defer ( "name" -- )}, which creates a word that stores an
                   3231: XT (at the start the XT of @code{abort}), and upon execution
                   3232: @code{execute}s the XT.  Define @code{is ( xt "name" -- )} that stores
                   3233: @code{xt} into @code{name}, a word defined with @code{defer}.  Indirect
                   3234: recursion is one application of @code{defer}.
                   3235: @endassignment
1.29      crook    3236: 
1.66      anton    3237: Reference: @ref{User-defined Defining Words}.
                   3238: 
                   3239: 
1.48      anton    3240: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
                   3241: @section Arrays and Records
1.66      anton    3242: @cindex arrays tutorial
                   3243: @cindex records tutorial
                   3244: @cindex structs tutorial
1.29      crook    3245: 
1.48      anton    3246: Forth has no standard words for defining data structures such as arrays
                   3247: and records (structs in C terminology), but you can build them yourself
                   3248: based on address arithmetic.  You can also define words for defining
                   3249: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
1.29      crook    3250: 
1.48      anton    3251: One of the first projects a Forth newcomer sets out upon when learning
                   3252: about defining words is an array defining word (possibly for
                   3253: n-dimensional arrays).  Go ahead and do it, I did it, too; you will
                   3254: learn something from it.  However, don't be disappointed when you later
                   3255: learn that you have little use for these words (inappropriate use would
                   3256: be even worse).  I have not yet found a set of useful array words yet;
                   3257: the needs are just too diverse, and named, global arrays (the result of
                   3258: naive use of defining words) are often not flexible enough (e.g.,
1.66      anton    3259: consider how to pass them as parameters).  Another such project is a set
                   3260: of words to help dealing with strings.
1.29      crook    3261: 
1.48      anton    3262: On the other hand, there is a useful set of record words, and it has
                   3263: been defined in @file{compat/struct.fs}; these words are predefined in
                   3264: Gforth.  They are explained in depth elsewhere in this manual (see
                   3265: @pxref{Structures}).  The @code{simple-field} example above is
                   3266: simplified variant of fields in this package.
1.21      crook    3267: 
                   3268: 
1.48      anton    3269: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
                   3270: @section @code{POSTPONE}
1.66      anton    3271: @cindex postpone tutorial
1.21      crook    3272: 
1.48      anton    3273: You can compile the compilation semantics (instead of compiling the
                   3274: interpretation semantics) of a word with @code{POSTPONE}:
1.21      crook    3275: 
1.48      anton    3276: @example
                   3277: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
1.51      pazsan   3278:  POSTPONE + ; immediate
1.48      anton    3279: : foo ( n1 n2 -- n )
                   3280:  MY-+ ;
                   3281: 1 2 foo .
                   3282: see foo
                   3283: @end example
1.21      crook    3284: 
1.48      anton    3285: During the definition of @code{foo} the text interpreter performs the
                   3286: compilation semantics of @code{MY-+}, which performs the compilation
                   3287: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
                   3288: 
                   3289: This example also displays separate stack comments for the compilation
                   3290: semantics and for the stack effect of the compiled code.  For words with
                   3291: default compilation semantics these stack effects are usually not
                   3292: displayed; the stack effect of the compilation semantics is always
                   3293: @code{( -- )} for these words, the stack effect for the compiled code is
                   3294: the stack effect of the interpretation semantics.
                   3295: 
                   3296: Note that the state of the interpreter does not come into play when
                   3297: performing the compilation semantics in this way.  You can also perform
                   3298: it interpretively, e.g.:
                   3299: 
                   3300: @example
                   3301: : foo2 ( n1 n2 -- n )
                   3302:  [ MY-+ ] ;
                   3303: 1 2 foo .
                   3304: see foo
                   3305: @end example
1.21      crook    3306: 
1.48      anton    3307: However, there are some broken Forth systems where this does not always
1.62      crook    3308: work, and therefore this practice was been declared non-standard in
1.48      anton    3309: 1999.
                   3310: @c !! repair.fs
                   3311: 
                   3312: Here is another example for using @code{POSTPONE}:
1.44      crook    3313: 
1.48      anton    3314: @example
                   3315: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
                   3316:  POSTPONE negate POSTPONE + ; immediate compile-only
                   3317: : bar ( n1 n2 -- n )
                   3318:   MY-- ;
                   3319: 2 1 bar .
                   3320: see bar
                   3321: @end example
1.21      crook    3322: 
1.48      anton    3323: You can define @code{ENDIF} in this way:
1.21      crook    3324: 
1.48      anton    3325: @example
                   3326: : ENDIF ( Compilation: orig -- )
                   3327:   POSTPONE then ; immediate
                   3328: @end example
1.21      crook    3329: 
1.48      anton    3330: @assignment
                   3331: Write @code{MY-2DUP} that has compilation semantics equivalent to
                   3332: @code{2dup}, but compiles @code{over over}.
                   3333: @endassignment
1.29      crook    3334: 
1.66      anton    3335: @c !! @xref{Macros} for reference
                   3336: 
                   3337: 
1.48      anton    3338: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
                   3339: @section @code{Literal}
1.66      anton    3340: @cindex literal tutorial
1.29      crook    3341: 
1.48      anton    3342: You cannot @code{POSTPONE} numbers:
1.21      crook    3343: 
1.48      anton    3344: @example
                   3345: : [FOO] POSTPONE 500 ; immediate
1.21      crook    3346: @end example
                   3347: 
1.48      anton    3348: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
1.29      crook    3349: 
1.48      anton    3350: @example
                   3351: : [FOO] ( compilation: --; run-time: -- n )
                   3352:   500 POSTPONE literal ; immediate
1.29      crook    3353: 
1.60      anton    3354: : flip [FOO] ;
1.48      anton    3355: flip .
                   3356: see flip
                   3357: @end example
1.29      crook    3358: 
1.48      anton    3359: @code{LITERAL} consumes a number at compile-time (when it's compilation
                   3360: semantics are executed) and pushes it at run-time (when the code it
                   3361: compiled is executed).  A frequent use of @code{LITERAL} is to compile a
                   3362: number computed at compile time into the current word:
1.29      crook    3363: 
1.48      anton    3364: @example
                   3365: : bar ( -- n )
                   3366:   [ 2 2 + ] literal ;
                   3367: see bar
                   3368: @end example
1.29      crook    3369: 
1.48      anton    3370: @assignment
                   3371: Write @code{]L} which allows writing the example above as @code{: bar (
                   3372: -- n ) [ 2 2 + ]L ;}
                   3373: @endassignment
                   3374: 
1.66      anton    3375: @c !! @xref{Macros} for reference
                   3376: 
1.48      anton    3377: 
                   3378: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
                   3379: @section Advanced macros
1.66      anton    3380: @cindex macros, advanced tutorial
                   3381: @cindex run-time code generation, tutorial
1.48      anton    3382: 
1.66      anton    3383: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
                   3384: Execution Tokens}.  It frequently performs @code{execute}, a relatively
                   3385: expensive operation in some Forth implementations.  You can use
1.48      anton    3386: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
                   3387: and produce a word that contains the word to be performed directly:
                   3388: 
                   3389: @c use ]] ... [[
                   3390: @example
                   3391: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
                   3392: \ at run-time, execute xt ( ... x -- ... ) for each element of the
                   3393: \ array beginning at addr and containing u elements
                   3394:   @{ xt @}
                   3395:   POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
1.50      anton    3396:     POSTPONE i POSTPONE @@ xt compile,
1.48      anton    3397:   1 cells POSTPONE literal POSTPONE +loop ;
                   3398: 
                   3399: : sum-array ( addr u -- n )
                   3400:  0 rot rot [ ' + compile-map-array ] ;
                   3401: see sum-array
                   3402: a 5 sum-array .
                   3403: @end example
                   3404: 
                   3405: You can use the full power of Forth for generating the code; here's an
                   3406: example where the code is generated in a loop:
                   3407: 
                   3408: @example
                   3409: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
                   3410: \ n2=n1+(addr1)*n, addr2=addr1+cell
1.50      anton    3411:   POSTPONE tuck POSTPONE @@
1.48      anton    3412:   POSTPONE literal POSTPONE * POSTPONE +
                   3413:   POSTPONE swap POSTPONE cell+ ;
                   3414: 
                   3415: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
1.51      pazsan   3416: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
1.48      anton    3417:   0 postpone literal postpone swap
                   3418:   [ ' compile-vmul-step compile-map-array ]
                   3419:   postpone drop ;
                   3420: see compile-vmul
                   3421: 
                   3422: : a-vmul ( addr -- n )
1.51      pazsan   3423: \ n=a*v, where v is a vector that's as long as a and starts at addr
1.48      anton    3424:  [ a 5 compile-vmul ] ;
                   3425: see a-vmul
                   3426: a a-vmul .
                   3427: @end example
                   3428: 
                   3429: This example uses @code{compile-map-array} to show off, but you could
1.66      anton    3430: also use @code{map-array} instead (try it now!).
1.48      anton    3431: 
                   3432: You can use this technique for efficient multiplication of large
                   3433: matrices.  In matrix multiplication, you multiply every line of one
                   3434: matrix with every column of the other matrix.  You can generate the code
                   3435: for one line once, and use it for every column.  The only downside of
                   3436: this technique is that it is cumbersome to recover the memory consumed
                   3437: by the generated code when you are done (and in more complicated cases
                   3438: it is not possible portably).
                   3439: 
1.66      anton    3440: @c !! @xref{Macros} for reference
                   3441: 
                   3442: 
1.48      anton    3443: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
                   3444: @section Compilation Tokens
1.66      anton    3445: @cindex compilation tokens, tutorial
                   3446: @cindex CT, tutorial
1.48      anton    3447: 
                   3448: This section is Gforth-specific.  You can skip it.
                   3449: 
                   3450: @code{' word compile,} compiles the interpretation semantics.  For words
                   3451: with default compilation semantics this is the same as performing the
                   3452: compilation semantics.  To represent the compilation semantics of other
                   3453: words (e.g., words like @code{if} that have no interpretation
                   3454: semantics), Gforth has the concept of a compilation token (CT,
                   3455: consisting of two cells), and words @code{comp'} and @code{[comp']}.
                   3456: You can perform the compilation semantics represented by a CT with
                   3457: @code{execute}:
1.29      crook    3458: 
1.48      anton    3459: @example
                   3460: : foo2 ( n1 n2 -- n )
                   3461:    [ comp' + execute ] ;
                   3462: see foo
                   3463: @end example
1.29      crook    3464: 
1.48      anton    3465: You can compile the compilation semantics represented by a CT with
                   3466: @code{postpone,}:
1.30      anton    3467: 
1.48      anton    3468: @example
                   3469: : foo3 ( -- )
                   3470:   [ comp' + postpone, ] ;
                   3471: see foo3
                   3472: @end example
1.30      anton    3473: 
1.51      pazsan   3474: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
1.48      anton    3475: @code{comp'} is particularly useful for words that have no
                   3476: interpretation semantics:
1.29      crook    3477: 
1.30      anton    3478: @example
1.48      anton    3479: ' if
1.60      anton    3480: comp' if .s 2drop
1.30      anton    3481: @end example
                   3482: 
1.66      anton    3483: Reference: @ref{Tokens for Words}.
                   3484: 
1.29      crook    3485: 
1.48      anton    3486: @node Wordlists and Search Order Tutorial,  , Compilation Tokens Tutorial, Tutorial
                   3487: @section Wordlists and Search Order
1.66      anton    3488: @cindex wordlists tutorial
                   3489: @cindex search order, tutorial
1.48      anton    3490: 
                   3491: The dictionary is not just a memory area that allows you to allocate
                   3492: memory with @code{allot}, it also contains the Forth words, arranged in
                   3493: several wordlists.  When searching for a word in a wordlist,
                   3494: conceptually you start searching at the youngest and proceed towards
                   3495: older words (in reality most systems nowadays use hash-tables); i.e., if
                   3496: you define a word with the same name as an older word, the new word
                   3497: shadows the older word.
                   3498: 
                   3499: Which wordlists are searched in which order is determined by the search
                   3500: order.  You can display the search order with @code{order}.  It displays
                   3501: first the search order, starting with the wordlist searched first, then
                   3502: it displays the wordlist that will contain newly defined words.
1.21      crook    3503: 
1.48      anton    3504: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
1.21      crook    3505: 
1.48      anton    3506: @example
                   3507: wordlist constant mywords
                   3508: @end example
1.21      crook    3509: 
1.48      anton    3510: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
                   3511: defined words (the @emph{current} wordlist):
1.21      crook    3512: 
1.48      anton    3513: @example
                   3514: mywords set-current
                   3515: order
                   3516: @end example
1.26      crook    3517: 
1.48      anton    3518: Gforth does not display a name for the wordlist in @code{mywords}
                   3519: because this wordlist was created anonymously with @code{wordlist}.
1.21      crook    3520: 
1.48      anton    3521: You can get the current wordlist with @code{get-current ( -- wid)}.  If
                   3522: you want to put something into a specific wordlist without overall
                   3523: effect on the current wordlist, this typically looks like this:
1.21      crook    3524: 
1.48      anton    3525: @example
                   3526: get-current mywords set-current ( wid )
                   3527: create someword
                   3528: ( wid ) set-current
                   3529: @end example
1.21      crook    3530: 
1.48      anton    3531: You can write the search order with @code{set-order ( wid1 .. widn n --
                   3532: )} and read it with @code{get-order ( -- wid1 .. widn n )}.  The first
                   3533: searched wordlist is topmost.
1.21      crook    3534: 
1.48      anton    3535: @example
                   3536: get-order mywords swap 1+ set-order
                   3537: order
                   3538: @end example
1.21      crook    3539: 
1.48      anton    3540: Yes, the order of wordlists in the output of @code{order} is reversed
                   3541: from stack comments and the output of @code{.s} and thus unintuitive.
1.21      crook    3542: 
1.48      anton    3543: @assignment
                   3544: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
                   3545: wordlist to the search order.  Define @code{previous ( -- )}, which
                   3546: removes the first searched wordlist from the search order.  Experiment
                   3547: with boundary conditions (you will see some crashes or situations that
                   3548: are hard or impossible to leave).
                   3549: @endassignment
1.21      crook    3550: 
1.48      anton    3551: The search order is a powerful foundation for providing features similar
                   3552: to Modula-2 modules and C++ namespaces.  However, trying to modularize
                   3553: programs in this way has disadvantages for debugging and reuse/factoring
                   3554: that overcome the advantages in my experience (I don't do huge projects,
1.55      anton    3555: though).  These disadvantages are not so clear in other
1.82      anton    3556: languages/programming environments, because these languages are not so
1.48      anton    3557: strong in debugging and reuse.
1.21      crook    3558: 
1.66      anton    3559: @c !! example
                   3560: 
                   3561: Reference: @ref{Word Lists}.
1.21      crook    3562: 
1.29      crook    3563: @c ******************************************************************
1.48      anton    3564: @node Introduction, Words, Tutorial, Top
1.29      crook    3565: @comment node-name,     next,           previous, up
                   3566: @chapter An Introduction to ANS Forth
                   3567: @cindex Forth - an introduction
1.21      crook    3568: 
1.83      anton    3569: The difference of this chapter from the Tutorial (@pxref{Tutorial}) is
                   3570: that it is slower-paced in its examples, but uses them to dive deep into
                   3571: explaining Forth internals (not covered by the Tutorial).  Apart from
                   3572: that, this chapter covers far less material.  It is suitable for reading
                   3573: without using a computer.
                   3574: 
1.29      crook    3575: The primary purpose of this manual is to document Gforth. However, since
                   3576: Forth is not a widely-known language and there is a lack of up-to-date
                   3577: teaching material, it seems worthwhile to provide some introductory
1.49      anton    3578: material.  For other sources of Forth-related
                   3579: information, see @ref{Forth-related information}.
1.21      crook    3580: 
1.29      crook    3581: The examples in this section should work on any ANS Forth; the
                   3582: output shown was produced using Gforth. Each example attempts to
                   3583: reproduce the exact output that Gforth produces. If you try out the
                   3584: examples (and you should), what you should type is shown @kbd{like this}
                   3585: and Gforth's response is shown @code{like this}. The single exception is
1.30      anton    3586: that, where the example shows @key{RET} it means that you should
1.29      crook    3587: press the ``carriage return'' key. Unfortunately, some output formats for
                   3588: this manual cannot show the difference between @kbd{this} and
                   3589: @code{this} which will make trying out the examples harder (but not
                   3590: impossible).
1.21      crook    3591: 
1.29      crook    3592: Forth is an unusual language. It provides an interactive development
                   3593: environment which includes both an interpreter and compiler. Forth
                   3594: programming style encourages you to break a problem down into many
                   3595: @cindex factoring
                   3596: small fragments (@dfn{factoring}), and then to develop and test each
                   3597: fragment interactively. Forth advocates assert that breaking the
                   3598: edit-compile-test cycle used by conventional programming languages can
                   3599: lead to great productivity improvements.
1.21      crook    3600: 
1.29      crook    3601: @menu
1.67      anton    3602: * Introducing the Text Interpreter::  
                   3603: * Stacks and Postfix notation::  
                   3604: * Your first definition::       
                   3605: * How does that work?::         
                   3606: * Forth is written in Forth::   
                   3607: * Review - elements of a Forth system::  
                   3608: * Where to go next::            
                   3609: * Exercises::                   
1.29      crook    3610: @end menu
1.21      crook    3611: 
1.29      crook    3612: @comment ----------------------------------------------
                   3613: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
                   3614: @section Introducing the Text Interpreter
                   3615: @cindex text interpreter
                   3616: @cindex outer interpreter
1.21      crook    3617: 
1.30      anton    3618: @c IMO this is too detailed and the pace is too slow for
                   3619: @c an introduction.  If you know German, take a look at
                   3620: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html 
                   3621: @c to see how I do it - anton 
                   3622: 
1.44      crook    3623: @c nac-> Where I have accepted your comments 100% and modified the text
                   3624: @c accordingly, I have deleted your comments. Elsewhere I have added a
                   3625: @c response like this to attempt to rationalise what I have done. Of
                   3626: @c course, this is a very clumsy mechanism for something that would be
                   3627: @c done far more efficiently over a beer. Please delete any dialogue
                   3628: @c you consider closed.
                   3629: 
1.29      crook    3630: When you invoke the Forth image, you will see a startup banner printed
                   3631: and nothing else (if you have Gforth installed on your system, try
1.30      anton    3632: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
1.29      crook    3633: its command line interpreter, which is called the @dfn{Text Interpreter}
                   3634: (also known as the @dfn{Outer Interpreter}).  (You will learn a lot
1.49      anton    3635: about the text interpreter as you read through this chapter, for more
                   3636: detail @pxref{The Text Interpreter}).
1.21      crook    3637: 
1.29      crook    3638: Although it's not obvious, Forth is actually waiting for your
1.30      anton    3639: input. Type a number and press the @key{RET} key:
1.21      crook    3640: 
1.26      crook    3641: @example
1.30      anton    3642: @kbd{45@key{RET}}  ok
1.26      crook    3643: @end example
1.21      crook    3644: 
1.29      crook    3645: Rather than give you a prompt to invite you to input something, the text
                   3646: interpreter prints a status message @i{after} it has processed a line
                   3647: of input. The status message in this case (``@code{ ok}'' followed by
                   3648: carriage-return) indicates that the text interpreter was able to process
                   3649: all of your input successfully. Now type something illegal:
                   3650: 
                   3651: @example
1.30      anton    3652: @kbd{qwer341@key{RET}}
1.29      crook    3653: :1: Undefined word
                   3654: qwer341
                   3655: ^^^^^^^
                   3656: $400D2BA8 Bounce
                   3657: $400DBDA8 no.extensions
                   3658: @end example
1.23      crook    3659: 
1.29      crook    3660: The exact text, other than the ``Undefined word'' may differ slightly on
                   3661: your system, but the effect is the same; when the text interpreter
                   3662: detects an error, it discards any remaining text on a line, resets
1.49      anton    3663: certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
                   3664: messages}.
1.23      crook    3665: 
1.29      crook    3666: The text interpreter waits for you to press carriage-return, and then
                   3667: processes your input line. Starting at the beginning of the line, it
                   3668: breaks the line into groups of characters separated by spaces. For each
                   3669: group of characters in turn, it makes two attempts to do something:
1.23      crook    3670: 
1.29      crook    3671: @itemize @bullet
                   3672: @item
1.44      crook    3673: @cindex name dictionary
1.29      crook    3674: It tries to treat it as a command. It does this by searching a @dfn{name
                   3675: dictionary}. If the group of characters matches an entry in the name
                   3676: dictionary, the name dictionary provides the text interpreter with
                   3677: information that allows the text interpreter perform some actions. In
                   3678: Forth jargon, we say that the group
                   3679: @cindex word
                   3680: @cindex definition
                   3681: @cindex execution token
                   3682: @cindex xt
                   3683: of characters names a @dfn{word}, that the dictionary search returns an
                   3684: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
                   3685: word, and that the text interpreter executes the xt. Often, the terms
                   3686: @dfn{word} and @dfn{definition} are used interchangeably.
                   3687: @item
                   3688: If the text interpreter fails to find a match in the name dictionary, it
                   3689: tries to treat the group of characters as a number in the current number
                   3690: base (when you start up Forth, the current number base is base 10). If
                   3691: the group of characters legitimately represents a number, the text
                   3692: interpreter pushes the number onto a stack (we'll learn more about that
                   3693: in the next section).
                   3694: @end itemize
1.23      crook    3695: 
1.29      crook    3696: If the text interpreter is unable to do either of these things with any
                   3697: group of characters, it discards the group of characters and the rest of
                   3698: the line, then prints an error message. If the text interpreter reaches
                   3699: the end of the line without error, it prints the status message ``@code{ ok}''
                   3700: followed by carriage-return.
1.21      crook    3701: 
1.29      crook    3702: This is the simplest command we can give to the text interpreter:
1.23      crook    3703: 
                   3704: @example
1.30      anton    3705: @key{RET}  ok
1.23      crook    3706: @end example
1.21      crook    3707: 
1.29      crook    3708: The text interpreter did everything we asked it to do (nothing) without
                   3709: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
                   3710: command:
1.21      crook    3711: 
1.23      crook    3712: @example
1.30      anton    3713: @kbd{12 dup fred dup@key{RET}}
1.29      crook    3714: :1: Undefined word
                   3715: 12 dup fred dup
                   3716:        ^^^^
                   3717: $400D2BA8 Bounce
                   3718: $400DBDA8 no.extensions
1.23      crook    3719: @end example
1.21      crook    3720: 
1.29      crook    3721: When you press the carriage-return key, the text interpreter starts to
                   3722: work its way along the line:
1.21      crook    3723: 
1.29      crook    3724: @itemize @bullet
                   3725: @item
                   3726: When it gets to the space after the @code{2}, it takes the group of
                   3727: characters @code{12} and looks them up in the name
                   3728: dictionary@footnote{We can't tell if it found them or not, but assume
                   3729: for now that it did not}. There is no match for this group of characters
                   3730: in the name dictionary, so it tries to treat them as a number. It is
                   3731: able to do this successfully, so it puts the number, 12, ``on the stack''
                   3732: (whatever that means).
                   3733: @item
                   3734: The text interpreter resumes scanning the line and gets the next group
                   3735: of characters, @code{dup}. It looks it up in the name dictionary and
                   3736: (you'll have to take my word for this) finds it, and executes the word
                   3737: @code{dup} (whatever that means).
                   3738: @item
                   3739: Once again, the text interpreter resumes scanning the line and gets the
                   3740: group of characters @code{fred}. It looks them up in the name
                   3741: dictionary, but can't find them. It tries to treat them as a number, but
                   3742: they don't represent any legal number.
                   3743: @end itemize
1.21      crook    3744: 
1.29      crook    3745: At this point, the text interpreter gives up and prints an error
                   3746: message. The error message shows exactly how far the text interpreter
                   3747: got in processing the line. In particular, it shows that the text
                   3748: interpreter made no attempt to do anything with the final character
                   3749: group, @code{dup}, even though we have good reason to believe that the
                   3750: text interpreter would have no problem looking that word up and
                   3751: executing it a second time.
1.21      crook    3752: 
                   3753: 
1.29      crook    3754: @comment ----------------------------------------------
                   3755: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
                   3756: @section Stacks, postfix notation and parameter passing
                   3757: @cindex text interpreter
                   3758: @cindex outer interpreter
1.21      crook    3759: 
1.29      crook    3760: In procedural programming languages (like C and Pascal), the
                   3761: building-block of programs is the @dfn{function} or @dfn{procedure}. These
                   3762: functions or procedures are called with @dfn{explicit parameters}. For
                   3763: example, in C we might write:
1.21      crook    3764: 
1.23      crook    3765: @example
1.29      crook    3766: total = total + new_volume(length,height,depth);
1.23      crook    3767: @end example
1.21      crook    3768: 
1.23      crook    3769: @noindent
1.29      crook    3770: where new_volume is a function-call to another piece of code, and total,
                   3771: length, height and depth are all variables. length, height and depth are
                   3772: parameters to the function-call.
1.21      crook    3773: 
1.29      crook    3774: In Forth, the equivalent of the function or procedure is the
                   3775: @dfn{definition} and parameters are implicitly passed between
                   3776: definitions using a shared stack that is visible to the
                   3777: programmer. Although Forth does support variables, the existence of the
                   3778: stack means that they are used far less often than in most other
                   3779: programming languages. When the text interpreter encounters a number, it
                   3780: will place (@dfn{push}) it on the stack. There are several stacks (the
1.30      anton    3781: actual number is implementation-dependent ...) and the particular stack
1.29      crook    3782: used for any operation is implied unambiguously by the operation being
                   3783: performed. The stack used for all integer operations is called the @dfn{data
                   3784: stack} and, since this is the stack used most commonly, references to
                   3785: ``the data stack'' are often abbreviated to ``the stack''.
1.21      crook    3786: 
1.29      crook    3787: The stacks have a last-in, first-out (LIFO) organisation. If you type:
1.21      crook    3788: 
1.23      crook    3789: @example
1.30      anton    3790: @kbd{1 2 3@key{RET}}  ok
1.23      crook    3791: @end example
1.21      crook    3792: 
1.29      crook    3793: Then this instructs the text interpreter to placed three numbers on the
                   3794: (data) stack. An analogy for the behaviour of the stack is to take a
                   3795: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
                   3796: the table. The 3 was the last card onto the pile (``last-in'') and if
                   3797: you take a card off the pile then, unless you're prepared to fiddle a
                   3798: bit, the card that you take off will be the 3 (``first-out''). The
                   3799: number that will be first-out of the stack is called the @dfn{top of
                   3800: stack}, which
                   3801: @cindex TOS definition
                   3802: is often abbreviated to @dfn{TOS}.
1.21      crook    3803: 
1.29      crook    3804: To understand how parameters are passed in Forth, consider the
                   3805: behaviour of the definition @code{+} (pronounced ``plus''). You will not
                   3806: be surprised to learn that this definition performs addition. More
                   3807: precisely, it adds two number together and produces a result. Where does
                   3808: it get the two numbers from? It takes the top two numbers off the
                   3809: stack. Where does it place the result? On the stack. You can act-out the
                   3810: behaviour of @code{+} with your playing cards like this:
1.21      crook    3811: 
                   3812: @itemize @bullet
                   3813: @item
1.29      crook    3814: Pick up two cards from the stack on the table
1.21      crook    3815: @item
1.29      crook    3816: Stare at them intently and ask yourself ``what @i{is} the sum of these two
                   3817: numbers''
1.21      crook    3818: @item
1.29      crook    3819: Decide that the answer is 5
1.21      crook    3820: @item
1.29      crook    3821: Shuffle the two cards back into the pack and find a 5
1.21      crook    3822: @item
1.29      crook    3823: Put a 5 on the remaining ace that's on the table.
1.21      crook    3824: @end itemize
                   3825: 
1.29      crook    3826: If you don't have a pack of cards handy but you do have Forth running,
                   3827: you can use the definition @code{.s} to show the current state of the stack,
                   3828: without affecting the stack. Type:
1.21      crook    3829: 
                   3830: @example
1.30      anton    3831: @kbd{clearstack 1 2 3@key{RET}} ok
                   3832: @kbd{.s@key{RET}} <3> 1 2 3  ok
1.23      crook    3833: @end example
                   3834: 
1.29      crook    3835: The text interpreter looks up the word @code{clearstack} and executes
                   3836: it; it tidies up the stack and removes any entries that may have been
                   3837: left on it by earlier examples. The text interpreter pushes each of the
                   3838: three numbers in turn onto the stack. Finally, the text interpreter
                   3839: looks up the word @code{.s} and executes it. The effect of executing
                   3840: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
                   3841: followed by a list of all the items on the stack; the item on the far
                   3842: right-hand side is the TOS.
1.21      crook    3843: 
1.29      crook    3844: You can now type:
1.21      crook    3845: 
1.29      crook    3846: @example
1.30      anton    3847: @kbd{+ .s@key{RET}} <2> 1 5  ok
1.29      crook    3848: @end example
1.21      crook    3849: 
1.29      crook    3850: @noindent
                   3851: which is correct; there are now 2 items on the stack and the result of
                   3852: the addition is 5.
1.23      crook    3853: 
1.29      crook    3854: If you're playing with cards, try doing a second addition: pick up the
                   3855: two cards, work out that their sum is 6, shuffle them into the pack,
                   3856: look for a 6 and place that on the table. You now have just one item on
                   3857: the stack. What happens if you try to do a third addition? Pick up the
                   3858: first card, pick up the second card -- ah! There is no second card. This
                   3859: is called a @dfn{stack underflow} and consitutes an error. If you try to
                   3860: do the same thing with Forth it will report an error (probably a Stack
                   3861: Underflow or an Invalid Memory Address error).
1.23      crook    3862: 
1.29      crook    3863: The opposite situation to a stack underflow is a @dfn{stack overflow},
                   3864: which simply accepts that there is a finite amount of storage space
                   3865: reserved for the stack. To stretch the playing card analogy, if you had
                   3866: enough packs of cards and you piled the cards up on the table, you would
                   3867: eventually be unable to add another card; you'd hit the ceiling. Gforth
                   3868: allows you to set the maximum size of the stacks. In general, the only
                   3869: time that you will get a stack overflow is because a definition has a
                   3870: bug in it and is generating data on the stack uncontrollably.
1.23      crook    3871: 
1.29      crook    3872: There's one final use for the playing card analogy. If you model your
                   3873: stack using a pack of playing cards, the maximum number of items on
                   3874: your stack will be 52 (I assume you didn't use the Joker). The maximum
                   3875: @i{value} of any item on the stack is 13 (the King). In fact, the only
                   3876: possible numbers are positive integer numbers 1 through 13; you can't
                   3877: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
                   3878: think about some of the cards, you can accommodate different
                   3879: numbers. For example, you could think of the Jack as representing 0,
                   3880: the Queen as representing -1 and the King as representing -2. Your
1.45      crook    3881: @i{range} remains unchanged (you can still only represent a total of 13
1.29      crook    3882: numbers) but the numbers that you can represent are -2 through 10.
1.28      crook    3883: 
1.29      crook    3884: In that analogy, the limit was the amount of information that a single
                   3885: stack entry could hold, and Forth has a similar limit. In Forth, the
                   3886: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
                   3887: implementation dependent and affects the maximum value that a stack
                   3888: entry can hold. A Standard Forth provides a cell size of at least
                   3889: 16-bits, and most desktop systems use a cell size of 32-bits.
1.21      crook    3890: 
1.29      crook    3891: Forth does not do any type checking for you, so you are free to
                   3892: manipulate and combine stack items in any way you wish. A convenient way
                   3893: of treating stack items is as 2's complement signed integers, and that
                   3894: is what Standard words like @code{+} do. Therefore you can type:
1.21      crook    3895: 
1.29      crook    3896: @example
1.30      anton    3897: @kbd{-5 12 + .s@key{RET}} <1> 7  ok
1.29      crook    3898: @end example
1.21      crook    3899: 
1.29      crook    3900: If you use numbers and definitions like @code{+} in order to turn Forth
                   3901: into a great big pocket calculator, you will realise that it's rather
                   3902: different from a normal calculator. Rather than typing 2 + 3 = you had
                   3903: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
                   3904: result). The terminology used to describe this difference is to say that
                   3905: your calculator uses @dfn{Infix Notation} (parameters and operators are
                   3906: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
                   3907: operators are separate), also called @dfn{Reverse Polish Notation}.
1.21      crook    3908: 
1.29      crook    3909: Whilst postfix notation might look confusing to begin with, it has
                   3910: several important advantages:
1.21      crook    3911: 
1.23      crook    3912: @itemize @bullet
                   3913: @item
1.29      crook    3914: it is unambiguous
1.23      crook    3915: @item
1.29      crook    3916: it is more concise
1.23      crook    3917: @item
1.29      crook    3918: it fits naturally with a stack-based system
1.23      crook    3919: @end itemize
1.21      crook    3920: 
1.29      crook    3921: To examine these claims in more detail, consider these sums:
1.21      crook    3922: 
1.29      crook    3923: @example
                   3924: 6 + 5 * 4 =
                   3925: 4 * 5 + 6 =
                   3926: @end example
1.21      crook    3927: 
1.29      crook    3928: If you're just learning maths or your maths is very rusty, you will
                   3929: probably come up with the answer 44 for the first and 26 for the
                   3930: second. If you are a bit of a whizz at maths you will remember the
                   3931: @i{convention} that multiplication takes precendence over addition, and
                   3932: you'd come up with the answer 26 both times. To explain the answer 26
                   3933: to someone who got the answer 44, you'd probably rewrite the first sum
                   3934: like this:
1.21      crook    3935: 
1.29      crook    3936: @example
                   3937: 6 + (5 * 4) =
                   3938: @end example
1.21      crook    3939: 
1.29      crook    3940: If what you really wanted was to perform the addition before the
                   3941: multiplication, you would have to use parentheses to force it.
1.21      crook    3942: 
1.29      crook    3943: If you did the first two sums on a pocket calculator you would probably
                   3944: get the right answers, unless you were very cautious and entered them using
                   3945: these keystroke sequences:
1.21      crook    3946: 
1.29      crook    3947: 6 + 5 = * 4 =
                   3948: 4 * 5 = + 6 =
1.21      crook    3949: 
1.29      crook    3950: Postfix notation is unambiguous because the order that the operators
                   3951: are applied is always explicit; that also means that parentheses are
                   3952: never required. The operators are @i{active} (the act of quoting the
                   3953: operator makes the operation occur) which removes the need for ``=''.
1.28      crook    3954: 
1.29      crook    3955: The sum 6 + 5 * 4 can be written (in postfix notation) in two
                   3956: equivalent ways:
1.26      crook    3957: 
                   3958: @example
1.29      crook    3959: 6 5 4 * +      or:
                   3960: 5 4 * 6 +
1.26      crook    3961: @end example
1.23      crook    3962: 
1.29      crook    3963: An important thing that you should notice about this notation is that
                   3964: the @i{order} of the numbers does not change; if you want to subtract
                   3965: 2 from 10 you type @code{10 2 -}.
1.1       anton    3966: 
1.29      crook    3967: The reason that Forth uses postfix notation is very simple to explain: it
                   3968: makes the implementation extremely simple, and it follows naturally from
                   3969: using the stack as a mechanism for passing parameters. Another way of
                   3970: thinking about this is to realise that all Forth definitions are
                   3971: @i{active}; they execute as they are encountered by the text
                   3972: interpreter. The result of this is that the syntax of Forth is trivially
                   3973: simple.
1.1       anton    3974: 
                   3975: 
                   3976: 
1.29      crook    3977: @comment ----------------------------------------------
                   3978: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
                   3979: @section Your first Forth definition
                   3980: @cindex first definition
1.1       anton    3981: 
1.29      crook    3982: Until now, the examples we've seen have been trivial; we've just been
                   3983: using Forth as a bigger-than-pocket calculator. Also, each calculation
                   3984: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
                   3985: again@footnote{That's not quite true. If you press the up-arrow key on
                   3986: your keyboard you should be able to scroll back to any earlier command,
                   3987: edit it and re-enter it.} In this section we'll see how to add new
                   3988: words to Forth's vocabulary.
1.1       anton    3989: 
1.29      crook    3990: The easiest way to create a new word is to use a @dfn{colon
                   3991: definition}. We'll define a few and try them out before worrying too
                   3992: much about how they work. Try typing in these examples; be careful to
                   3993: copy the spaces accurately:
1.1       anton    3994: 
1.29      crook    3995: @example
                   3996: : add-two 2 + . ;
                   3997: : greet ." Hello and welcome" ;
                   3998: : demo 5 add-two ;
                   3999: @end example
1.1       anton    4000: 
1.29      crook    4001: @noindent
                   4002: Now try them out:
1.1       anton    4003: 
1.29      crook    4004: @example
1.30      anton    4005: @kbd{greet@key{RET}} Hello and welcome  ok
                   4006: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome  ok
                   4007: @kbd{4 add-two@key{RET}} 6  ok
                   4008: @kbd{demo@key{RET}} 7  ok
                   4009: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11  ok
1.29      crook    4010: @end example
1.1       anton    4011: 
1.29      crook    4012: The first new thing that we've introduced here is the pair of words
                   4013: @code{:} and @code{;}. These are used to start and terminate a new
                   4014: definition, respectively. The first word after the @code{:} is the name
                   4015: for the new definition.
1.1       anton    4016: 
1.29      crook    4017: As you can see from the examples, a definition is built up of words that
                   4018: have already been defined; Forth makes no distinction between
                   4019: definitions that existed when you started the system up, and those that
                   4020: you define yourself.
1.1       anton    4021: 
1.29      crook    4022: The examples also introduce the words @code{.} (dot), @code{."}
                   4023: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
                   4024: the stack and displays it. It's like @code{.s} except that it only
                   4025: displays the top item of the stack and it is destructive; after it has
                   4026: executed, the number is no longer on the stack. There is always one
                   4027: space printed after the number, and no spaces before it. Dot-quote
                   4028: defines a string (a sequence of characters) that will be printed when
                   4029: the word is executed. The string can contain any printable characters
                   4030: except @code{"}. A @code{"} has a special function; it is not a Forth
                   4031: word but it acts as a delimiter (the way that delimiters work is
                   4032: described in the next section). Finally, @code{dup} duplicates the value
                   4033: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
1.1       anton    4034: 
1.29      crook    4035: We already know that the text interpreter searches through the
                   4036: dictionary to locate names. If you've followed the examples earlier, you
                   4037: will already have a definition called @code{add-two}. Lets try modifying
                   4038: it by typing in a new definition:
1.1       anton    4039: 
1.29      crook    4040: @example
1.30      anton    4041: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two  ok
1.29      crook    4042: @end example
1.5       anton    4043: 
1.29      crook    4044: Forth recognised that we were defining a word that already exists, and
                   4045: printed a message to warn us of that fact. Let's try out the new
                   4046: definition:
1.5       anton    4047: 
1.29      crook    4048: @example
1.30      anton    4049: @kbd{9 add-two@key{RET}} 9 + 2 =11  ok
1.29      crook    4050: @end example
1.1       anton    4051: 
1.29      crook    4052: @noindent
                   4053: All that we've actually done here, though, is to create a new
                   4054: definition, with a particular name. The fact that there was already a
                   4055: definition with the same name did not make any difference to the way
                   4056: that the new definition was created (except that Forth printed a warning
                   4057: message). The old definition of add-two still exists (try @code{demo}
                   4058: again to see that this is true). Any new definition will use the new
                   4059: definition of @code{add-two}, but old definitions continue to use the
                   4060: version that already existed at the time that they were @code{compiled}.
1.1       anton    4061: 
1.29      crook    4062: Before you go on to the next section, try defining and redefining some
                   4063: words of your own.
1.1       anton    4064: 
1.29      crook    4065: @comment ----------------------------------------------
                   4066: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
                   4067: @section How does that work?
                   4068: @cindex parsing words
1.1       anton    4069: 
1.30      anton    4070: @c That's pretty deep (IMO way too deep) for an introduction. - anton
                   4071: 
                   4072: @c Is it a good idea to talk about the interpretation semantics of a
                   4073: @c number? We don't have an xt to go along with it. - anton
                   4074: 
                   4075: @c Now that I have eliminated execution semantics, I wonder if it would not
                   4076: @c be better to keep them (or add run-time semantics), to make it easier to
                   4077: @c explain what compilation semantics usually does. - anton
                   4078: 
1.44      crook    4079: @c nac-> I removed the term ``default compilation sematics'' from the
                   4080: @c introductory chapter. Removing ``execution semantics'' was making
                   4081: @c everything simpler to explain, then I think the use of this term made
                   4082: @c everything more complex again. I replaced it with ``default
                   4083: @c semantics'' (which is used elsewhere in the manual) by which I mean
                   4084: @c ``a definition that has neither the immediate nor the compile-only
1.83      anton    4085: @c flag set''.
                   4086: 
                   4087: @c anton: I have eliminated default semantics (except in one place where it
                   4088: @c means "default interpretation and compilation semantics"), because it
                   4089: @c makes no sense in the presence of combined words.  I reverted to
                   4090: @c "execution semantics" where necessary.
                   4091: 
                   4092: @c nac-> I reworded big chunks of the ``how does that work''
1.44      crook    4093: @c section (and, unusually for me, I think I even made it shorter!).  See
                   4094: @c what you think -- I know I have not addressed your primary concern
                   4095: @c that it is too heavy-going for an introduction. From what I understood
                   4096: @c of your course notes it looks as though they might be a good framework. 
                   4097: @c Things that I've tried to capture here are some things that came as a
                   4098: @c great revelation here when I first understood them. Also, I like the
                   4099: @c fact that a very simple code example shows up almost all of the issues
                   4100: @c that you need to understand to see how Forth works. That's unique and
                   4101: @c worthwhile to emphasise.
                   4102: 
1.83      anton    4103: @c anton: I think it's a good idea to present the details, especially those
                   4104: @c that you found to be a revelation, and probably the tutorial tries to be
                   4105: @c too superficial and does not get some of the things across that make
                   4106: @c Forth special.  I do believe that most of the time these things should
                   4107: @c be discussed at the end of a section or in separate sections instead of
                   4108: @c in the middle of a section (e.g., the stuff you added in "User-defined
                   4109: @c defining words" leads in a completely different direction from the rest
                   4110: @c of the section).
                   4111: 
1.29      crook    4112: Now we're going to take another look at the definition of @code{add-two}
                   4113: from the previous section. From our knowledge of the way that the text
                   4114: interpreter works, we would have expected this result when we tried to
                   4115: define @code{add-two}:
1.21      crook    4116: 
1.29      crook    4117: @example
1.44      crook    4118: @kbd{: add-two 2 + . ;@key{RET}}
1.29      crook    4119:   ^^^^^^^
                   4120: Error: Undefined word
                   4121: @end example
1.28      crook    4122: 
1.29      crook    4123: The reason that this didn't happen is bound up in the way that @code{:}
                   4124: works. The word @code{:} does two special things. The first special
                   4125: thing that it does prevents the text interpreter from ever seeing the
                   4126: characters @code{add-two}. The text interpreter uses a variable called
                   4127: @cindex modifying >IN
1.44      crook    4128: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
1.29      crook    4129: input line. When it encounters the word @code{:} it behaves in exactly
                   4130: the same way as it does for any other word; it looks it up in the name
                   4131: dictionary, finds its xt and executes it. When @code{:} executes, it
                   4132: looks at the input buffer, finds the word @code{add-two} and advances the
                   4133: value of @code{>IN} to point past it. It then does some other stuff
                   4134: associated with creating the new definition (including creating an entry
                   4135: for @code{add-two} in the name dictionary). When the execution of @code{:}
                   4136: completes, control returns to the text interpreter, which is oblivious
                   4137: to the fact that it has been tricked into ignoring part of the input
                   4138: line.
1.21      crook    4139: 
1.29      crook    4140: @cindex parsing words
                   4141: Words like @code{:} -- words that advance the value of @code{>IN} and so
                   4142: prevent the text interpreter from acting on the whole of the input line
                   4143: -- are called @dfn{parsing words}.
1.21      crook    4144: 
1.29      crook    4145: @cindex @code{state} - effect on the text interpreter
                   4146: @cindex text interpreter - effect of state
                   4147: The second special thing that @code{:} does is change the value of a
                   4148: variable called @code{state}, which affects the way that the text
                   4149: interpreter behaves. When Gforth starts up, @code{state} has the value
                   4150: 0, and the text interpreter is said to be @dfn{interpreting}. During a
                   4151: colon definition (started with @code{:}), @code{state} is set to -1 and
1.44      crook    4152: the text interpreter is said to be @dfn{compiling}.
                   4153: 
                   4154: In this example, the text interpreter is compiling when it processes the
                   4155: string ``@code{2 + . ;}''. It still breaks the string down into
                   4156: character sequences in the same way. However, instead of pushing the
                   4157: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
                   4158: into the definition of @code{add-two} that will make the number @code{2} get
                   4159: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
                   4160: the behaviours of @code{+} and @code{.} are also compiled into the
                   4161: definition.
                   4162: 
                   4163: One category of words don't get compiled. These so-called @dfn{immediate
                   4164: words} get executed (performed @i{now}) regardless of whether the text
                   4165: interpreter is interpreting or compiling. The word @code{;} is an
                   4166: immediate word. Rather than being compiled into the definition, it
                   4167: executes. Its effect is to terminate the current definition, which
                   4168: includes changing the value of @code{state} back to 0.
                   4169: 
                   4170: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
                   4171: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
                   4172: definition.
1.28      crook    4173: 
1.30      anton    4174: In Forth, every word or number can be described in terms of two
1.29      crook    4175: properties:
1.28      crook    4176: 
                   4177: @itemize @bullet
                   4178: @item
1.29      crook    4179: @cindex interpretation semantics
1.44      crook    4180: Its @dfn{interpretation semantics} describe how it will behave when the
                   4181: text interpreter encounters it in @dfn{interpret} state. The
                   4182: interpretation semantics of a word are represented by an @dfn{execution
                   4183: token}.
1.28      crook    4184: @item
1.29      crook    4185: @cindex compilation semantics
1.44      crook    4186: Its @dfn{compilation semantics} describe how it will behave when the
                   4187: text interpreter encounters it in @dfn{compile} state. The compilation
                   4188: semantics of a word are represented in an implementation-dependent way;
                   4189: Gforth uses a @dfn{compilation token}.
1.29      crook    4190: @end itemize
                   4191: 
                   4192: @noindent
                   4193: Numbers are always treated in a fixed way:
                   4194: 
                   4195: @itemize @bullet
1.28      crook    4196: @item
1.44      crook    4197: When the number is @dfn{interpreted}, its behaviour is to push the
                   4198: number onto the stack.
1.28      crook    4199: @item
1.30      anton    4200: When the number is @dfn{compiled}, a piece of code is appended to the
                   4201: current definition that pushes the number when it runs. (In other words,
                   4202: the compilation semantics of a number are to postpone its interpretation
                   4203: semantics until the run-time of the definition that it is being compiled
                   4204: into.)
1.29      crook    4205: @end itemize
                   4206: 
1.44      crook    4207: Words don't behave in such a regular way, but most have @i{default
                   4208: semantics} which means that they behave like this:
1.29      crook    4209: 
                   4210: @itemize @bullet
1.28      crook    4211: @item
1.30      anton    4212: The @dfn{interpretation semantics} of the word are to do something useful.
                   4213: @item
1.29      crook    4214: The @dfn{compilation semantics} of the word are to append its
1.30      anton    4215: @dfn{interpretation semantics} to the current definition (so that its
                   4216: run-time behaviour is to do something useful).
1.28      crook    4217: @end itemize
                   4218: 
1.30      anton    4219: @cindex immediate words
1.44      crook    4220: The actual behaviour of any particular word can be controlled by using
                   4221: the words @code{immediate} and @code{compile-only} when the word is
                   4222: defined. These words set flags in the name dictionary entry of the most
                   4223: recently defined word, and these flags are retrieved by the text
                   4224: interpreter when it finds the word in the name dictionary.
                   4225: 
                   4226: A word that is marked as @dfn{immediate} has compilation semantics that
                   4227: are identical to its interpretation semantics. In other words, it
                   4228: behaves like this:
1.29      crook    4229: 
                   4230: @itemize @bullet
                   4231: @item
1.30      anton    4232: The @dfn{interpretation semantics} of the word are to do something useful.
1.29      crook    4233: @item
1.30      anton    4234: The @dfn{compilation semantics} of the word are to do something useful
                   4235: (and actually the same thing); i.e., it is executed during compilation.
1.29      crook    4236: @end itemize
1.28      crook    4237: 
1.44      crook    4238: Marking a word as @dfn{compile-only} prohibits the text interpreter from
                   4239: performing the interpretation semantics of the word directly; an attempt
                   4240: to do so will generate an error. It is never necessary to use
                   4241: @code{compile-only} (and it is not even part of ANS Forth, though it is
                   4242: provided by many implementations) but it is good etiquette to apply it
                   4243: to a word that will not behave correctly (and might have unexpected
                   4244: side-effects) in interpret state. For example, it is only legal to use
                   4245: the conditional word @code{IF} within a definition. If you forget this
                   4246: and try to use it elsewhere, the fact that (in Gforth) it is marked as
                   4247: @code{compile-only} allows the text interpreter to generate a helpful
                   4248: error message rather than subjecting you to the consequences of your
                   4249: folly.
                   4250: 
1.29      crook    4251: This example shows the difference between an immediate and a
                   4252: non-immediate word:
1.28      crook    4253: 
1.29      crook    4254: @example
                   4255: : show-state state @@ . ;
                   4256: : show-state-now show-state ; immediate
                   4257: : word1 show-state ;
                   4258: : word2 show-state-now ;
1.28      crook    4259: @end example
1.23      crook    4260: 
1.29      crook    4261: The word @code{immediate} after the definition of @code{show-state-now}
                   4262: makes that word an immediate word. These definitions introduce a new
                   4263: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
                   4264: variable, and leaves it on the stack. Therefore, the behaviour of
                   4265: @code{show-state} is to print a number that represents the current value
                   4266: of @code{state}.
1.28      crook    4267: 
1.29      crook    4268: When you execute @code{word1}, it prints the number 0, indicating that
                   4269: the system is interpreting. When the text interpreter compiled the
                   4270: definition of @code{word1}, it encountered @code{show-state} whose
1.30      anton    4271: compilation semantics are to append its interpretation semantics to the
1.29      crook    4272: current definition. When you execute @code{word1}, it performs the
1.30      anton    4273: interpretation semantics of @code{show-state}.  At the time that @code{word1}
1.29      crook    4274: (and therefore @code{show-state}) are executed, the system is
                   4275: interpreting.
1.28      crook    4276: 
1.30      anton    4277: When you pressed @key{RET} after entering the definition of @code{word2},
1.29      crook    4278: you should have seen the number -1 printed, followed by ``@code{
                   4279: ok}''. When the text interpreter compiled the definition of
                   4280: @code{word2}, it encountered @code{show-state-now}, an immediate word,
1.30      anton    4281: whose compilation semantics are therefore to perform its interpretation
1.29      crook    4282: semantics. It is executed straight away (even before the text
                   4283: interpreter has moved on to process another group of characters; the
                   4284: @code{;} in this example). The effect of executing it are to display the
                   4285: value of @code{state} @i{at the time that the definition of}
                   4286: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
                   4287: system is compiling at this time. If you execute @code{word2} it does
                   4288: nothing at all.
1.28      crook    4289: 
1.29      crook    4290: @cindex @code{."}, how it works
                   4291: Before leaving the subject of immediate words, consider the behaviour of
                   4292: @code{."} in the definition of @code{greet}, in the previous
                   4293: section. This word is both a parsing word and an immediate word. Notice
                   4294: that there is a space between @code{."} and the start of the text
                   4295: @code{Hello and welcome}, but that there is no space between the last
                   4296: letter of @code{welcome} and the @code{"} character. The reason for this
                   4297: is that @code{."} is a Forth word; it must have a space after it so that
                   4298: the text interpreter can identify it. The @code{"} is not a Forth word;
                   4299: it is a @dfn{delimiter}. The examples earlier show that, when the string
                   4300: is displayed, there is neither a space before the @code{H} nor after the
                   4301: @code{e}. Since @code{."} is an immediate word, it executes at the time
                   4302: that @code{greet} is defined. When it executes, its behaviour is to
                   4303: search forward in the input line looking for the delimiter. When it
                   4304: finds the delimiter, it updates @code{>IN} to point past the
                   4305: delimiter. It also compiles some magic code into the definition of
                   4306: @code{greet}; the xt of a run-time routine that prints a text string. It
                   4307: compiles the string @code{Hello and welcome} into memory so that it is
                   4308: available to be printed later. When the text interpreter gains control,
                   4309: the next word it finds in the input stream is @code{;} and so it
                   4310: terminates the definition of @code{greet}.
1.28      crook    4311: 
                   4312: 
                   4313: @comment ----------------------------------------------
1.29      crook    4314: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
                   4315: @section Forth is written in Forth
                   4316: @cindex structure of Forth programs
                   4317: 
                   4318: When you start up a Forth compiler, a large number of definitions
                   4319: already exist. In Forth, you develop a new application using bottom-up
                   4320: programming techniques to create new definitions that are defined in
                   4321: terms of existing definitions. As you create each definition you can
                   4322: test and debug it interactively.
                   4323: 
                   4324: If you have tried out the examples in this section, you will probably
                   4325: have typed them in by hand; when you leave Gforth, your definitions will
                   4326: be lost. You can avoid this by using a text editor to enter Forth source
                   4327: code into a file, and then loading code from the file using
1.49      anton    4328: @code{include} (@pxref{Forth source files}). A Forth source file is
1.29      crook    4329: processed by the text interpreter, just as though you had typed it in by
                   4330: hand@footnote{Actually, there are some subtle differences -- see
                   4331: @ref{The Text Interpreter}.}.
                   4332: 
                   4333: Gforth also supports the traditional Forth alternative to using text
1.49      anton    4334: files for program entry (@pxref{Blocks}).
1.28      crook    4335: 
1.29      crook    4336: In common with many, if not most, Forth compilers, most of Gforth is
                   4337: actually written in Forth. All of the @file{.fs} files in the
                   4338: installation directory@footnote{For example,
1.30      anton    4339: @file{/usr/local/share/gforth...}} are Forth source files, which you can
1.29      crook    4340: study to see examples of Forth programming.
1.28      crook    4341: 
1.29      crook    4342: Gforth maintains a history file that records every line that you type to
                   4343: the text interpreter. This file is preserved between sessions, and is
                   4344: used to provide a command-line recall facility. If you enter long
                   4345: definitions by hand, you can use a text editor to paste them out of the
                   4346: history file into a Forth source file for reuse at a later time
1.49      anton    4347: (for more information @pxref{Command-line editing}).
1.28      crook    4348: 
                   4349: 
                   4350: @comment ----------------------------------------------
1.29      crook    4351: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
                   4352: @section Review - elements of a Forth system
                   4353: @cindex elements of a Forth system
1.28      crook    4354: 
1.29      crook    4355: To summarise this chapter:
1.28      crook    4356: 
                   4357: @itemize @bullet
                   4358: @item
1.29      crook    4359: Forth programs use @dfn{factoring} to break a problem down into small
                   4360: fragments called @dfn{words} or @dfn{definitions}.
                   4361: @item
                   4362: Forth program development is an interactive process.
                   4363: @item
                   4364: The main command loop that accepts input, and controls both
                   4365: interpretation and compilation, is called the @dfn{text interpreter}
                   4366: (also known as the @dfn{outer interpreter}).
                   4367: @item
                   4368: Forth has a very simple syntax, consisting of words and numbers
                   4369: separated by spaces or carriage-return characters. Any additional syntax
                   4370: is imposed by @dfn{parsing words}.
                   4371: @item
                   4372: Forth uses a stack to pass parameters between words. As a result, it
                   4373: uses postfix notation.
                   4374: @item
                   4375: To use a word that has previously been defined, the text interpreter
                   4376: searches for the word in the @dfn{name dictionary}.
                   4377: @item
1.30      anton    4378: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
1.28      crook    4379: @item
1.29      crook    4380: The text interpreter uses the value of @code{state} to select between
                   4381: the use of the @dfn{interpretation semantics} and the  @dfn{compilation
                   4382: semantics} of a word that it encounters.
1.28      crook    4383: @item
1.30      anton    4384: The relationship between the @dfn{interpretation semantics} and
                   4385: @dfn{compilation semantics} for a word
1.29      crook    4386: depend upon the way in which the word was defined (for example, whether
                   4387: it is an @dfn{immediate} word).
1.28      crook    4388: @item
1.29      crook    4389: Forth definitions can be implemented in Forth (called @dfn{high-level
                   4390: definitions}) or in some other way (usually a lower-level language and
                   4391: as a result often called @dfn{low-level definitions}, @dfn{code
                   4392: definitions} or @dfn{primitives}).
1.28      crook    4393: @item
1.29      crook    4394: Many Forth systems are implemented mainly in Forth.
1.28      crook    4395: @end itemize
                   4396: 
                   4397: 
1.29      crook    4398: @comment ----------------------------------------------
1.48      anton    4399: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
1.29      crook    4400: @section Where To Go Next
                   4401: @cindex where to go next
1.28      crook    4402: 
1.29      crook    4403: Amazing as it may seem, if you have read (and understood) this far, you
                   4404: know almost all the fundamentals about the inner workings of a Forth
                   4405: system. You certainly know enough to be able to read and understand the
                   4406: rest of this manual and the ANS Forth document, to learn more about the
                   4407: facilities that Forth in general and Gforth in particular provide. Even
                   4408: scarier, you know almost enough to implement your own Forth system.
1.30      anton    4409: However, that's not a good idea just yet... better to try writing some
1.29      crook    4410: programs in Gforth.
1.28      crook    4411: 
1.29      crook    4412: Forth has such a rich vocabulary that it can be hard to know where to
                   4413: start in learning it. This section suggests a few sets of words that are
                   4414: enough to write small but useful programs. Use the word index in this
                   4415: document to learn more about each word, then try it out and try to write
                   4416: small definitions using it. Start by experimenting with these words:
1.28      crook    4417: 
                   4418: @itemize @bullet
                   4419: @item
1.29      crook    4420: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
                   4421: @item
                   4422: Comparison: @code{MIN MAX =}
                   4423: @item
                   4424: Logic: @code{AND OR XOR NOT}
                   4425: @item
                   4426: Stack manipulation: @code{DUP DROP SWAP OVER}
1.28      crook    4427: @item
1.29      crook    4428: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
1.28      crook    4429: @item
1.29      crook    4430: Input/Output: @code{. ." EMIT CR KEY}
1.28      crook    4431: @item
1.29      crook    4432: Defining words: @code{: ; CREATE}
1.28      crook    4433: @item
1.29      crook    4434: Memory allocation words: @code{ALLOT ,}
1.28      crook    4435: @item
1.29      crook    4436: Tools: @code{SEE WORDS .S MARKER}
                   4437: @end itemize
                   4438: 
                   4439: When you have mastered those, go on to:
                   4440: 
                   4441: @itemize @bullet
1.28      crook    4442: @item
1.29      crook    4443: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
1.28      crook    4444: @item
1.29      crook    4445: Memory access: @code{@@ !}
1.28      crook    4446: @end itemize
1.23      crook    4447: 
1.29      crook    4448: When you have mastered these, there's nothing for it but to read through
                   4449: the whole of this manual and find out what you've missed.
                   4450: 
                   4451: @comment ----------------------------------------------
1.48      anton    4452: @node Exercises,  , Where to go next, Introduction
1.29      crook    4453: @section Exercises
                   4454: @cindex exercises
                   4455: 
                   4456: TODO: provide a set of programming excercises linked into the stuff done
                   4457: already and into other sections of the manual. Provide solutions to all
                   4458: the exercises in a .fs file in the distribution.
                   4459: 
                   4460: @c Get some inspiration from Starting Forth and Kelly&Spies.
                   4461: 
                   4462: @c excercises:
                   4463: @c 1. take inches and convert to feet and inches.
                   4464: @c 2. take temperature and convert from fahrenheight to celcius;
                   4465: @c    may need to care about symmetric vs floored??
                   4466: @c 3. take input line and do character substitution
                   4467: @c    to encipher or decipher
                   4468: @c 4. as above but work on a file for in and out
                   4469: @c 5. take input line and convert to pig-latin 
                   4470: @c
                   4471: @c thing of sets of things to exercise then come up with
                   4472: @c problems that need those things.
                   4473: 
                   4474: 
1.26      crook    4475: @c ******************************************************************
1.29      crook    4476: @node Words, Error messages, Introduction, Top
1.1       anton    4477: @chapter Forth Words
1.26      crook    4478: @cindex words
1.1       anton    4479: 
                   4480: @menu
                   4481: * Notation::                    
1.65      anton    4482: * Case insensitivity::          
                   4483: * Comments::                    
                   4484: * Boolean Flags::               
1.1       anton    4485: * Arithmetic::                  
                   4486: * Stack Manipulation::          
1.5       anton    4487: * Memory::                      
1.1       anton    4488: * Control Structures::          
                   4489: * Defining Words::              
1.65      anton    4490: * Interpretation and Compilation Semantics::  
1.47      crook    4491: * Tokens for Words::            
1.81      anton    4492: * Compiling words::             
1.65      anton    4493: * The Text Interpreter::        
                   4494: * Word Lists::                  
                   4495: * Environmental Queries::       
1.12      anton    4496: * Files::                       
                   4497: * Blocks::                      
                   4498: * Other I/O::                   
1.78      anton    4499: * Locals::                      
                   4500: * Structures::                  
                   4501: * Object-oriented Forth::       
1.12      anton    4502: * Programming Tools::           
                   4503: * Assembler and Code Words::    
                   4504: * Threading Words::             
1.65      anton    4505: * Passing Commands to the OS::  
                   4506: * Keeping track of Time::       
                   4507: * Miscellaneous Words::         
1.1       anton    4508: @end menu
                   4509: 
1.65      anton    4510: @node Notation, Case insensitivity, Words, Words
1.1       anton    4511: @section Notation
                   4512: @cindex notation of glossary entries
                   4513: @cindex format of glossary entries
                   4514: @cindex glossary notation format
                   4515: @cindex word glossary entry format
                   4516: 
                   4517: The Forth words are described in this section in the glossary notation
1.67      anton    4518: that has become a de-facto standard for Forth texts:
1.1       anton    4519: 
                   4520: @format
1.29      crook    4521: @i{word}     @i{Stack effect}   @i{wordset}   @i{pronunciation}
1.1       anton    4522: @end format
1.29      crook    4523: @i{Description}
1.1       anton    4524: 
                   4525: @table @var
                   4526: @item word
1.28      crook    4527: The name of the word.
1.1       anton    4528: 
                   4529: @item Stack effect
                   4530: @cindex stack effect
1.29      crook    4531: The stack effect is written in the notation @code{@i{before} --
                   4532: @i{after}}, where @i{before} and @i{after} describe the top of
1.1       anton    4533: stack entries before and after the execution of the word. The rest of
                   4534: the stack is not touched by the word. The top of stack is rightmost,
                   4535: i.e., a stack sequence is written as it is typed in. Note that Gforth
                   4536: uses a separate floating point stack, but a unified stack
1.29      crook    4537: notation. Also, return stack effects are not shown in @i{stack
                   4538: effect}, but in @i{Description}. The name of a stack item describes
1.1       anton    4539: the type and/or the function of the item. See below for a discussion of
                   4540: the types.
                   4541: 
                   4542: All words have two stack effects: A compile-time stack effect and a
                   4543: run-time stack effect. The compile-time stack-effect of most words is
1.29      crook    4544: @i{ -- }. If the compile-time stack-effect of a word deviates from
1.1       anton    4545: this standard behaviour, or the word does other unusual things at
                   4546: compile time, both stack effects are shown; otherwise only the run-time
                   4547: stack effect is shown.
                   4548: 
                   4549: @cindex pronounciation of words
                   4550: @item pronunciation
                   4551: How the word is pronounced.
                   4552: 
                   4553: @cindex wordset
1.67      anton    4554: @cindex environment wordset
1.1       anton    4555: @item wordset
1.21      crook    4556: The ANS Forth standard is divided into several word sets. A standard
                   4557: system need not support all of them. Therefore, in theory, the fewer
                   4558: word sets your program uses the more portable it will be. However, we
                   4559: suspect that most ANS Forth systems on personal machines will feature
1.26      crook    4560: all word sets. Words that are not defined in ANS Forth have
1.21      crook    4561: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
1.1       anton    4562: describes words that will work in future releases of Gforth;
                   4563: @code{gforth-internal} words are more volatile. Environmental query
                   4564: strings are also displayed like words; you can recognize them by the
1.21      crook    4565: @code{environment} in the word set field.
1.1       anton    4566: 
                   4567: @item Description
                   4568: A description of the behaviour of the word.
                   4569: @end table
                   4570: 
                   4571: @cindex types of stack items
                   4572: @cindex stack item types
                   4573: The type of a stack item is specified by the character(s) the name
                   4574: starts with:
                   4575: 
                   4576: @table @code
                   4577: @item f
                   4578: @cindex @code{f}, stack item type
                   4579: Boolean flags, i.e. @code{false} or @code{true}.
                   4580: @item c
                   4581: @cindex @code{c}, stack item type
                   4582: Char
                   4583: @item w
                   4584: @cindex @code{w}, stack item type
                   4585: Cell, can contain an integer or an address
                   4586: @item n
                   4587: @cindex @code{n}, stack item type
                   4588: signed integer
                   4589: @item u
                   4590: @cindex @code{u}, stack item type
                   4591: unsigned integer
                   4592: @item d
                   4593: @cindex @code{d}, stack item type
                   4594: double sized signed integer
                   4595: @item ud
                   4596: @cindex @code{ud}, stack item type
                   4597: double sized unsigned integer
                   4598: @item r
                   4599: @cindex @code{r}, stack item type
                   4600: Float (on the FP stack)
1.21      crook    4601: @item a-
1.1       anton    4602: @cindex @code{a_}, stack item type
                   4603: Cell-aligned address
1.21      crook    4604: @item c-
1.1       anton    4605: @cindex @code{c_}, stack item type
                   4606: Char-aligned address (note that a Char may have two bytes in Windows NT)
1.21      crook    4607: @item f-
1.1       anton    4608: @cindex @code{f_}, stack item type
                   4609: Float-aligned address
1.21      crook    4610: @item df-
1.1       anton    4611: @cindex @code{df_}, stack item type
                   4612: Address aligned for IEEE double precision float
1.21      crook    4613: @item sf-
1.1       anton    4614: @cindex @code{sf_}, stack item type
                   4615: Address aligned for IEEE single precision float
                   4616: @item xt
                   4617: @cindex @code{xt}, stack item type
                   4618: Execution token, same size as Cell
                   4619: @item wid
                   4620: @cindex @code{wid}, stack item type
1.21      crook    4621: Word list ID, same size as Cell
1.68      anton    4622: @item ior, wior
                   4623: @cindex ior type description
                   4624: @cindex wior type description
                   4625: I/O result code, cell-sized.  In Gforth, you can @code{throw} iors.
1.1       anton    4626: @item f83name
                   4627: @cindex @code{f83name}, stack item type
                   4628: Pointer to a name structure
                   4629: @item "
                   4630: @cindex @code{"}, stack item type
1.12      anton    4631: string in the input stream (not on the stack). The terminating character
                   4632: is a blank by default. If it is not a blank, it is shown in @code{<>}
1.1       anton    4633: quotes.
                   4634: @end table
                   4635: 
1.65      anton    4636: @comment ----------------------------------------------
                   4637: @node Case insensitivity, Comments, Notation, Words
                   4638: @section Case insensitivity
                   4639: @cindex case sensitivity
                   4640: @cindex upper and lower case
                   4641: 
                   4642: Gforth is case-insensitive; you can enter definitions and invoke
                   4643: Standard words using upper, lower or mixed case (however,
                   4644: @pxref{core-idef, Implementation-defined options, Implementation-defined
                   4645: options}).
                   4646: 
                   4647: ANS Forth only @i{requires} implementations to recognise Standard words
                   4648: when they are typed entirely in upper case. Therefore, a Standard
                   4649: program must use upper case for all Standard words. You can use whatever
                   4650: case you like for words that you define, but in a Standard program you
                   4651: have to use the words in the same case that you defined them.
                   4652: 
                   4653: Gforth supports case sensitivity through @code{table}s (case-sensitive
                   4654: wordlists, @pxref{Word Lists}).
                   4655: 
                   4656: Two people have asked how to convert Gforth to be case-sensitive; while
                   4657: we think this is a bad idea, you can change all wordlists into tables
                   4658: like this:
                   4659: 
                   4660: @example
                   4661: ' table-find forth-wordlist wordlist-map @ !
                   4662: @end example
                   4663: 
                   4664: Note that you now have to type the predefined words in the same case
                   4665: that we defined them, which are varying.  You may want to convert them
                   4666: to your favourite case before doing this operation (I won't explain how,
                   4667: because if you are even contemplating doing this, you'd better have
                   4668: enough knowledge of Forth systems to know this already).
                   4669: 
                   4670: @node Comments, Boolean Flags, Case insensitivity, Words
1.21      crook    4671: @section Comments
1.26      crook    4672: @cindex comments
1.21      crook    4673: 
1.29      crook    4674: Forth supports two styles of comment; the traditional @i{in-line} comment,
                   4675: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
1.21      crook    4676: 
1.44      crook    4677: 
1.23      crook    4678: doc-(
1.21      crook    4679: doc-\
1.23      crook    4680: doc-\G
1.21      crook    4681: 
1.44      crook    4682: 
1.21      crook    4683: @node Boolean Flags, Arithmetic, Comments, Words
                   4684: @section Boolean Flags
1.26      crook    4685: @cindex Boolean flags
1.21      crook    4686: 
                   4687: A Boolean flag is cell-sized. A cell with all bits clear represents the
                   4688: flag @code{false} and a flag with all bits set represents the flag
1.26      crook    4689: @code{true}. Words that check a flag (for example, @code{IF}) will treat
1.29      crook    4690: a cell that has @i{any} bit set as @code{true}.
1.67      anton    4691: @c on and off to Memory? 
                   4692: @c true and false to "Bitwise operations" or "Numeric comparison"?
1.44      crook    4693: 
1.21      crook    4694: doc-true
                   4695: doc-false
1.29      crook    4696: doc-on
                   4697: doc-off
1.21      crook    4698: 
1.44      crook    4699: 
1.21      crook    4700: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
1.1       anton    4701: @section Arithmetic
                   4702: @cindex arithmetic words
                   4703: 
                   4704: @cindex division with potentially negative operands
                   4705: Forth arithmetic is not checked, i.e., you will not hear about integer
                   4706: overflow on addition or multiplication, you may hear about division by
                   4707: zero if you are lucky. The operator is written after the operands, but
                   4708: the operands are still in the original order. I.e., the infix @code{2-1}
                   4709: corresponds to @code{2 1 -}. Forth offers a variety of division
                   4710: operators. If you perform division with potentially negative operands,
                   4711: you do not want to use @code{/} or @code{/mod} with its undefined
                   4712: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
                   4713: former, @pxref{Mixed precision}).
1.26      crook    4714: @comment TODO discuss the different division forms and the std approach
1.1       anton    4715: 
                   4716: @menu
                   4717: * Single precision::            
1.67      anton    4718: * Double precision::            Double-cell integer arithmetic
1.1       anton    4719: * Bitwise operations::          
1.67      anton    4720: * Numeric comparison::          
1.29      crook    4721: * Mixed precision::             Operations with single and double-cell integers
1.1       anton    4722: * Floating Point::              
                   4723: @end menu
                   4724: 
1.67      anton    4725: @node Single precision, Double precision, Arithmetic, Arithmetic
1.1       anton    4726: @subsection Single precision
                   4727: @cindex single precision arithmetic words
                   4728: 
1.67      anton    4729: @c !! cell undefined
                   4730: 
                   4731: By default, numbers in Forth are single-precision integers that are one
1.26      crook    4732: cell in size. They can be signed or unsigned, depending upon how you
1.49      anton    4733: treat them. For the rules used by the text interpreter for recognising
                   4734: single-precision integers see @ref{Number Conversion}.
1.21      crook    4735: 
1.67      anton    4736: These words are all defined for signed operands, but some of them also
                   4737: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
                   4738: @code{*}.
1.44      crook    4739: 
1.1       anton    4740: doc-+
1.21      crook    4741: doc-1+
1.1       anton    4742: doc--
1.21      crook    4743: doc-1-
1.1       anton    4744: doc-*
                   4745: doc-/
                   4746: doc-mod
                   4747: doc-/mod
                   4748: doc-negate
                   4749: doc-abs
                   4750: doc-min
                   4751: doc-max
1.27      crook    4752: doc-floored
1.1       anton    4753: 
1.44      crook    4754: 
1.67      anton    4755: @node Double precision, Bitwise operations, Single precision, Arithmetic
1.21      crook    4756: @subsection Double precision
                   4757: @cindex double precision arithmetic words
                   4758: 
1.49      anton    4759: For the rules used by the text interpreter for
                   4760: recognising double-precision integers, see @ref{Number Conversion}.
1.21      crook    4761: 
                   4762: A double precision number is represented by a cell pair, with the most
1.67      anton    4763: significant cell at the TOS. It is trivial to convert an unsigned single
                   4764: to a double: simply push a @code{0} onto the TOS. Since numbers are
                   4765: represented by Gforth using 2's complement arithmetic, converting a
                   4766: signed single to a (signed) double requires sign-extension across the
                   4767: most significant cell. This can be achieved using @code{s>d}. The moral
                   4768: of the story is that you cannot convert a number without knowing whether
                   4769: it represents an unsigned or a signed number.
1.21      crook    4770: 
1.67      anton    4771: These words are all defined for signed operands, but some of them also
                   4772: work for unsigned numbers: @code{d+}, @code{d-}.
1.44      crook    4773: 
1.21      crook    4774: doc-s>d
1.67      anton    4775: doc-d>s
1.21      crook    4776: doc-d+
                   4777: doc-d-
                   4778: doc-dnegate
                   4779: doc-dabs
                   4780: doc-dmin
                   4781: doc-dmax
                   4782: 
1.44      crook    4783: 
1.67      anton    4784: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
                   4785: @subsection Bitwise operations
                   4786: @cindex bitwise operation words
                   4787: 
                   4788: 
                   4789: doc-and
                   4790: doc-or
                   4791: doc-xor
                   4792: doc-invert
                   4793: doc-lshift
                   4794: doc-rshift
                   4795: doc-2*
                   4796: doc-d2*
                   4797: doc-2/
                   4798: doc-d2/
                   4799: 
                   4800: 
                   4801: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
1.21      crook    4802: @subsection Numeric comparison
                   4803: @cindex numeric comparison words
                   4804: 
1.67      anton    4805: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
                   4806: d0= d0<>}) work for for both signed and unsigned numbers.
1.44      crook    4807: 
1.28      crook    4808: doc-<
                   4809: doc-<=
                   4810: doc-<>
                   4811: doc-=
                   4812: doc->
                   4813: doc->=
                   4814: 
1.21      crook    4815: doc-0<
1.23      crook    4816: doc-0<=
1.21      crook    4817: doc-0<>
                   4818: doc-0=
1.23      crook    4819: doc-0>
                   4820: doc-0>=
1.28      crook    4821: 
                   4822: doc-u<
                   4823: doc-u<=
1.44      crook    4824: @c u<> and u= exist but are the same as <> and =
1.31      anton    4825: @c doc-u<>
                   4826: @c doc-u=
1.28      crook    4827: doc-u>
                   4828: doc-u>=
                   4829: 
                   4830: doc-within
                   4831: 
                   4832: doc-d<
                   4833: doc-d<=
                   4834: doc-d<>
                   4835: doc-d=
                   4836: doc-d>
                   4837: doc-d>=
1.23      crook    4838: 
1.21      crook    4839: doc-d0<
1.23      crook    4840: doc-d0<=
                   4841: doc-d0<>
1.21      crook    4842: doc-d0=
1.23      crook    4843: doc-d0>
                   4844: doc-d0>=
                   4845: 
1.21      crook    4846: doc-du<
1.28      crook    4847: doc-du<=
1.44      crook    4848: @c du<> and du= exist but are the same as d<> and d=
1.31      anton    4849: @c doc-du<>
                   4850: @c doc-du=
1.28      crook    4851: doc-du>
                   4852: doc-du>=
1.1       anton    4853: 
1.44      crook    4854: 
1.21      crook    4855: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
1.1       anton    4856: @subsection Mixed precision
                   4857: @cindex mixed precision arithmetic words
                   4858: 
1.44      crook    4859: 
1.1       anton    4860: doc-m+
                   4861: doc-*/
                   4862: doc-*/mod
                   4863: doc-m*
                   4864: doc-um*
                   4865: doc-m*/
                   4866: doc-um/mod
                   4867: doc-fm/mod
                   4868: doc-sm/rem
                   4869: 
1.44      crook    4870: 
1.21      crook    4871: @node Floating Point,  , Mixed precision, Arithmetic
1.1       anton    4872: @subsection Floating Point
                   4873: @cindex floating point arithmetic words
                   4874: 
1.49      anton    4875: For the rules used by the text interpreter for
                   4876: recognising floating-point numbers see @ref{Number Conversion}.
1.1       anton    4877: 
1.67      anton    4878: Gforth has a separate floating point stack, but the documentation uses
                   4879: the unified notation.@footnote{It's easy to generate the separate
                   4880: notation from that by just separating the floating-point numbers out:
                   4881: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
                   4882: r3 )}.}
1.1       anton    4883: 
                   4884: @cindex floating-point arithmetic, pitfalls
                   4885: Floating point numbers have a number of unpleasant surprises for the
                   4886: unwary (e.g., floating point addition is not associative) and even a few
                   4887: for the wary. You should not use them unless you know what you are doing
                   4888: or you don't care that the results you get are totally bogus. If you
                   4889: want to learn about the problems of floating point numbers (and how to
1.66      anton    4890: avoid them), you might start with @cite{David Goldberg,
                   4891: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
                   4892: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
                   4893: Surveys 23(1):5@minus{}48, March 1991}.
1.1       anton    4894: 
1.44      crook    4895: 
1.21      crook    4896: doc-d>f
                   4897: doc-f>d
1.1       anton    4898: doc-f+
                   4899: doc-f-
                   4900: doc-f*
                   4901: doc-f/
                   4902: doc-fnegate
                   4903: doc-fabs
                   4904: doc-fmax
                   4905: doc-fmin
                   4906: doc-floor
                   4907: doc-fround
                   4908: doc-f**
                   4909: doc-fsqrt
                   4910: doc-fexp
                   4911: doc-fexpm1
                   4912: doc-fln
                   4913: doc-flnp1
                   4914: doc-flog
                   4915: doc-falog
1.32      anton    4916: doc-f2*
                   4917: doc-f2/
                   4918: doc-1/f
                   4919: doc-precision
                   4920: doc-set-precision
                   4921: 
                   4922: @cindex angles in trigonometric operations
                   4923: @cindex trigonometric operations
                   4924: Angles in floating point operations are given in radians (a full circle
                   4925: has 2 pi radians).
                   4926: 
1.1       anton    4927: doc-fsin
                   4928: doc-fcos
                   4929: doc-fsincos
                   4930: doc-ftan
                   4931: doc-fasin
                   4932: doc-facos
                   4933: doc-fatan
                   4934: doc-fatan2
                   4935: doc-fsinh
                   4936: doc-fcosh
                   4937: doc-ftanh
                   4938: doc-fasinh
                   4939: doc-facosh
                   4940: doc-fatanh
1.21      crook    4941: doc-pi
1.28      crook    4942: 
1.32      anton    4943: @cindex equality of floats
                   4944: @cindex floating-point comparisons
1.31      anton    4945: One particular problem with floating-point arithmetic is that comparison
                   4946: for equality often fails when you would expect it to succeed.  For this
                   4947: reason approximate equality is often preferred (but you still have to
1.67      anton    4948: know what you are doing).  Also note that IEEE NaNs may compare
1.68      anton    4949: differently from what you might expect.  The comparison words are:
1.31      anton    4950: 
                   4951: doc-f~rel
                   4952: doc-f~abs
1.68      anton    4953: doc-f~
1.31      anton    4954: doc-f=
                   4955: doc-f<>
                   4956: 
                   4957: doc-f<
                   4958: doc-f<=
                   4959: doc-f>
                   4960: doc-f>=
                   4961: 
1.21      crook    4962: doc-f0<
1.28      crook    4963: doc-f0<=
                   4964: doc-f0<>
1.21      crook    4965: doc-f0=
1.28      crook    4966: doc-f0>
                   4967: doc-f0>=
                   4968: 
1.1       anton    4969: 
                   4970: @node Stack Manipulation, Memory, Arithmetic, Words
                   4971: @section Stack Manipulation
                   4972: @cindex stack manipulation words
                   4973: 
                   4974: @cindex floating-point stack in the standard
1.21      crook    4975: Gforth maintains a number of separate stacks:
                   4976: 
1.29      crook    4977: @cindex data stack
                   4978: @cindex parameter stack
1.21      crook    4979: @itemize @bullet
                   4980: @item
1.29      crook    4981: A data stack (also known as the @dfn{parameter stack}) -- for
                   4982: characters, cells, addresses, and double cells.
1.21      crook    4983: 
1.29      crook    4984: @cindex floating-point stack
1.21      crook    4985: @item
1.44      crook    4986: A floating point stack -- for holding floating point (FP) numbers.
1.21      crook    4987: 
1.29      crook    4988: @cindex return stack
1.21      crook    4989: @item
1.44      crook    4990: A return stack -- for holding the return addresses of colon
1.32      anton    4991: definitions and other (non-FP) data.
1.21      crook    4992: 
1.29      crook    4993: @cindex locals stack
1.21      crook    4994: @item
1.44      crook    4995: A locals stack -- for holding local variables.
1.21      crook    4996: @end itemize
                   4997: 
1.1       anton    4998: @menu
                   4999: * Data stack::                  
                   5000: * Floating point stack::        
                   5001: * Return stack::                
                   5002: * Locals stack::                
                   5003: * Stack pointer manipulation::  
                   5004: @end menu
                   5005: 
                   5006: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
                   5007: @subsection Data stack
                   5008: @cindex data stack manipulation words
                   5009: @cindex stack manipulations words, data stack
                   5010: 
1.44      crook    5011: 
1.1       anton    5012: doc-drop
                   5013: doc-nip
                   5014: doc-dup
                   5015: doc-over
                   5016: doc-tuck
                   5017: doc-swap
1.21      crook    5018: doc-pick
1.1       anton    5019: doc-rot
                   5020: doc--rot
                   5021: doc-?dup
                   5022: doc-roll
                   5023: doc-2drop
                   5024: doc-2nip
                   5025: doc-2dup
                   5026: doc-2over
                   5027: doc-2tuck
                   5028: doc-2swap
                   5029: doc-2rot
                   5030: 
1.44      crook    5031: 
1.1       anton    5032: @node Floating point stack, Return stack, Data stack, Stack Manipulation
                   5033: @subsection Floating point stack
                   5034: @cindex floating-point stack manipulation words
                   5035: @cindex stack manipulation words, floating-point stack
                   5036: 
1.32      anton    5037: Whilst every sane Forth has a separate floating-point stack, it is not
                   5038: strictly required; an ANS Forth system could theoretically keep
                   5039: floating-point numbers on the data stack. As an additional difficulty,
                   5040: you don't know how many cells a floating-point number takes. It is
                   5041: reportedly possible to write words in a way that they work also for a
                   5042: unified stack model, but we do not recommend trying it. Instead, just
                   5043: say that your program has an environmental dependency on a separate
                   5044: floating-point stack.
                   5045: 
                   5046: doc-floating-stack
                   5047: 
1.1       anton    5048: doc-fdrop
                   5049: doc-fnip
                   5050: doc-fdup
                   5051: doc-fover
                   5052: doc-ftuck
                   5053: doc-fswap
1.21      crook    5054: doc-fpick
1.1       anton    5055: doc-frot
                   5056: 
1.44      crook    5057: 
1.1       anton    5058: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
                   5059: @subsection Return stack
                   5060: @cindex return stack manipulation words
                   5061: @cindex stack manipulation words, return stack
                   5062: 
1.32      anton    5063: @cindex return stack and locals
                   5064: @cindex locals and return stack
                   5065: A Forth system is allowed to keep local variables on the
                   5066: return stack. This is reasonable, as local variables usually eliminate
                   5067: the need to use the return stack explicitly. So, if you want to produce
                   5068: a standard compliant program and you are using local variables in a
                   5069: word, forget about return stack manipulations in that word (refer to the
                   5070: standard document for the exact rules).
                   5071: 
1.1       anton    5072: doc->r
                   5073: doc-r>
                   5074: doc-r@
                   5075: doc-rdrop
                   5076: doc-2>r
                   5077: doc-2r>
                   5078: doc-2r@
                   5079: doc-2rdrop
                   5080: 
1.44      crook    5081: 
1.1       anton    5082: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
                   5083: @subsection Locals stack
                   5084: 
1.78      anton    5085: Gforth uses an extra locals stack.  It is described, along with the
                   5086: reasons for its existence, in @ref{Locals implementation}.
1.21      crook    5087: 
1.1       anton    5088: @node Stack pointer manipulation,  , Locals stack, Stack Manipulation
                   5089: @subsection Stack pointer manipulation
                   5090: @cindex stack pointer manipulation words
                   5091: 
1.44      crook    5092: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
1.21      crook    5093: doc-sp0
1.1       anton    5094: doc-sp@
                   5095: doc-sp!
1.21      crook    5096: doc-fp0
1.1       anton    5097: doc-fp@
                   5098: doc-fp!
1.21      crook    5099: doc-rp0
1.1       anton    5100: doc-rp@
                   5101: doc-rp!
1.21      crook    5102: doc-lp0
1.1       anton    5103: doc-lp@
                   5104: doc-lp!
                   5105: 
1.44      crook    5106: 
1.1       anton    5107: @node Memory, Control Structures, Stack Manipulation, Words
                   5108: @section Memory
1.26      crook    5109: @cindex memory words
1.1       anton    5110: 
1.32      anton    5111: @menu
                   5112: * Memory model::                
                   5113: * Dictionary allocation::       
                   5114: * Heap Allocation::             
                   5115: * Memory Access::               
                   5116: * Address arithmetic::          
                   5117: * Memory Blocks::               
                   5118: @end menu
                   5119: 
1.67      anton    5120: In addition to the standard Forth memory allocation words, there is also
                   5121: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
                   5122: garbage collector}.
                   5123: 
1.32      anton    5124: @node Memory model, Dictionary allocation, Memory, Memory
                   5125: @subsection ANS Forth and Gforth memory models
                   5126: 
                   5127: @c The ANS Forth description is a mess (e.g., is the heap part of
                   5128: @c the dictionary?), so let's not stick to closely with it.
                   5129: 
1.67      anton    5130: ANS Forth considers a Forth system as consisting of several address
                   5131: spaces, of which only @dfn{data space} is managed and accessible with
                   5132: the memory words.  Memory not necessarily in data space includes the
                   5133: stacks, the code (called code space) and the headers (called name
                   5134: space). In Gforth everything is in data space, but the code for the
                   5135: primitives is usually read-only.
1.32      anton    5136: 
                   5137: Data space is divided into a number of areas: The (data space portion of
                   5138: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
                   5139: refer to the search data structure embodied in word lists and headers,
                   5140: because it is used for looking up names, just as you would in a
                   5141: conventional dictionary.}, the heap, and a number of system-allocated
                   5142: buffers.
                   5143: 
1.68      anton    5144: @cindex address arithmetic restrictions, ANS vs. Gforth
                   5145: @cindex contiguous regions, ANS vs. Gforth
1.32      anton    5146: In ANS Forth data space is also divided into contiguous regions.  You
                   5147: can only use address arithmetic within a contiguous region, not between
                   5148: them.  Usually each allocation gives you one contiguous region, but the
1.33      anton    5149: dictionary allocation words have additional rules (@pxref{Dictionary
1.32      anton    5150: allocation}).
                   5151: 
                   5152: Gforth provides one big address space, and address arithmetic can be
                   5153: performed between any addresses. However, in the dictionary headers or
                   5154: code are interleaved with data, so almost the only contiguous data space
                   5155: regions there are those described by ANS Forth as contiguous; but you
                   5156: can be sure that the dictionary is allocated towards increasing
                   5157: addresses even between contiguous regions.  The memory order of
                   5158: allocations in the heap is platform-dependent (and possibly different
                   5159: from one run to the next).
                   5160: 
1.27      crook    5161: 
1.32      anton    5162: @node Dictionary allocation, Heap Allocation, Memory model, Memory
                   5163: @subsection Dictionary allocation
1.27      crook    5164: @cindex reserving data space
                   5165: @cindex data space - reserving some
                   5166: 
1.32      anton    5167: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
                   5168: you want to deallocate X, you also deallocate everything
                   5169: allocated after X.
                   5170: 
1.68      anton    5171: @cindex contiguous regions in dictionary allocation
1.32      anton    5172: The allocations using the words below are contiguous and grow the region
                   5173: towards increasing addresses.  Other words that allocate dictionary
                   5174: memory of any kind (i.e., defining words including @code{:noname}) end
                   5175: the contiguous region and start a new one.
                   5176: 
                   5177: In ANS Forth only @code{create}d words are guaranteed to produce an
                   5178: address that is the start of the following contiguous region.  In
                   5179: particular, the cell allocated by @code{variable} is not guaranteed to
                   5180: be contiguous with following @code{allot}ed memory.
                   5181: 
                   5182: You can deallocate memory by using @code{allot} with a negative argument
                   5183: (with some restrictions, see @code{allot}). For larger deallocations use
                   5184: @code{marker}.
1.27      crook    5185: 
1.29      crook    5186: 
1.27      crook    5187: doc-here
                   5188: doc-unused
                   5189: doc-allot
                   5190: doc-c,
1.29      crook    5191: doc-f,
1.27      crook    5192: doc-,
                   5193: doc-2,
                   5194: 
1.32      anton    5195: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
                   5196: course you should allocate memory in an aligned way, too. I.e., before
                   5197: allocating allocating a cell, @code{here} must be cell-aligned, etc.
                   5198: The words below align @code{here} if it is not already.  Basically it is
                   5199: only already aligned for a type, if the last allocation was a multiple
                   5200: of the size of this type and if @code{here} was aligned for this type
                   5201: before.
                   5202: 
                   5203: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
                   5204: ANS Forth (@code{maxalign}ed in Gforth).
                   5205: 
                   5206: doc-align
                   5207: doc-falign
                   5208: doc-sfalign
                   5209: doc-dfalign
                   5210: doc-maxalign
                   5211: doc-cfalign
                   5212: 
                   5213: 
                   5214: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
                   5215: @subsection Heap allocation
                   5216: @cindex heap allocation
                   5217: @cindex dynamic allocation of memory
                   5218: @cindex memory-allocation word set
                   5219: 
1.68      anton    5220: @cindex contiguous regions and heap allocation
1.32      anton    5221: Heap allocation supports deallocation of allocated memory in any
                   5222: order. Dictionary allocation is not affected by it (i.e., it does not
                   5223: end a contiguous region). In Gforth, these words are implemented using
                   5224: the standard C library calls malloc(), free() and resize().
                   5225: 
1.68      anton    5226: The memory region produced by one invocation of @code{allocate} or
                   5227: @code{resize} is internally contiguous.  There is no contiguity between
                   5228: such a region and any other region (including others allocated from the
                   5229: heap).
                   5230: 
1.32      anton    5231: doc-allocate
                   5232: doc-free
                   5233: doc-resize
                   5234: 
1.27      crook    5235: 
1.32      anton    5236: @node Memory Access, Address arithmetic, Heap Allocation, Memory
1.1       anton    5237: @subsection Memory Access
                   5238: @cindex memory access words
                   5239: 
                   5240: doc-@
                   5241: doc-!
                   5242: doc-+!
                   5243: doc-c@
                   5244: doc-c!
                   5245: doc-2@
                   5246: doc-2!
                   5247: doc-f@
                   5248: doc-f!
                   5249: doc-sf@
                   5250: doc-sf!
                   5251: doc-df@
                   5252: doc-df!
                   5253: 
1.68      anton    5254: 
1.32      anton    5255: @node Address arithmetic, Memory Blocks, Memory Access, Memory
                   5256: @subsection Address arithmetic
1.1       anton    5257: @cindex address arithmetic words
                   5258: 
1.67      anton    5259: Address arithmetic is the foundation on which you can build data
                   5260: structures like arrays, records (@pxref{Structures}) and objects
                   5261: (@pxref{Object-oriented Forth}).
1.32      anton    5262: 
1.68      anton    5263: @cindex address unit
                   5264: @cindex au (address unit)
1.1       anton    5265: ANS Forth does not specify the sizes of the data types. Instead, it
                   5266: offers a number of words for computing sizes and doing address
1.29      crook    5267: arithmetic. Address arithmetic is performed in terms of address units
                   5268: (aus); on most systems the address unit is one byte. Note that a
                   5269: character may have more than one au, so @code{chars} is no noop (on
1.68      anton    5270: platforms where it is a noop, it compiles to nothing).
1.1       anton    5271: 
1.67      anton    5272: The basic address arithmetic words are @code{+} and @code{-}.  E.g., if
                   5273: you have the address of a cell, perform @code{1 cells +}, and you will
                   5274: have the address of the next cell.
                   5275: 
1.68      anton    5276: @cindex contiguous regions and address arithmetic
1.67      anton    5277: In ANS Forth you can perform address arithmetic only within a contiguous
                   5278: region, i.e., if you have an address into one region, you can only add
                   5279: and subtract such that the result is still within the region; you can
                   5280: only subtract or compare addresses from within the same contiguous
                   5281: region.  Reasons: several contiguous regions can be arranged in memory
                   5282: in any way; on segmented systems addresses may have unusual
                   5283: representations, such that address arithmetic only works within a
                   5284: region.  Gforth provides a few more guarantees (linear address space,
                   5285: dictionary grows upwards), but in general I have found it easy to stay
                   5286: within contiguous regions (exception: computing and comparing to the
                   5287: address just beyond the end of an array).
                   5288: 
1.1       anton    5289: @cindex alignment of addresses for types
                   5290: ANS Forth also defines words for aligning addresses for specific
                   5291: types. Many computers require that accesses to specific data types
                   5292: must only occur at specific addresses; e.g., that cells may only be
                   5293: accessed at addresses divisible by 4. Even if a machine allows unaligned
                   5294: accesses, it can usually perform aligned accesses faster. 
                   5295: 
                   5296: For the performance-conscious: alignment operations are usually only
                   5297: necessary during the definition of a data structure, not during the
                   5298: (more frequent) accesses to it.
                   5299: 
                   5300: ANS Forth defines no words for character-aligning addresses. This is not
                   5301: an oversight, but reflects the fact that addresses that are not
                   5302: char-aligned have no use in the standard and therefore will not be
                   5303: created.
                   5304: 
                   5305: @cindex @code{CREATE} and alignment
1.29      crook    5306: ANS Forth guarantees that addresses returned by @code{CREATE}d words
1.1       anton    5307: are cell-aligned; in addition, Gforth guarantees that these addresses
                   5308: are aligned for all purposes.
                   5309: 
1.26      crook    5310: Note that the ANS Forth word @code{char} has nothing to do with address
                   5311: arithmetic.
1.1       anton    5312: 
1.44      crook    5313: 
1.1       anton    5314: doc-chars
                   5315: doc-char+
                   5316: doc-cells
                   5317: doc-cell+
                   5318: doc-cell
                   5319: doc-aligned
                   5320: doc-floats
                   5321: doc-float+
                   5322: doc-float
                   5323: doc-faligned
                   5324: doc-sfloats
                   5325: doc-sfloat+
                   5326: doc-sfaligned
                   5327: doc-dfloats
                   5328: doc-dfloat+
                   5329: doc-dfaligned
                   5330: doc-maxaligned
                   5331: doc-cfaligned
                   5332: doc-address-unit-bits
                   5333: 
1.44      crook    5334: 
1.32      anton    5335: @node Memory Blocks,  , Address arithmetic, Memory
1.1       anton    5336: @subsection Memory Blocks
                   5337: @cindex memory block words
1.27      crook    5338: @cindex character strings - moving and copying
                   5339: 
1.49      anton    5340: Memory blocks often represent character strings; For ways of storing
                   5341: character strings in memory see @ref{String Formats}.  For other
                   5342: string-processing words see @ref{Displaying characters and strings}.
1.1       anton    5343: 
1.67      anton    5344: A few of these words work on address unit blocks.  In that case, you
                   5345: usually have to insert @code{CHARS} before the word when working on
                   5346: character strings.  Most words work on character blocks, and expect a
                   5347: char-aligned address.
                   5348: 
                   5349: When copying characters between overlapping memory regions, use
                   5350: @code{chars move} or choose carefully between @code{cmove} and
                   5351: @code{cmove>}.
1.44      crook    5352: 
1.1       anton    5353: doc-move
                   5354: doc-erase
                   5355: doc-cmove
                   5356: doc-cmove>
                   5357: doc-fill
                   5358: doc-blank
1.21      crook    5359: doc-compare
                   5360: doc-search
1.27      crook    5361: doc--trailing
                   5362: doc-/string
1.82      anton    5363: doc-bounds
1.44      crook    5364: 
1.27      crook    5365: @comment TODO examples
                   5366: 
1.1       anton    5367: 
1.26      crook    5368: @node Control Structures, Defining Words, Memory, Words
1.1       anton    5369: @section Control Structures
                   5370: @cindex control structures
                   5371: 
1.33      anton    5372: Control structures in Forth cannot be used interpretively, only in a
                   5373: colon definition@footnote{To be precise, they have no interpretation
                   5374: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
                   5375: not like this limitation, but have not seen a satisfying way around it
                   5376: yet, although many schemes have been proposed.
1.1       anton    5377: 
                   5378: @menu
1.33      anton    5379: * Selection::                   IF ... ELSE ... ENDIF
                   5380: * Simple Loops::                BEGIN ...
1.29      crook    5381: * Counted Loops::               DO
1.67      anton    5382: * Arbitrary control structures::  
                   5383: * Calls and returns::           
1.1       anton    5384: * Exception Handling::          
                   5385: @end menu
                   5386: 
                   5387: @node Selection, Simple Loops, Control Structures, Control Structures
                   5388: @subsection Selection
                   5389: @cindex selection control structures
                   5390: @cindex control structures for selection
                   5391: 
                   5392: @cindex @code{IF} control structure
                   5393: @example
1.29      crook    5394: @i{flag}
1.1       anton    5395: IF
1.29      crook    5396:   @i{code}
1.1       anton    5397: ENDIF
                   5398: @end example
1.21      crook    5399: @noindent
1.33      anton    5400: 
1.44      crook    5401: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
                   5402: with any bit set represents truth) @i{code} is executed.
1.33      anton    5403: 
1.1       anton    5404: @example
1.29      crook    5405: @i{flag}
1.1       anton    5406: IF
1.29      crook    5407:   @i{code1}
1.1       anton    5408: ELSE
1.29      crook    5409:   @i{code2}
1.1       anton    5410: ENDIF
                   5411: @end example
                   5412: 
1.44      crook    5413: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
                   5414: executed.
1.33      anton    5415: 
1.1       anton    5416: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
                   5417: standard, and @code{ENDIF} is not, although it is quite popular. We
                   5418: recommend using @code{ENDIF}, because it is less confusing for people
                   5419: who also know other languages (and is not prone to reinforcing negative
                   5420: prejudices against Forth in these people). Adding @code{ENDIF} to a
                   5421: system that only supplies @code{THEN} is simple:
                   5422: @example
1.82      anton    5423: : ENDIF   POSTPONE then ; immediate
1.1       anton    5424: @end example
                   5425: 
                   5426: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
                   5427: (adv.)}  has the following meanings:
                   5428: @quotation
                   5429: ... 2b: following next after in order ... 3d: as a necessary consequence
                   5430: (if you were there, then you saw them).
                   5431: @end quotation
                   5432: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
                   5433: and many other programming languages has the meaning 3d.]
                   5434: 
1.21      crook    5435: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
1.1       anton    5436: you can avoid using @code{?dup}. Using these alternatives is also more
1.26      crook    5437: efficient than using @code{?dup}. Definitions in ANS Forth
1.1       anton    5438: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
                   5439: @file{compat/control.fs}.
                   5440: 
                   5441: @cindex @code{CASE} control structure
                   5442: @example
1.29      crook    5443: @i{n}
1.1       anton    5444: CASE
1.29      crook    5445:   @i{n1} OF @i{code1} ENDOF
                   5446:   @i{n2} OF @i{code2} ENDOF
1.1       anton    5447:   @dots{}
1.68      anton    5448:   ( n ) @i{default-code} ( n )
1.1       anton    5449: ENDCASE
                   5450: @end example
                   5451: 
1.68      anton    5452: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}.  If no
                   5453: @i{ni} matches, the optional @i{default-code} is executed. The optional
                   5454: default case can be added by simply writing the code after the last
                   5455: @code{ENDOF}. It may use @i{n}, which is on top of the stack, but must
                   5456: not consume it.
1.1       anton    5457: 
1.69      anton    5458: @progstyle
                   5459: To keep the code understandable, you should ensure that on all paths
                   5460: through a selection construct the stack is changed in the same way
                   5461: (wrt. number and types of stack items consumed and pushed).
                   5462: 
1.1       anton    5463: @node Simple Loops, Counted Loops, Selection, Control Structures
                   5464: @subsection Simple Loops
                   5465: @cindex simple loops
                   5466: @cindex loops without count 
                   5467: 
                   5468: @cindex @code{WHILE} loop
                   5469: @example
                   5470: BEGIN
1.29      crook    5471:   @i{code1}
                   5472:   @i{flag}
1.1       anton    5473: WHILE
1.29      crook    5474:   @i{code2}
1.1       anton    5475: REPEAT
                   5476: @end example
                   5477: 
1.29      crook    5478: @i{code1} is executed and @i{flag} is computed. If it is true,
                   5479: @i{code2} is executed and the loop is restarted; If @i{flag} is
1.1       anton    5480: false, execution continues after the @code{REPEAT}.
                   5481: 
                   5482: @cindex @code{UNTIL} loop
                   5483: @example
                   5484: BEGIN
1.29      crook    5485:   @i{code}
                   5486:   @i{flag}
1.1       anton    5487: UNTIL
                   5488: @end example
                   5489: 
1.29      crook    5490: @i{code} is executed. The loop is restarted if @code{flag} is false.
1.1       anton    5491: 
1.69      anton    5492: @progstyle
                   5493: To keep the code understandable, a complete iteration of the loop should
                   5494: not change the number and types of the items on the stacks.
                   5495: 
1.1       anton    5496: @cindex endless loop
                   5497: @cindex loops, endless
                   5498: @example
                   5499: BEGIN
1.29      crook    5500:   @i{code}
1.1       anton    5501: AGAIN
                   5502: @end example
                   5503: 
                   5504: This is an endless loop.
                   5505: 
                   5506: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
                   5507: @subsection Counted Loops
                   5508: @cindex counted loops
                   5509: @cindex loops, counted
                   5510: @cindex @code{DO} loops
                   5511: 
                   5512: The basic counted loop is:
                   5513: @example
1.29      crook    5514: @i{limit} @i{start}
1.1       anton    5515: ?DO
1.29      crook    5516:   @i{body}
1.1       anton    5517: LOOP
                   5518: @end example
                   5519: 
1.29      crook    5520: This performs one iteration for every integer, starting from @i{start}
                   5521: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
1.21      crook    5522: accessed with @code{i}. For example, the loop:
1.1       anton    5523: @example
                   5524: 10 0 ?DO
                   5525:   i .
                   5526: LOOP
                   5527: @end example
1.21      crook    5528: @noindent
                   5529: prints @code{0 1 2 3 4 5 6 7 8 9}
                   5530: 
1.1       anton    5531: The index of the innermost loop can be accessed with @code{i}, the index
                   5532: of the next loop with @code{j}, and the index of the third loop with
                   5533: @code{k}.
                   5534: 
1.44      crook    5535: 
1.1       anton    5536: doc-i
                   5537: doc-j
                   5538: doc-k
                   5539: 
1.44      crook    5540: 
1.1       anton    5541: The loop control data are kept on the return stack, so there are some
1.21      crook    5542: restrictions on mixing return stack accesses and counted loop words. In
                   5543: particuler, if you put values on the return stack outside the loop, you
                   5544: cannot read them inside the loop@footnote{well, not in a way that is
                   5545: portable.}. If you put values on the return stack within a loop, you
                   5546: have to remove them before the end of the loop and before accessing the
                   5547: index of the loop.
1.1       anton    5548: 
                   5549: There are several variations on the counted loop:
                   5550: 
1.21      crook    5551: @itemize @bullet
                   5552: @item
                   5553: @code{LEAVE} leaves the innermost counted loop immediately; execution
                   5554: continues after the associated @code{LOOP} or @code{NEXT}. For example:
                   5555: 
                   5556: @example
                   5557: 10 0 ?DO  i DUP . 3 = IF LEAVE THEN LOOP
                   5558: @end example
                   5559: prints @code{0 1 2 3}
                   5560: 
1.1       anton    5561: 
1.21      crook    5562: @item
                   5563: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
                   5564: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
                   5565: return stack so @code{EXIT} can get to its return address. For example:
                   5566: 
                   5567: @example
                   5568: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
                   5569: @end example
                   5570: prints @code{0 1 2 3}
                   5571: 
                   5572: 
                   5573: @item
1.29      crook    5574: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
1.1       anton    5575: (and @code{LOOP} iterates until they become equal by wrap-around
                   5576: arithmetic). This behaviour is usually not what you want. Therefore,
                   5577: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1.29      crook    5578: @code{?DO}), which do not enter the loop if @i{start} is greater than
                   5579: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1.1       anton    5580: unsigned loop parameters.
                   5581: 
1.21      crook    5582: @item
                   5583: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
                   5584: the loop, independent of the loop parameters. Do not use @code{DO}, even
                   5585: if you know that the loop is entered in any case. Such knowledge tends
                   5586: to become invalid during maintenance of a program, and then the
                   5587: @code{DO} will make trouble.
                   5588: 
                   5589: @item
1.29      crook    5590: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
                   5591: index by @i{n} instead of by 1. The loop is terminated when the border
                   5592: between @i{limit-1} and @i{limit} is crossed. E.g.:
1.1       anton    5593: 
1.21      crook    5594: @example
                   5595: 4 0 +DO  i .  2 +LOOP
                   5596: @end example
                   5597: @noindent
                   5598: prints @code{0 2}
                   5599: 
                   5600: @example
                   5601: 4 1 +DO  i .  2 +LOOP
                   5602: @end example
                   5603: @noindent
                   5604: prints @code{1 3}
1.1       anton    5605: 
1.68      anton    5606: @item
1.1       anton    5607: @cindex negative increment for counted loops
                   5608: @cindex counted loops with negative increment
1.29      crook    5609: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
1.1       anton    5610: 
1.21      crook    5611: @example
                   5612: -1 0 ?DO  i .  -1 +LOOP
                   5613: @end example
                   5614: @noindent
                   5615: prints @code{0 -1}
1.1       anton    5616: 
1.21      crook    5617: @example
                   5618: 0 0 ?DO  i .  -1 +LOOP
                   5619: @end example
                   5620: prints nothing.
1.1       anton    5621: 
1.29      crook    5622: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
                   5623: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
                   5624: index by @i{u} each iteration. The loop is terminated when the border
                   5625: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
1.1       anton    5626: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
                   5627: 
1.21      crook    5628: @example
                   5629: -2 0 -DO  i .  1 -LOOP
                   5630: @end example
                   5631: @noindent
                   5632: prints @code{0 -1}
1.1       anton    5633: 
1.21      crook    5634: @example
                   5635: -1 0 -DO  i .  1 -LOOP
                   5636: @end example
                   5637: @noindent
                   5638: prints @code{0}
                   5639: 
                   5640: @example
                   5641: 0 0 -DO  i .  1 -LOOP
                   5642: @end example
                   5643: @noindent
                   5644: prints nothing.
1.1       anton    5645: 
1.21      crook    5646: @end itemize
1.1       anton    5647: 
                   5648: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
1.26      crook    5649: @code{-LOOP} are not defined in ANS Forth. However, an implementation
                   5650: for these words that uses only standard words is provided in
                   5651: @file{compat/loops.fs}.
1.1       anton    5652: 
                   5653: 
                   5654: @cindex @code{FOR} loops
1.26      crook    5655: Another counted loop is:
1.1       anton    5656: @example
1.29      crook    5657: @i{n}
1.1       anton    5658: FOR
1.29      crook    5659:   @i{body}
1.1       anton    5660: NEXT
                   5661: @end example
                   5662: This is the preferred loop of native code compiler writers who are too
1.26      crook    5663: lazy to optimize @code{?DO} loops properly. This loop structure is not
1.29      crook    5664: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
                   5665: @code{i} produces values starting with @i{n} and ending with 0. Other
1.26      crook    5666: Forth systems may behave differently, even if they support @code{FOR}
                   5667: loops. To avoid problems, don't use @code{FOR} loops.
1.1       anton    5668: 
                   5669: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
                   5670: @subsection Arbitrary control structures
                   5671: @cindex control structures, user-defined
                   5672: 
                   5673: @cindex control-flow stack
                   5674: ANS Forth permits and supports using control structures in a non-nested
                   5675: way. Information about incomplete control structures is stored on the
                   5676: control-flow stack. This stack may be implemented on the Forth data
                   5677: stack, and this is what we have done in Gforth.
                   5678: 
                   5679: @cindex @code{orig}, control-flow stack item
                   5680: @cindex @code{dest}, control-flow stack item
                   5681: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
                   5682: entry represents a backward branch target. A few words are the basis for
                   5683: building any control structure possible (except control structures that
                   5684: need storage, like calls, coroutines, and backtracking).
                   5685: 
1.44      crook    5686: 
1.1       anton    5687: doc-if
                   5688: doc-ahead
                   5689: doc-then
                   5690: doc-begin
                   5691: doc-until
                   5692: doc-again
                   5693: doc-cs-pick
                   5694: doc-cs-roll
                   5695: 
1.44      crook    5696: 
1.21      crook    5697: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
                   5698: manipulate the control-flow stack in a portable way. Without them, you
                   5699: would need to know how many stack items are occupied by a control-flow
                   5700: entry (many systems use one cell. In Gforth they currently take three,
                   5701: but this may change in the future).
                   5702: 
1.1       anton    5703: Some standard control structure words are built from these words:
                   5704: 
1.44      crook    5705: 
1.1       anton    5706: doc-else
                   5707: doc-while
                   5708: doc-repeat
                   5709: 
1.44      crook    5710: 
                   5711: @noindent
1.1       anton    5712: Gforth adds some more control-structure words:
                   5713: 
1.44      crook    5714: 
1.1       anton    5715: doc-endif
                   5716: doc-?dup-if
                   5717: doc-?dup-0=-if
                   5718: 
1.44      crook    5719: 
                   5720: @noindent
1.1       anton    5721: Counted loop words constitute a separate group of words:
                   5722: 
1.44      crook    5723: 
1.1       anton    5724: doc-?do
                   5725: doc-+do
                   5726: doc-u+do
                   5727: doc--do
                   5728: doc-u-do
                   5729: doc-do
                   5730: doc-for
                   5731: doc-loop
                   5732: doc-+loop
                   5733: doc--loop
                   5734: doc-next
                   5735: doc-leave
                   5736: doc-?leave
                   5737: doc-unloop
                   5738: doc-done
                   5739: 
1.44      crook    5740: 
1.21      crook    5741: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
                   5742: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
1.1       anton    5743: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
                   5744: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
                   5745: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
                   5746: resolved (by using one of the loop-ending words or @code{DONE}).
                   5747: 
1.44      crook    5748: @noindent
1.26      crook    5749: Another group of control structure words are:
1.1       anton    5750: 
1.44      crook    5751: 
1.1       anton    5752: doc-case
                   5753: doc-endcase
                   5754: doc-of
                   5755: doc-endof
                   5756: 
1.44      crook    5757: 
1.21      crook    5758: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
                   5759: @code{CS-ROLL}.
1.1       anton    5760: 
                   5761: @subsubsection Programming Style
1.47      crook    5762: @cindex control structures programming style
                   5763: @cindex programming style, arbitrary control structures
1.1       anton    5764: 
                   5765: In order to ensure readability we recommend that you do not create
                   5766: arbitrary control structures directly, but define new control structure
                   5767: words for the control structure you want and use these words in your
1.26      crook    5768: program. For example, instead of writing:
1.1       anton    5769: 
                   5770: @example
1.26      crook    5771: BEGIN
1.1       anton    5772:   ...
1.26      crook    5773: IF [ 1 CS-ROLL ]
1.1       anton    5774:   ...
1.26      crook    5775: AGAIN THEN
1.1       anton    5776: @end example
                   5777: 
1.21      crook    5778: @noindent
1.1       anton    5779: we recommend defining control structure words, e.g.,
                   5780: 
                   5781: @example
1.26      crook    5782: : WHILE ( DEST -- ORIG DEST )
                   5783:  POSTPONE IF
                   5784:  1 CS-ROLL ; immediate
                   5785: 
                   5786: : REPEAT ( orig dest -- )
                   5787:  POSTPONE AGAIN
                   5788:  POSTPONE THEN ; immediate
1.1       anton    5789: @end example
                   5790: 
1.21      crook    5791: @noindent
1.1       anton    5792: and then using these to create the control structure:
                   5793: 
                   5794: @example
1.26      crook    5795: BEGIN
1.1       anton    5796:   ...
1.26      crook    5797: WHILE
1.1       anton    5798:   ...
1.26      crook    5799: REPEAT
1.1       anton    5800: @end example
                   5801: 
                   5802: That's much easier to read, isn't it? Of course, @code{REPEAT} and
                   5803: @code{WHILE} are predefined, so in this example it would not be
                   5804: necessary to define them.
                   5805: 
                   5806: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
                   5807: @subsection Calls and returns
                   5808: @cindex calling a definition
                   5809: @cindex returning from a definition
                   5810: 
1.3       anton    5811: @cindex recursive definitions
                   5812: A definition can be called simply be writing the name of the definition
1.26      crook    5813: to be called. Normally a definition is invisible during its own
1.3       anton    5814: definition. If you want to write a directly recursive definition, you
1.26      crook    5815: can use @code{recursive} to make the current definition visible, or
                   5816: @code{recurse} to call the current definition directly.
1.3       anton    5817: 
1.44      crook    5818: 
1.3       anton    5819: doc-recursive
                   5820: doc-recurse
                   5821: 
1.44      crook    5822: 
1.21      crook    5823: @comment TODO add example of the two recursion methods
1.12      anton    5824: @quotation
                   5825: @progstyle
                   5826: I prefer using @code{recursive} to @code{recurse}, because calling the
                   5827: definition by name is more descriptive (if the name is well-chosen) than
                   5828: the somewhat cryptic @code{recurse}.  E.g., in a quicksort
                   5829: implementation, it is much better to read (and think) ``now sort the
                   5830: partitions'' than to read ``now do a recursive call''.
                   5831: @end quotation
1.3       anton    5832: 
1.29      crook    5833: For mutual recursion, use @code{Defer}red words, like this:
1.3       anton    5834: 
                   5835: @example
1.28      crook    5836: Defer foo
1.3       anton    5837: 
                   5838: : bar ( ... -- ... )
                   5839:  ... foo ... ;
                   5840: 
                   5841: :noname ( ... -- ... )
                   5842:  ... bar ... ;
                   5843: IS foo
                   5844: @end example
                   5845: 
1.44      crook    5846: Deferred words are discussed in more detail in @ref{Deferred words}.
1.33      anton    5847: 
1.26      crook    5848: The current definition returns control to the calling definition when
1.33      anton    5849: the end of the definition is reached or @code{EXIT} is encountered.
1.1       anton    5850: 
                   5851: doc-exit
                   5852: doc-;s
                   5853: 
1.44      crook    5854: 
1.1       anton    5855: @node Exception Handling,  , Calls and returns, Control Structures
                   5856: @subsection Exception Handling
1.26      crook    5857: @cindex exceptions
1.1       anton    5858: 
1.68      anton    5859: @c quit is a very bad idea for error handling, 
                   5860: @c because it does not translate into a THROW
                   5861: @c it also does not belong into this chapter
                   5862: 
                   5863: If a word detects an error condition that it cannot handle, it can
                   5864: @code{throw} an exception.  In the simplest case, this will terminate
                   5865: your program, and report an appropriate error.
1.21      crook    5866: 
1.68      anton    5867: doc-throw
1.1       anton    5868: 
1.69      anton    5869: @code{Throw} consumes a cell-sized error number on the stack. There are
                   5870: some predefined error numbers in ANS Forth (see @file{errors.fs}).  In
                   5871: Gforth (and most other systems) you can use the iors produced by various
                   5872: words as error numbers (e.g., a typical use of @code{allocate} is
                   5873: @code{allocate throw}).  Gforth also provides the word @code{exception}
                   5874: to define your own error numbers (with decent error reporting); an ANS
                   5875: Forth version of this word (but without the error messages) is available
                   5876: in @code{compat/except.fs}.  And finally, you can use your own error
1.68      anton    5877: numbers (anything outside the range -4095..0), but won't get nice error
                   5878: messages, only numbers.  For example, try:
                   5879: 
                   5880: @example
1.69      anton    5881: -10 throw                    \ ANS defined
                   5882: -267 throw                   \ system defined
                   5883: s" my error" exception throw \ user defined
                   5884: 7 throw                      \ arbitrary number
1.68      anton    5885: @end example
                   5886: 
                   5887: doc---exception-exception
1.1       anton    5888: 
1.69      anton    5889: A common idiom to @code{THROW} a specific error if a flag is true is
                   5890: this:
                   5891: 
                   5892: @example
                   5893: @code{( flag ) 0<> @i{errno} and throw}
                   5894: @end example
                   5895: 
                   5896: Your program can provide exception handlers to catch exceptions.  An
                   5897: exception handler can be used to correct the problem, or to clean up
                   5898: some data structures and just throw the exception to the next exception
                   5899: handler.  Note that @code{throw} jumps to the dynamically innermost
                   5900: exception handler.  The system's exception handler is outermost, and just
                   5901: prints an error and restarts command-line interpretation (or, in batch
                   5902: mode (i.e., while processing the shell command line), leaves Gforth).
1.1       anton    5903: 
1.68      anton    5904: The ANS Forth way to catch exceptions is @code{catch}:
1.1       anton    5905: 
1.68      anton    5906: doc-catch
                   5907: 
                   5908: The most common use of exception handlers is to clean up the state when
                   5909: an error happens.  E.g.,
1.1       anton    5910: 
1.26      crook    5911: @example
1.68      anton    5912: base @ >r hex \ actually the hex should be inside foo, or we h
                   5913: ['] foo catch ( nerror|0 )
                   5914: r> base !
1.69      anton    5915: ( nerror|0 ) throw \ pass it on
1.26      crook    5916: @end example
1.1       anton    5917: 
1.69      anton    5918: A use of @code{catch} for handling the error @code{myerror} might look
                   5919: like this:
1.44      crook    5920: 
1.68      anton    5921: @example
1.69      anton    5922: ['] foo catch
                   5923: CASE
                   5924:   myerror OF ... ( do something about it ) ENDOF
                   5925:   dup throw \ default: pass other errors on, do nothing on non-errors
                   5926: ENDCASE
1.68      anton    5927: @end example
1.44      crook    5928: 
1.68      anton    5929: Having to wrap the code into a separate word is often cumbersome,
                   5930: therefore Gforth provides an alternative syntax:
1.1       anton    5931: 
                   5932: @example
1.69      anton    5933: TRY
1.68      anton    5934:   @i{code1}
1.69      anton    5935: RECOVER     \ optional
1.68      anton    5936:   @i{code2} \ optional
1.69      anton    5937: ENDTRY
1.1       anton    5938: @end example
                   5939: 
1.68      anton    5940: This performs @i{Code1}.  If @i{code1} completes normally, execution
                   5941: continues after the @code{endtry}.  If @i{Code1} throws, the stacks are
                   5942: reset to the state during @code{try}, the throw value is pushed on the
                   5943: data stack, and execution constinues at @i{code2}, and finally falls
                   5944: through the @code{endtry} into the following code. If there is no
                   5945: @code{recover} clause, this works like an empty recover clause.
1.26      crook    5946: 
1.68      anton    5947: doc-try
                   5948: doc-recover
                   5949: doc-endtry
1.26      crook    5950: 
1.69      anton    5951: The cleanup example from above in this syntax:
1.26      crook    5952: 
1.68      anton    5953: @example
1.69      anton    5954: base @ >r TRY
1.68      anton    5955:   hex foo \ now the hex is placed correctly
1.69      anton    5956:   0       \ value for throw
                   5957: ENDTRY
1.68      anton    5958: r> base ! throw
1.1       anton    5959: @end example
                   5960: 
1.69      anton    5961: And here's the error handling example:
1.1       anton    5962: 
1.68      anton    5963: @example
1.69      anton    5964: TRY
1.68      anton    5965:   foo
1.69      anton    5966: RECOVER
                   5967:   CASE
                   5968:     myerror OF ... ( do something about it ) ENDOF
                   5969:     throw \ pass other errors on
                   5970:   ENDCASE
                   5971: ENDTRY
1.68      anton    5972: @end example
1.1       anton    5973: 
1.69      anton    5974: @progstyle
                   5975: As usual, you should ensure that the stack depth is statically known at
                   5976: the end: either after the @code{throw} for passing on errors, or after
                   5977: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
                   5978: selection construct for handling the error).
                   5979: 
1.68      anton    5980: There are two alternatives to @code{throw}: @code{Abort"} is conditional
                   5981: and you can provide an error message.  @code{Abort} just produces an
                   5982: ``Aborted'' error.
1.1       anton    5983: 
1.68      anton    5984: The problem with these words is that exception handlers cannot
                   5985: differentiate between different @code{abort"}s; they just look like
                   5986: @code{-2 throw} to them (the error message cannot be accessed by
                   5987: standard programs).  Similar @code{abort} looks like @code{-1 throw} to
                   5988: exception handlers.
1.44      crook    5989: 
1.68      anton    5990: doc-abort"
1.26      crook    5991: doc-abort
1.29      crook    5992: 
                   5993: 
1.44      crook    5994: 
1.29      crook    5995: @c -------------------------------------------------------------
1.47      crook    5996: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
1.29      crook    5997: @section Defining Words
                   5998: @cindex defining words
                   5999: 
1.47      crook    6000: Defining words are used to extend Forth by creating new entries in the dictionary.
                   6001: 
1.29      crook    6002: @menu
1.67      anton    6003: * CREATE::                      
1.44      crook    6004: * Variables::                   Variables and user variables
1.67      anton    6005: * Constants::                   
1.44      crook    6006: * Values::                      Initialised variables
1.67      anton    6007: * Colon Definitions::           
1.44      crook    6008: * Anonymous Definitions::       Definitions without names
1.69      anton    6009: * Supplying names::             Passing definition names as strings
1.67      anton    6010: * User-defined Defining Words::  
1.44      crook    6011: * Deferred words::              Allow forward references
1.67      anton    6012: * Aliases::                     
1.29      crook    6013: @end menu
                   6014: 
1.44      crook    6015: @node CREATE, Variables, Defining Words, Defining Words
                   6016: @subsection @code{CREATE}
1.29      crook    6017: @cindex simple defining words
                   6018: @cindex defining words, simple
                   6019: 
                   6020: Defining words are used to create new entries in the dictionary. The
                   6021: simplest defining word is @code{CREATE}. @code{CREATE} is used like
                   6022: this:
                   6023: 
                   6024: @example
                   6025: CREATE new-word1
                   6026: @end example
                   6027: 
1.69      anton    6028: @code{CREATE} is a parsing word, i.e., it takes an argument from the
                   6029: input stream (@code{new-word1} in our example).  It generates a
                   6030: dictionary entry for @code{new-word1}. When @code{new-word1} is
                   6031: executed, all that it does is leave an address on the stack. The address
                   6032: represents the value of the data space pointer (@code{HERE}) at the time
                   6033: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
                   6034: associating a name with the address of a region of memory.
1.29      crook    6035: 
1.34      anton    6036: doc-create
                   6037: 
1.69      anton    6038: Note that in ANS Forth guarantees only for @code{create} that its body
                   6039: is in dictionary data space (i.e., where @code{here}, @code{allot}
                   6040: etc. work, @pxref{Dictionary allocation}).  Also, in ANS Forth only
                   6041: @code{create}d words can be modified with @code{does>}
                   6042: (@pxref{User-defined Defining Words}).  And in ANS Forth @code{>body}
                   6043: can only be applied to @code{create}d words.
                   6044: 
1.29      crook    6045: By extending this example to reserve some memory in data space, we end
1.69      anton    6046: up with something like a @i{variable}. Here are two different ways to do
                   6047: it:
1.29      crook    6048: 
                   6049: @example
                   6050: CREATE new-word2 1 cells allot  \ reserve 1 cell - initial value undefined
                   6051: CREATE new-word3 4 ,            \ reserve 1 cell and initialise it (to 4)
                   6052: @end example
                   6053: 
                   6054: The variable can be examined and modified using @code{@@} (``fetch'') and
                   6055: @code{!} (``store'') like this:
                   6056: 
                   6057: @example
                   6058: new-word2 @@ .      \ get address, fetch from it and display
                   6059: 1234 new-word2 !   \ new value, get address, store to it
                   6060: @end example
                   6061: 
1.44      crook    6062: @cindex arrays
                   6063: A similar mechanism can be used to create arrays. For example, an
                   6064: 80-character text input buffer:
1.29      crook    6065: 
                   6066: @example
1.44      crook    6067: CREATE text-buf 80 chars allot
                   6068: 
                   6069: text-buf 0 chars c@@ \ the 1st character (offset 0)
                   6070: text-buf 3 chars c@@ \ the 4th character (offset 3)
                   6071: @end example
1.29      crook    6072: 
1.44      crook    6073: You can build arbitrarily complex data structures by allocating
1.49      anton    6074: appropriate areas of memory. For further discussions of this, and to
1.66      anton    6075: learn about some Gforth tools that make it easier,
1.49      anton    6076: @xref{Structures}.
1.44      crook    6077: 
                   6078: 
                   6079: @node Variables, Constants, CREATE, Defining Words
                   6080: @subsection Variables
                   6081: @cindex variables
                   6082: 
                   6083: The previous section showed how a sequence of commands could be used to
                   6084: generate a variable.  As a final refinement, the whole code sequence can
                   6085: be wrapped up in a defining word (pre-empting the subject of the next
                   6086: section), making it easier to create new variables:
                   6087: 
                   6088: @example
                   6089: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
                   6090: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
                   6091: 
                   6092: myvariableX foo \ variable foo starts off with an unknown value
                   6093: myvariable0 joe \ whilst joe is initialised to 0
1.29      crook    6094: 
                   6095: 45 3 * foo !   \ set foo to 135
                   6096: 1234 joe !     \ set joe to 1234
                   6097: 3 joe +!       \ increment joe by 3.. to 1237
                   6098: @end example
                   6099: 
                   6100: Not surprisingly, there is no need to define @code{myvariable}, since
1.44      crook    6101: Forth already has a definition @code{Variable}. ANS Forth does not
1.69      anton    6102: guarantee that a @code{Variable} is initialised when it is created
                   6103: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
                   6104: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
                   6105: like @code{myvariable0}). Forth also provides @code{2Variable} and
1.47      crook    6106: @code{fvariable} for double and floating-point variables, respectively
1.69      anton    6107: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
1.47      crook    6108: store a boolean, you can use @code{on} and @code{off} to toggle its
                   6109: state.
1.29      crook    6110: 
1.34      anton    6111: doc-variable
                   6112: doc-2variable
                   6113: doc-fvariable
                   6114: 
1.29      crook    6115: @cindex user variables
                   6116: @cindex user space
                   6117: The defining word @code{User} behaves in the same way as @code{Variable}.
                   6118: The difference is that it reserves space in @i{user (data) space} rather
                   6119: than normal data space. In a Forth system that has a multi-tasker, each
                   6120: task has its own set of user variables.
                   6121: 
1.34      anton    6122: doc-user
1.67      anton    6123: @c doc-udp
                   6124: @c doc-uallot
1.34      anton    6125: 
1.29      crook    6126: @comment TODO is that stuff about user variables strictly correct? Is it
                   6127: @comment just terminal tasks that have user variables?
                   6128: @comment should document tasker.fs (with some examples) elsewhere
                   6129: @comment in this manual, then expand on user space and user variables.
                   6130: 
1.44      crook    6131: @node Constants, Values, Variables, Defining Words
                   6132: @subsection Constants
                   6133: @cindex constants
                   6134: 
                   6135: @code{Constant} allows you to declare a fixed value and refer to it by
                   6136: name. For example:
1.29      crook    6137: 
                   6138: @example
                   6139: 12 Constant INCHES-PER-FOOT
                   6140: 3E+08 fconstant SPEED-O-LIGHT
                   6141: @end example
                   6142: 
                   6143: A @code{Variable} can be both read and written, so its run-time
                   6144: behaviour is to supply an address through which its current value can be
                   6145: manipulated. In contrast, the value of a @code{Constant} cannot be
                   6146: changed once it has been declared@footnote{Well, often it can be -- but
                   6147: not in a Standard, portable way. It's safer to use a @code{Value} (read
                   6148: on).} so it's not necessary to supply the address -- it is more
                   6149: efficient to return the value of the constant directly. That's exactly
                   6150: what happens; the run-time effect of a constant is to put its value on
1.49      anton    6151: the top of the stack (You can find one
                   6152: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
1.29      crook    6153: 
1.69      anton    6154: Forth also provides @code{2Constant} and @code{fconstant} for defining
1.29      crook    6155: double and floating-point constants, respectively.
                   6156: 
1.34      anton    6157: doc-constant
                   6158: doc-2constant
                   6159: doc-fconstant
                   6160: 
                   6161: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
1.44      crook    6162: @c nac-> How could that not be true in an ANS Forth? You can't define a
                   6163: @c constant, use it and then delete the definition of the constant..
1.69      anton    6164: 
                   6165: @c anton->An ANS Forth system can compile a constant to a literal; On
                   6166: @c decompilation you would see only the number, just as if it had been used
                   6167: @c in the first place.  The word will stay, of course, but it will only be
                   6168: @c used by the text interpreter (no run-time duties, except when it is 
                   6169: @c POSTPONEd or somesuch).
                   6170: 
                   6171: @c nac:
1.44      crook    6172: @c I agree that it's rather deep, but IMO it is an important difference
                   6173: @c relative to other programming languages.. often it's annoying: it
                   6174: @c certainly changes my programming style relative to C.
                   6175: 
1.69      anton    6176: @c anton: In what way?
                   6177: 
1.29      crook    6178: Constants in Forth behave differently from their equivalents in other
                   6179: programming languages. In other languages, a constant (such as an EQU in
                   6180: assembler or a #define in C) only exists at compile-time; in the
                   6181: executable program the constant has been translated into an absolute
                   6182: number and, unless you are using a symbolic debugger, it's impossible to
                   6183: know what abstract thing that number represents. In Forth a constant has
1.44      crook    6184: an entry in the header space and remains there after the code that uses
                   6185: it has been defined. In fact, it must remain in the dictionary since it
                   6186: has run-time duties to perform. For example:
1.29      crook    6187: 
                   6188: @example
                   6189: 12 Constant INCHES-PER-FOOT
                   6190: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
                   6191: @end example
                   6192: 
                   6193: @cindex in-lining of constants
                   6194: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
                   6195: associated with the constant @code{INCHES-PER-FOOT}. If you use
                   6196: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
                   6197: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
                   6198: attempt to optimise constants by in-lining them where they are used. You
                   6199: can force Gforth to in-line a constant like this:
                   6200: 
                   6201: @example
                   6202: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
                   6203: @end example
                   6204: 
                   6205: If you use @code{see} to decompile @i{this} version of
                   6206: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
1.49      anton    6207: longer present. To understand how this works, read
                   6208: @ref{Interpret/Compile states}, and @ref{Literals}.
1.29      crook    6209: 
                   6210: In-lining constants in this way might improve execution time
                   6211: fractionally, and can ensure that a constant is now only referenced at
                   6212: compile-time. However, the definition of the constant still remains in
                   6213: the dictionary. Some Forth compilers provide a mechanism for controlling
                   6214: a second dictionary for holding transient words such that this second
                   6215: dictionary can be deleted later in order to recover memory
                   6216: space. However, there is no standard way of doing this.
                   6217: 
                   6218: 
1.44      crook    6219: @node Values, Colon Definitions, Constants, Defining Words
                   6220: @subsection Values
                   6221: @cindex values
1.34      anton    6222: 
1.69      anton    6223: A @code{Value} behaves like a @code{Constant}, but it can be changed.
                   6224: @code{TO} is a parsing word that changes a @code{Values}.  In Gforth
                   6225: (not in ANS Forth) you can access (and change) a @code{value} also with
                   6226: @code{>body}.
                   6227: 
                   6228: Here are some
                   6229: examples:
1.29      crook    6230: 
                   6231: @example
1.69      anton    6232: 12 Value APPLES     \ Define APPLES with an initial value of 12
                   6233: 34 TO APPLES        \ Change the value of APPLES. TO is a parsing word
                   6234: 1 ' APPLES >body +! \ Increment APPLES.  Non-standard usage.
                   6235: APPLES              \ puts 35 on the top of the stack.
1.29      crook    6236: @end example
                   6237: 
1.44      crook    6238: doc-value
                   6239: doc-to
1.29      crook    6240: 
1.35      anton    6241: 
1.69      anton    6242: 
1.44      crook    6243: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
                   6244: @subsection Colon Definitions
                   6245: @cindex colon definitions
1.35      anton    6246: 
                   6247: @example
1.44      crook    6248: : name ( ... -- ... )
                   6249:     word1 word2 word3 ;
1.29      crook    6250: @end example
                   6251: 
1.44      crook    6252: @noindent
                   6253: Creates a word called @code{name} that, upon execution, executes
                   6254: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
1.29      crook    6255: 
1.49      anton    6256: The explanation above is somewhat superficial. For simple examples of
                   6257: colon definitions see @ref{Your first definition}.  For an in-depth
1.66      anton    6258: discussion of some of the issues involved, @xref{Interpretation and
1.49      anton    6259: Compilation Semantics}.
1.29      crook    6260: 
1.44      crook    6261: doc-:
                   6262: doc-;
1.1       anton    6263: 
1.34      anton    6264: 
1.69      anton    6265: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
1.44      crook    6266: @subsection Anonymous Definitions
                   6267: @cindex colon definitions
                   6268: @cindex defining words without name
1.34      anton    6269: 
1.44      crook    6270: Sometimes you want to define an @dfn{anonymous word}; a word without a
                   6271: name. You can do this with:
1.1       anton    6272: 
1.44      crook    6273: doc-:noname
1.1       anton    6274: 
1.44      crook    6275: This leaves the execution token for the word on the stack after the
                   6276: closing @code{;}. Here's an example in which a deferred word is
                   6277: initialised with an @code{xt} from an anonymous colon definition:
1.1       anton    6278: 
1.29      crook    6279: @example
1.44      crook    6280: Defer deferred
                   6281: :noname ( ... -- ... )
                   6282:   ... ;
                   6283: IS deferred
1.29      crook    6284: @end example
1.26      crook    6285: 
1.44      crook    6286: @noindent
                   6287: Gforth provides an alternative way of doing this, using two separate
                   6288: words:
1.27      crook    6289: 
1.44      crook    6290: doc-noname
                   6291: @cindex execution token of last defined word
                   6292: doc-lastxt
1.1       anton    6293: 
1.44      crook    6294: @noindent
                   6295: The previous example can be rewritten using @code{noname} and
                   6296: @code{lastxt}:
1.1       anton    6297: 
1.26      crook    6298: @example
1.44      crook    6299: Defer deferred
                   6300: noname : ( ... -- ... )
                   6301:   ... ;
                   6302: lastxt IS deferred
1.26      crook    6303: @end example
1.1       anton    6304: 
1.29      crook    6305: @noindent
1.44      crook    6306: @code{noname} works with any defining word, not just @code{:}.
                   6307: 
                   6308: @code{lastxt} also works when the last word was not defined as
1.71      anton    6309: @code{noname}.  It does not work for combined words, though.  It also has
                   6310: the useful property that is is valid as soon as the header for a
                   6311: definition has been built. Thus:
1.44      crook    6312: 
                   6313: @example
                   6314: lastxt . : foo [ lastxt . ] ; ' foo .
                   6315: @end example
1.1       anton    6316: 
1.44      crook    6317: @noindent
                   6318: prints 3 numbers; the last two are the same.
1.26      crook    6319: 
1.69      anton    6320: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
                   6321: @subsection Supplying the name of a defined word
                   6322: @cindex names for defined words
                   6323: @cindex defining words, name given in a string
                   6324: 
                   6325: By default, a defining word takes the name for the defined word from the
                   6326: input stream. Sometimes you want to supply the name from a string. You
                   6327: can do this with:
                   6328: 
                   6329: doc-nextname
                   6330: 
                   6331: For example:
                   6332: 
                   6333: @example
                   6334: s" foo" nextname create
                   6335: @end example
                   6336: 
                   6337: @noindent
                   6338: is equivalent to:
                   6339: 
                   6340: @example
                   6341: create foo
                   6342: @end example
                   6343: 
                   6344: @noindent
                   6345: @code{nextname} works with any defining word.
                   6346: 
1.1       anton    6347: 
1.69      anton    6348: @node User-defined Defining Words, Deferred words, Supplying names, Defining Words
1.26      crook    6349: @subsection User-defined Defining Words
                   6350: @cindex user-defined defining words
                   6351: @cindex defining words, user-defined
1.1       anton    6352: 
1.29      crook    6353: You can create a new defining word by wrapping defining-time code around
                   6354: an existing defining word and putting the sequence in a colon
1.69      anton    6355: definition. 
                   6356: 
                   6357: @c anton: This example is very complex and leads in a quite different
                   6358: @c direction from the CREATE-DOES> stuff that follows.  It should probably
                   6359: @c be done elsewhere, or as a subsubsection of this subsection (or as a
                   6360: @c subsection of Defining Words)
                   6361: 
                   6362: For example, suppose that you have a word @code{stats} that
1.29      crook    6363: gathers statistics about colon definitions given the @i{xt} of the
                   6364: definition, and you want every colon definition in your application to
                   6365: make a call to @code{stats}. You can define and use a new version of
                   6366: @code{:} like this:
                   6367: 
                   6368: @example
                   6369: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
                   6370:   ... ;  \ other code
                   6371: 
                   6372: : my: : lastxt postpone literal ['] stats compile, ;
                   6373: 
                   6374: my: foo + - ;
                   6375: @end example
                   6376: 
                   6377: When @code{foo} is defined using @code{my:} these steps occur:
                   6378: 
                   6379: @itemize @bullet
                   6380: @item
                   6381: @code{my:} is executed.
                   6382: @item
                   6383: The @code{:} within the definition (the one between @code{my:} and
                   6384: @code{lastxt}) is executed, and does just what it always does; it parses
                   6385: the input stream for a name, builds a dictionary header for the name
                   6386: @code{foo} and switches @code{state} from interpret to compile.
                   6387: @item
                   6388: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
                   6389: being defined -- @code{foo} -- onto the stack.
                   6390: @item
                   6391: The code that was produced by @code{postpone literal} is executed; this
                   6392: causes the value on the stack to be compiled as a literal in the code
                   6393: area of @code{foo}.
                   6394: @item
                   6395: The code @code{['] stats} compiles a literal into the definition of
                   6396: @code{my:}. When @code{compile,} is executed, that literal -- the
                   6397: execution token for @code{stats} -- is layed down in the code area of
                   6398: @code{foo} , following the literal@footnote{Strictly speaking, the
                   6399: mechanism that @code{compile,} uses to convert an @i{xt} into something
                   6400: in the code area is implementation-dependent. A threaded implementation
                   6401: might spit out the execution token directly whilst another
                   6402: implementation might spit out a native code sequence.}.
                   6403: @item
                   6404: At this point, the execution of @code{my:} is complete, and control
                   6405: returns to the text interpreter. The text interpreter is in compile
                   6406: state, so subsequent text @code{+ -} is compiled into the definition of
                   6407: @code{foo} and the @code{;} terminates the definition as always.
                   6408: @end itemize
                   6409: 
                   6410: You can use @code{see} to decompile a word that was defined using
                   6411: @code{my:} and see how it is different from a normal @code{:}
                   6412: definition. For example:
                   6413: 
                   6414: @example
                   6415: : bar + - ;  \ like foo but using : rather than my:
                   6416: see bar
                   6417: : bar
                   6418:   + - ;
                   6419: see foo
                   6420: : foo
                   6421:   107645672 stats + - ;
                   6422: 
                   6423: \ use ' stats . to show that 107645672 is the xt for stats
                   6424: @end example
                   6425: 
                   6426: You can use techniques like this to make new defining words in terms of
                   6427: @i{any} existing defining word.
1.1       anton    6428: 
                   6429: 
1.29      crook    6430: @cindex defining defining words
1.26      crook    6431: @cindex @code{CREATE} ... @code{DOES>}
                   6432: If you want the words defined with your defining words to behave
                   6433: differently from words defined with standard defining words, you can
                   6434: write your defining word like this:
1.1       anton    6435: 
                   6436: @example
1.26      crook    6437: : def-word ( "name" -- )
1.29      crook    6438:     CREATE @i{code1}
1.26      crook    6439: DOES> ( ... -- ... )
1.29      crook    6440:     @i{code2} ;
1.26      crook    6441: 
                   6442: def-word name
1.1       anton    6443: @end example
                   6444: 
1.29      crook    6445: @cindex child words
                   6446: This fragment defines a @dfn{defining word} @code{def-word} and then
                   6447: executes it.  When @code{def-word} executes, it @code{CREATE}s a new
                   6448: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
                   6449: is not executed at this time. The word @code{name} is sometimes called a
                   6450: @dfn{child} of @code{def-word}.
                   6451: 
                   6452: When you execute @code{name}, the address of the body of @code{name} is
                   6453: put on the data stack and @i{code2} is executed (the address of the body
                   6454: of @code{name} is the address @code{HERE} returns immediately after the
1.69      anton    6455: @code{CREATE}, i.e., the address a @code{create}d word returns by
                   6456: default).
                   6457: 
                   6458: @c anton:
                   6459: @c www.dictionary.com says:
                   6460: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
                   6461: @c several generations of absence, usually caused by the chance
                   6462: @c recombination of genes.  2.An individual or a part that exhibits
                   6463: @c atavism. Also called throwback.  3.The return of a trait or recurrence
                   6464: @c of previous behavior after a period of absence.
                   6465: @c
                   6466: @c Doesn't seem to fit.
1.29      crook    6467: 
1.69      anton    6468: @c @cindex atavism in child words
1.33      anton    6469: You can use @code{def-word} to define a set of child words that behave
1.69      anton    6470: similarly; they all have a common run-time behaviour determined by
                   6471: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
                   6472: body of the child word. The structure of the data is common to all
                   6473: children of @code{def-word}, but the data values are specific -- and
                   6474: private -- to each child word. When a child word is executed, the
                   6475: address of its private data area is passed as a parameter on TOS to be
                   6476: used and manipulated@footnote{It is legitimate both to read and write to
                   6477: this data area.} by @i{code2}.
1.29      crook    6478: 
                   6479: The two fragments of code that make up the defining words act (are
                   6480: executed) at two completely separate times:
1.1       anton    6481: 
1.29      crook    6482: @itemize @bullet
                   6483: @item
                   6484: At @i{define time}, the defining word executes @i{code1} to generate a
                   6485: child word
                   6486: @item
                   6487: At @i{child execution time}, when a child word is invoked, @i{code2}
                   6488: is executed, using parameters (data) that are private and specific to
                   6489: the child word.
                   6490: @end itemize
                   6491: 
1.44      crook    6492: Another way of understanding the behaviour of @code{def-word} and
                   6493: @code{name} is to say that, if you make the following definitions:
1.33      anton    6494: @example
                   6495: : def-word1 ( "name" -- )
                   6496:     CREATE @i{code1} ;
                   6497: 
                   6498: : action1 ( ... -- ... )
                   6499:     @i{code2} ;
                   6500: 
                   6501: def-word1 name1
                   6502: @end example
                   6503: 
1.44      crook    6504: @noindent
                   6505: Then using @code{name1 action1} is equivalent to using @code{name}.
1.1       anton    6506: 
1.29      crook    6507: The classic example is that you can define @code{CONSTANT} in this way:
1.26      crook    6508: 
1.1       anton    6509: @example
1.29      crook    6510: : CONSTANT ( w "name" -- )
                   6511:     CREATE ,
1.26      crook    6512: DOES> ( -- w )
                   6513:     @@ ;
1.1       anton    6514: @end example
                   6515: 
1.29      crook    6516: @comment There is a beautiful description of how this works and what
                   6517: @comment it does in the Forthwrite 100th edition.. as well as an elegant
                   6518: @comment commentary on the Counting Fruits problem.
                   6519: 
                   6520: When you create a constant with @code{5 CONSTANT five}, a set of
                   6521: define-time actions take place; first a new word @code{five} is created,
                   6522: then the value 5 is laid down in the body of @code{five} with
1.44      crook    6523: @code{,}. When @code{five} is executed, the address of the body is put on
1.29      crook    6524: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
                   6525: no code of its own; it simply contains a data field and a pointer to the
                   6526: code that follows @code{DOES>} in its defining word. That makes words
                   6527: created in this way very compact.
                   6528: 
                   6529: The final example in this section is intended to remind you that space
                   6530: reserved in @code{CREATE}d words is @i{data} space and therefore can be
                   6531: both read and written by a Standard program@footnote{Exercise: use this
                   6532: example as a starting point for your own implementation of @code{Value}
                   6533: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
                   6534: @code{[']}.}:
                   6535: 
                   6536: @example
                   6537: : foo ( "name" -- )
                   6538:     CREATE -1 ,
                   6539: DOES> ( -- )
1.33      anton    6540:     @@ . ;
1.29      crook    6541: 
                   6542: foo first-word
                   6543: foo second-word
                   6544: 
                   6545: 123 ' first-word >BODY !
                   6546: @end example
                   6547: 
                   6548: If @code{first-word} had been a @code{CREATE}d word, we could simply
                   6549: have executed it to get the address of its data field. However, since it
                   6550: was defined to have @code{DOES>} actions, its execution semantics are to
                   6551: perform those @code{DOES>} actions. To get the address of its data field
                   6552: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
                   6553: translate the xt into the address of the data field.  When you execute
                   6554: @code{first-word}, it will display @code{123}. When you execute
                   6555: @code{second-word} it will display @code{-1}.
1.26      crook    6556: 
                   6557: @cindex stack effect of @code{DOES>}-parts
                   6558: @cindex @code{DOES>}-parts, stack effect
1.29      crook    6559: In the examples above the stack comment after the @code{DOES>} specifies
1.26      crook    6560: the stack effect of the defined words, not the stack effect of the
                   6561: following code (the following code expects the address of the body on
                   6562: the top of stack, which is not reflected in the stack comment). This is
                   6563: the convention that I use and recommend (it clashes a bit with using
                   6564: locals declarations for stack effect specification, though).
1.1       anton    6565: 
1.53      anton    6566: @menu
                   6567: * CREATE..DOES> applications::  
                   6568: * CREATE..DOES> details::       
1.63      anton    6569: * Advanced does> usage example::  
1.91    ! anton    6570: * @code{Const-does>}::          
1.53      anton    6571: @end menu
                   6572: 
                   6573: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
1.26      crook    6574: @subsubsection Applications of @code{CREATE..DOES>}
                   6575: @cindex @code{CREATE} ... @code{DOES>}, applications
1.1       anton    6576: 
1.26      crook    6577: You may wonder how to use this feature. Here are some usage patterns:
1.1       anton    6578: 
1.26      crook    6579: @cindex factoring similar colon definitions
                   6580: When you see a sequence of code occurring several times, and you can
                   6581: identify a meaning, you will factor it out as a colon definition. When
                   6582: you see similar colon definitions, you can factor them using
                   6583: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
                   6584: that look very similar:
1.1       anton    6585: @example
1.26      crook    6586: : ori, ( reg-target reg-source n -- )
                   6587:     0 asm-reg-reg-imm ;
                   6588: : andi, ( reg-target reg-source n -- )
                   6589:     1 asm-reg-reg-imm ;
1.1       anton    6590: @end example
                   6591: 
1.26      crook    6592: @noindent
                   6593: This could be factored with:
                   6594: @example
                   6595: : reg-reg-imm ( op-code -- )
                   6596:     CREATE ,
                   6597: DOES> ( reg-target reg-source n -- )
                   6598:     @@ asm-reg-reg-imm ;
                   6599: 
                   6600: 0 reg-reg-imm ori,
                   6601: 1 reg-reg-imm andi,
                   6602: @end example
1.1       anton    6603: 
1.26      crook    6604: @cindex currying
                   6605: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
                   6606: supply a part of the parameters for a word (known as @dfn{currying} in
                   6607: the functional language community). E.g., @code{+} needs two
                   6608: parameters. Creating versions of @code{+} with one parameter fixed can
                   6609: be done like this:
1.82      anton    6610: 
1.1       anton    6611: @example
1.82      anton    6612: : curry+ ( n1 "name" -- )
1.26      crook    6613:     CREATE ,
                   6614: DOES> ( n2 -- n1+n2 )
                   6615:     @@ + ;
                   6616: 
                   6617:  3 curry+ 3+
                   6618: -2 curry+ 2-
1.1       anton    6619: @end example
                   6620: 
1.91    ! anton    6621: 
1.63      anton    6622: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
1.26      crook    6623: @subsubsection The gory details of @code{CREATE..DOES>}
                   6624: @cindex @code{CREATE} ... @code{DOES>}, details
1.1       anton    6625: 
1.26      crook    6626: doc-does>
1.1       anton    6627: 
1.26      crook    6628: @cindex @code{DOES>} in a separate definition
                   6629: This means that you need not use @code{CREATE} and @code{DOES>} in the
                   6630: same definition; you can put the @code{DOES>}-part in a separate
1.29      crook    6631: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
1.26      crook    6632: @example
                   6633: : does1 
                   6634: DOES> ( ... -- ... )
1.44      crook    6635:     ... ;
                   6636: 
                   6637: : does2
                   6638: DOES> ( ... -- ... )
                   6639:     ... ;
                   6640: 
                   6641: : def-word ( ... -- ... )
                   6642:     create ...
                   6643:     IF
                   6644:        does1
                   6645:     ELSE
                   6646:        does2
                   6647:     ENDIF ;
                   6648: @end example
                   6649: 
                   6650: In this example, the selection of whether to use @code{does1} or
1.69      anton    6651: @code{does2} is made at definition-time; at the time that the child word is
1.44      crook    6652: @code{CREATE}d.
                   6653: 
                   6654: @cindex @code{DOES>} in interpretation state
                   6655: In a standard program you can apply a @code{DOES>}-part only if the last
                   6656: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
                   6657: will override the behaviour of the last word defined in any case. In a
                   6658: standard program, you can use @code{DOES>} only in a colon
                   6659: definition. In Gforth, you can also use it in interpretation state, in a
                   6660: kind of one-shot mode; for example:
                   6661: @example
                   6662: CREATE name ( ... -- ... )
                   6663:   @i{initialization}
                   6664: DOES>
                   6665:   @i{code} ;
                   6666: @end example
                   6667: 
                   6668: @noindent
                   6669: is equivalent to the standard:
                   6670: @example
                   6671: :noname
                   6672: DOES>
                   6673:     @i{code} ;
                   6674: CREATE name EXECUTE ( ... -- ... )
                   6675:     @i{initialization}
                   6676: @end example
                   6677: 
1.53      anton    6678: doc->body
                   6679: 
1.91    ! anton    6680: @node Advanced does> usage example, @code{Const-does>}, CREATE..DOES> details, User-defined Defining Words
1.63      anton    6681: @subsubsection Advanced does> usage example
                   6682: 
                   6683: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
                   6684: for disassembling instructions, that follow a very repetetive scheme:
                   6685: 
                   6686: @example
                   6687: :noname @var{disasm-operands} s" @var{inst-name}" type ;
                   6688: @var{entry-num} cells @var{table} + !
                   6689: @end example
                   6690: 
                   6691: Of course, this inspires the idea to factor out the commonalities to
                   6692: allow a definition like
                   6693: 
                   6694: @example
                   6695: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
                   6696: @end example
                   6697: 
                   6698: The parameters @var{disasm-operands} and @var{table} are usually
1.69      anton    6699: correlated.  Moreover, before I wrote the disassembler, there already
                   6700: existed code that defines instructions like this:
1.63      anton    6701: 
                   6702: @example
                   6703: @var{entry-num} @var{inst-format} @var{inst-name}
                   6704: @end example
                   6705: 
                   6706: This code comes from the assembler and resides in
                   6707: @file{arch/mips/insts.fs}.
                   6708: 
                   6709: So I had to define the @var{inst-format} words that performed the scheme
                   6710: above when executed.  At first I chose to use run-time code-generation:
                   6711: 
                   6712: @example
                   6713: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
                   6714:   :noname Postpone @var{disasm-operands}
                   6715:   name Postpone sliteral Postpone type Postpone ;
                   6716:   swap cells @var{table} + ! ;
                   6717: @end example
                   6718: 
                   6719: Note that this supplies the other two parameters of the scheme above.
1.44      crook    6720: 
1.63      anton    6721: An alternative would have been to write this using
                   6722: @code{create}/@code{does>}:
                   6723: 
                   6724: @example
                   6725: : @var{inst-format} ( entry-num "name" -- )
                   6726:   here name string, ( entry-num c-addr ) \ parse and save "name"
                   6727:   noname create , ( entry-num )
                   6728:   lastxt swap cells @var{table} + !
                   6729: does> ( addr w -- )
                   6730:   \ disassemble instruction w at addr
                   6731:   @@ >r 
                   6732:   @var{disasm-operands}
                   6733:   r> count type ;
                   6734: @end example
                   6735: 
                   6736: Somehow the first solution is simpler, mainly because it's simpler to
                   6737: shift a string from definition-time to use-time with @code{sliteral}
                   6738: than with @code{string,} and friends.
                   6739: 
                   6740: I wrote a lot of words following this scheme and soon thought about
                   6741: factoring out the commonalities among them.  Note that this uses a
                   6742: two-level defining word, i.e., a word that defines ordinary defining
                   6743: words.
                   6744: 
                   6745: This time a solution involving @code{postpone} and friends seemed more
                   6746: difficult (try it as an exercise), so I decided to use a
                   6747: @code{create}/@code{does>} word; since I was already at it, I also used
                   6748: @code{create}/@code{does>} for the lower level (try using
                   6749: @code{postpone} etc. as an exercise), resulting in the following
                   6750: definition:
                   6751: 
                   6752: @example
                   6753: : define-format ( disasm-xt table-xt -- )
                   6754:     \ define an instruction format that uses disasm-xt for
                   6755:     \ disassembling and enters the defined instructions into table
                   6756:     \ table-xt
                   6757:     create 2,
                   6758: does> ( u "inst" -- )
                   6759:     \ defines an anonymous word for disassembling instruction inst,
                   6760:     \ and enters it as u-th entry into table-xt
                   6761:     2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
                   6762:     noname create 2,      \ define anonymous word
                   6763:     execute lastxt swap ! \ enter xt of defined word into table-xt
                   6764: does> ( addr w -- )
                   6765:     \ disassemble instruction w at addr
                   6766:     2@@ >r ( addr w disasm-xt R: c-addr )
                   6767:     execute ( R: c-addr ) \ disassemble operands
                   6768:     r> count type ; \ print name 
                   6769: @end example
                   6770: 
                   6771: Note that the tables here (in contrast to above) do the @code{cells +}
                   6772: by themselves (that's why you have to pass an xt).  This word is used in
                   6773: the following way:
                   6774: 
                   6775: @example
                   6776: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
                   6777: @end example
                   6778: 
1.71      anton    6779: As shown above, the defined instruction format is then used like this:
                   6780: 
                   6781: @example
                   6782: @var{entry-num} @var{inst-format} @var{inst-name}
                   6783: @end example
                   6784: 
1.63      anton    6785: In terms of currying, this kind of two-level defining word provides the
                   6786: parameters in three stages: first @var{disasm-operands} and @var{table},
                   6787: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
                   6788: the instruction to be disassembled.  
                   6789: 
                   6790: Of course this did not quite fit all the instruction format names used
                   6791: in @file{insts.fs}, so I had to define a few wrappers that conditioned
                   6792: the parameters into the right form.
                   6793: 
                   6794: If you have trouble following this section, don't worry.  First, this is
                   6795: involved and takes time (and probably some playing around) to
                   6796: understand; second, this is the first two-level
                   6797: @code{create}/@code{does>} word I have written in seventeen years of
                   6798: Forth; and if I did not have @file{insts.fs} to start with, I may well
                   6799: have elected to use just a one-level defining word (with some repeating
                   6800: of parameters when using the defining word). So it is not necessary to
                   6801: understand this, but it may improve your understanding of Forth.
1.44      crook    6802: 
                   6803: 
1.91    ! anton    6804: @node @code{Const-does>},  , Advanced does> usage example, User-defined Defining Words
        !          6805: @subsubsection @code{Const-does>}
        !          6806: 
        !          6807: A frequent use of @code{create}...@code{does>} is for transferring some
        !          6808: values from definition-time to run-time.  Gforth supports this use with
        !          6809: 
        !          6810: doc-const-does>
        !          6811: 
        !          6812: A typical use of this word is:
        !          6813: 
        !          6814: @example
        !          6815: : curry+ ( n1 "name" -- )
        !          6816: 1 0 CONST-DOES> ( n2 -- n1+n2 )
        !          6817:     + ;
        !          6818: 
        !          6819: 3 curry+ 3+
        !          6820: @end example
        !          6821: 
        !          6822: Here the @code{1 0} means that 1 cell and 0 floats are transferred from
        !          6823: definition to run-time.
        !          6824: 
        !          6825: The advantages of using @code{const-does>} are:
        !          6826: 
        !          6827: @itemize
        !          6828: 
        !          6829: @item
        !          6830: You don't have to deal with storing and retrieving the values, i.e.,
        !          6831: your program becomes more writable and readable.
        !          6832: 
        !          6833: @item
        !          6834: When using @code{does>}, you have to introduce a @code{@@} that cannot
        !          6835: be optimized away (because you could change the data using
        !          6836: @code{>body}...@code{!}); @code{const-does>} avoids this problem.
        !          6837: 
        !          6838: @end itemize
        !          6839: 
        !          6840: An ANS Forth implementation of @code{const-does>} is available in
        !          6841: @file{compat/const-does.fs}.
        !          6842: 
        !          6843: 
1.44      crook    6844: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
                   6845: @subsection Deferred words
                   6846: @cindex deferred words
                   6847: 
                   6848: The defining word @code{Defer} allows you to define a word by name
                   6849: without defining its behaviour; the definition of its behaviour is
                   6850: deferred. Here are two situation where this can be useful:
                   6851: 
                   6852: @itemize @bullet
                   6853: @item
                   6854: Where you want to allow the behaviour of a word to be altered later, and
                   6855: for all precompiled references to the word to change when its behaviour
                   6856: is changed.
                   6857: @item
                   6858: For mutual recursion; @xref{Calls and returns}.
                   6859: @end itemize
                   6860: 
                   6861: In the following example, @code{foo} always invokes the version of
                   6862: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
                   6863: always invokes the version that prints ``@code{Hello}''. There is no way
                   6864: of getting @code{foo} to use the later version without re-ordering the
                   6865: source code and recompiling it.
                   6866: 
                   6867: @example
                   6868: : greet ." Good morning" ;
                   6869: : foo ... greet ... ;
                   6870: : greet ." Hello" ;
                   6871: : bar ... greet ... ;
                   6872: @end example
                   6873: 
                   6874: This problem can be solved by defining @code{greet} as a @code{Defer}red
                   6875: word. The behaviour of a @code{Defer}red word can be defined and
                   6876: redefined at any time by using @code{IS} to associate the xt of a
                   6877: previously-defined word with it. The previous example becomes:
                   6878: 
                   6879: @example
1.69      anton    6880: Defer greet ( -- )
1.44      crook    6881: : foo ... greet ... ;
                   6882: : bar ... greet ... ;
1.69      anton    6883: : greet1 ( -- ) ." Good morning" ;
                   6884: : greet2 ( -- ) ." Hello" ;
1.44      crook    6885: ' greet2 <IS> greet  \ make greet behave like greet2
                   6886: @end example
                   6887: 
1.69      anton    6888: @progstyle
                   6889: You should write a stack comment for every deferred word, and put only
                   6890: XTs into deferred words that conform to this stack effect.  Otherwise
                   6891: it's too difficult to use the deferred word.
                   6892: 
1.44      crook    6893: A deferred word can be used to improve the statistics-gathering example
                   6894: from @ref{User-defined Defining Words}; rather than edit the
                   6895: application's source code to change every @code{:} to a @code{my:}, do
                   6896: this:
                   6897: 
                   6898: @example
                   6899: : real: : ;     \ retain access to the original
                   6900: defer :         \ redefine as a deferred word
1.69      anton    6901: ' my: <IS> :      \ use special version of :
1.44      crook    6902: \
                   6903: \ load application here
                   6904: \
1.69      anton    6905: ' real: <IS> :    \ go back to the original
1.44      crook    6906: @end example
                   6907: 
                   6908: 
                   6909: One thing to note is that @code{<IS>} consumes its name when it is
                   6910: executed.  If you want to specify the name at compile time, use
                   6911: @code{[IS]}:
                   6912: 
                   6913: @example
                   6914: : set-greet ( xt -- )
                   6915:   [IS] greet ;
                   6916: 
                   6917: ' greet1 set-greet
                   6918: @end example
                   6919: 
1.69      anton    6920: A deferred word can only inherit execution semantics from the xt
                   6921: (because that is all that an xt can represent -- for more discussion of
                   6922: this @pxref{Tokens for Words}); by default it will have default
                   6923: interpretation and compilation semantics deriving from this execution
                   6924: semantics.  However, you can change the interpretation and compilation
                   6925: semantics of the deferred word in the usual ways:
1.44      crook    6926: 
                   6927: @example
                   6928: : bar .... ; compile-only
                   6929: Defer fred immediate
                   6930: Defer jim
                   6931: 
                   6932: ' bar <IS> jim  \ jim has default semantics
                   6933: ' bar <IS> fred \ fred is immediate
                   6934: @end example
                   6935: 
                   6936: doc-defer
                   6937: doc-<is>
                   6938: doc-[is]
                   6939: doc-is
                   6940: @comment TODO document these: what's defers [is]
                   6941: doc-what's
                   6942: doc-defers
                   6943: 
                   6944: @c Use @code{words-deferred} to see a list of deferred words.
                   6945: 
                   6946: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
                   6947: are provided in @file{compat/defer.fs}.
                   6948: 
                   6949: 
1.69      anton    6950: @node Aliases,  , Deferred words, Defining Words
1.44      crook    6951: @subsection Aliases
                   6952: @cindex aliases
1.1       anton    6953: 
1.44      crook    6954: The defining word @code{Alias} allows you to define a word by name that
                   6955: has the same behaviour as some other word. Here are two situation where
                   6956: this can be useful:
1.1       anton    6957: 
1.44      crook    6958: @itemize @bullet
                   6959: @item
                   6960: When you want access to a word's definition from a different word list
                   6961: (for an example of this, see the definition of the @code{Root} word list
                   6962: in the Gforth source).
                   6963: @item
                   6964: When you want to create a synonym; a definition that can be known by
                   6965: either of two names (for example, @code{THEN} and @code{ENDIF} are
                   6966: aliases).
                   6967: @end itemize
1.1       anton    6968: 
1.69      anton    6969: Like deferred words, an alias has default compilation and interpretation
                   6970: semantics at the beginning (not the modifications of the other word),
                   6971: but you can change them in the usual ways (@code{immediate},
                   6972: @code{compile-only}). For example:
1.1       anton    6973: 
                   6974: @example
1.44      crook    6975: : foo ... ; immediate
                   6976: 
                   6977: ' foo Alias bar \ bar is not an immediate word
                   6978: ' foo Alias fooby immediate \ fooby is an immediate word
1.1       anton    6979: @end example
                   6980: 
1.44      crook    6981: Words that are aliases have the same xt, different headers in the
                   6982: dictionary, and consequently different name tokens (@pxref{Tokens for
                   6983: Words}) and possibly different immediate flags.  An alias can only have
                   6984: default or immediate compilation semantics; you can define aliases for
                   6985: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
1.1       anton    6986: 
1.44      crook    6987: doc-alias
1.1       anton    6988: 
                   6989: 
1.47      crook    6990: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
                   6991: @section Interpretation and Compilation Semantics
1.26      crook    6992: @cindex semantics, interpretation and compilation
1.1       anton    6993: 
1.71      anton    6994: @c !! state and ' are used without explanation
                   6995: @c example for immediate/compile-only? or is the tutorial enough
                   6996: 
1.26      crook    6997: @cindex interpretation semantics
1.71      anton    6998: The @dfn{interpretation semantics} of a (named) word are what the text
1.26      crook    6999: interpreter does when it encounters the word in interpret state. It also
                   7000: appears in some other contexts, e.g., the execution token returned by
1.71      anton    7001: @code{' @i{word}} identifies the interpretation semantics of @i{word}
                   7002: (in other words, @code{' @i{word} execute} is equivalent to
1.29      crook    7003: interpret-state text interpretation of @code{@i{word}}).
1.1       anton    7004: 
1.26      crook    7005: @cindex compilation semantics
1.71      anton    7006: The @dfn{compilation semantics} of a (named) word are what the text
                   7007: interpreter does when it encounters the word in compile state. It also
                   7008: appears in other contexts, e.g, @code{POSTPONE @i{word}}
                   7009: compiles@footnote{In standard terminology, ``appends to the current
                   7010: definition''.} the compilation semantics of @i{word}.
1.1       anton    7011: 
1.26      crook    7012: @cindex execution semantics
                   7013: The standard also talks about @dfn{execution semantics}. They are used
                   7014: only for defining the interpretation and compilation semantics of many
                   7015: words. By default, the interpretation semantics of a word are to
                   7016: @code{execute} its execution semantics, and the compilation semantics of
                   7017: a word are to @code{compile,} its execution semantics.@footnote{In
                   7018: standard terminology: The default interpretation semantics are its
                   7019: execution semantics; the default compilation semantics are to append its
                   7020: execution semantics to the execution semantics of the current
                   7021: definition.}
                   7022: 
1.71      anton    7023: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
                   7024: the text interpreter, ticked, or @code{postpone}d, so they have no
                   7025: interpretation or compilation semantics.  Their behaviour is represented
                   7026: by their XT (@pxref{Tokens for Words}), and we call it execution
                   7027: semantics, too.
                   7028: 
1.26      crook    7029: @comment TODO expand, make it co-operate with new sections on text interpreter.
                   7030: 
                   7031: @cindex immediate words
                   7032: @cindex compile-only words
                   7033: You can change the semantics of the most-recently defined word:
                   7034: 
1.44      crook    7035: 
1.26      crook    7036: doc-immediate
                   7037: doc-compile-only
                   7038: doc-restrict
                   7039: 
1.82      anton    7040: By convention, words with non-default compilation semantics (e.g.,
                   7041: immediate words) often have names surrounded with brackets (e.g.,
                   7042: @code{[']}, @pxref{Execution token}).
1.44      crook    7043: 
1.26      crook    7044: Note that ticking (@code{'}) a compile-only word gives an error
                   7045: (``Interpreting a compile-only word'').
1.1       anton    7046: 
1.47      crook    7047: @menu
1.67      anton    7048: * Combined words::              
1.47      crook    7049: @end menu
1.44      crook    7050: 
1.71      anton    7051: 
1.48      anton    7052: @node Combined words,  , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
1.44      crook    7053: @subsection Combined Words
                   7054: @cindex combined words
                   7055: 
                   7056: Gforth allows you to define @dfn{combined words} -- words that have an
                   7057: arbitrary combination of interpretation and compilation semantics.
                   7058: 
1.26      crook    7059: doc-interpret/compile:
1.1       anton    7060: 
1.26      crook    7061: This feature was introduced for implementing @code{TO} and @code{S"}. I
                   7062: recommend that you do not define such words, as cute as they may be:
                   7063: they make it hard to get at both parts of the word in some contexts.
                   7064: E.g., assume you want to get an execution token for the compilation
                   7065: part. Instead, define two words, one that embodies the interpretation
                   7066: part, and one that embodies the compilation part.  Once you have done
                   7067: that, you can define a combined word with @code{interpret/compile:} for
                   7068: the convenience of your users.
1.1       anton    7069: 
1.26      crook    7070: You might try to use this feature to provide an optimizing
                   7071: implementation of the default compilation semantics of a word. For
                   7072: example, by defining:
1.1       anton    7073: @example
1.26      crook    7074: :noname
                   7075:    foo bar ;
                   7076: :noname
                   7077:    POSTPONE foo POSTPONE bar ;
1.29      crook    7078: interpret/compile: opti-foobar
1.1       anton    7079: @end example
1.26      crook    7080: 
1.23      crook    7081: @noindent
1.26      crook    7082: as an optimizing version of:
                   7083: 
1.1       anton    7084: @example
1.26      crook    7085: : foobar
                   7086:     foo bar ;
1.1       anton    7087: @end example
                   7088: 
1.26      crook    7089: Unfortunately, this does not work correctly with @code{[compile]},
                   7090: because @code{[compile]} assumes that the compilation semantics of all
                   7091: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
1.29      crook    7092: opti-foobar} would compile compilation semantics, whereas
                   7093: @code{[compile] foobar} would compile interpretation semantics.
1.1       anton    7094: 
1.26      crook    7095: @cindex state-smart words (are a bad idea)
1.82      anton    7096: @anchor{state-smartness}
1.29      crook    7097: Some people try to use @dfn{state-smart} words to emulate the feature provided
1.26      crook    7098: by @code{interpret/compile:} (words are state-smart if they check
                   7099: @code{STATE} during execution). E.g., they would try to code
                   7100: @code{foobar} like this:
1.1       anton    7101: 
1.26      crook    7102: @example
                   7103: : foobar
                   7104:   STATE @@
                   7105:   IF ( compilation state )
                   7106:     POSTPONE foo POSTPONE bar
                   7107:   ELSE
                   7108:     foo bar
                   7109:   ENDIF ; immediate
                   7110: @end example
1.1       anton    7111: 
1.26      crook    7112: Although this works if @code{foobar} is only processed by the text
                   7113: interpreter, it does not work in other contexts (like @code{'} or
                   7114: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
                   7115: for a state-smart word, not for the interpretation semantics of the
                   7116: original @code{foobar}; when you execute this execution token (directly
                   7117: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
                   7118: state, the result will not be what you expected (i.e., it will not
                   7119: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
                   7120: write them@footnote{For a more detailed discussion of this topic, see
1.66      anton    7121: M. Anton Ertl,
                   7122: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
                   7123: it is Evil and How to Exorcise it}}, EuroForth '98.}!
1.1       anton    7124: 
1.26      crook    7125: @cindex defining words with arbitrary semantics combinations
                   7126: It is also possible to write defining words that define words with
                   7127: arbitrary combinations of interpretation and compilation semantics. In
                   7128: general, they look like this:
1.1       anton    7129: 
1.26      crook    7130: @example
                   7131: : def-word
                   7132:     create-interpret/compile
1.29      crook    7133:     @i{code1}
1.26      crook    7134: interpretation>
1.29      crook    7135:     @i{code2}
1.26      crook    7136: <interpretation
                   7137: compilation>
1.29      crook    7138:     @i{code3}
1.26      crook    7139: <compilation ;
                   7140: @end example
1.1       anton    7141: 
1.29      crook    7142: For a @i{word} defined with @code{def-word}, the interpretation
                   7143: semantics are to push the address of the body of @i{word} and perform
                   7144: @i{code2}, and the compilation semantics are to push the address of
                   7145: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
1.26      crook    7146: can also be defined like this (except that the defined constants don't
                   7147: behave correctly when @code{[compile]}d):
1.1       anton    7148: 
1.26      crook    7149: @example
                   7150: : constant ( n "name" -- )
                   7151:     create-interpret/compile
                   7152:     ,
                   7153: interpretation> ( -- n )
                   7154:     @@
                   7155: <interpretation
                   7156: compilation> ( compilation. -- ; run-time. -- n )
                   7157:     @@ postpone literal
                   7158: <compilation ;
                   7159: @end example
1.1       anton    7160: 
1.44      crook    7161: 
1.26      crook    7162: doc-create-interpret/compile
                   7163: doc-interpretation>
                   7164: doc-<interpretation
                   7165: doc-compilation>
                   7166: doc-<compilation
1.1       anton    7167: 
1.44      crook    7168: 
1.29      crook    7169: Words defined with @code{interpret/compile:} and
1.26      crook    7170: @code{create-interpret/compile} have an extended header structure that
                   7171: differs from other words; however, unless you try to access them with
                   7172: plain address arithmetic, you should not notice this. Words for
                   7173: accessing the header structure usually know how to deal with this; e.g.,
1.29      crook    7174: @code{'} @i{word} @code{>body} also gives you the body of a word created
                   7175: with @code{create-interpret/compile}.
1.1       anton    7176: 
1.44      crook    7177: 
1.47      crook    7178: @c -------------------------------------------------------------
1.81      anton    7179: @node Tokens for Words, Compiling words, Interpretation and Compilation Semantics, Words
1.47      crook    7180: @section Tokens for Words
                   7181: @cindex tokens for words
                   7182: 
                   7183: This section describes the creation and use of tokens that represent
                   7184: words.
                   7185: 
1.71      anton    7186: @menu
                   7187: * Execution token::             represents execution/interpretation semantics
                   7188: * Compilation token::           represents compilation semantics
                   7189: * Name token::                  represents named words
                   7190: @end menu
1.47      crook    7191: 
1.71      anton    7192: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
                   7193: @subsection Execution token
1.47      crook    7194: 
                   7195: @cindex xt
                   7196: @cindex execution token
1.71      anton    7197: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
                   7198: You can use @code{execute} to invoke this behaviour.
1.47      crook    7199: 
1.71      anton    7200: @cindex tick (')
                   7201: You can use @code{'} to get an execution token that represents the
                   7202: interpretation semantics of a named word:
1.47      crook    7203: 
                   7204: @example
1.71      anton    7205: 5 ' .
                   7206: execute
                   7207: @end example
1.47      crook    7208: 
1.71      anton    7209: doc-'
                   7210: 
                   7211: @code{'} parses at run-time; there is also a word @code{[']} that parses
                   7212: when it is compiled, and compiles the resulting XT:
                   7213: 
                   7214: @example
                   7215: : foo ['] . execute ;
                   7216: 5 foo
                   7217: : bar ' execute ; \ by contrast,
                   7218: 5 bar .           \ ' parses "." when bar executes
                   7219: @end example
                   7220: 
                   7221: doc-[']
                   7222: 
                   7223: If you want the execution token of @i{word}, write @code{['] @i{word}}
                   7224: in compiled code and @code{' @i{word}} in interpreted code.  Gforth's
                   7225: @code{'} and @code{[']} behave somewhat unusually by complaining about
                   7226: compile-only words (because these words have no interpretation
                   7227: semantics).  You might get what you want by using @code{COMP' @i{word}
                   7228: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
                   7229: token}).
                   7230: 
                   7231: Another way to get an XT is @code{:noname} or @code{lastxt}
                   7232: (@pxref{Anonymous Definitions}).  For anonymous words this gives an xt
                   7233: for the only behaviour the word has (the execution semantics).  For
                   7234: named words, @code{lastxt} produces an XT for the same behaviour it
                   7235: would produce if the word was defined anonymously.
                   7236: 
                   7237: @example
                   7238: :noname ." hello" ;
                   7239: execute
1.47      crook    7240: @end example
                   7241: 
1.71      anton    7242: An XT occupies one cell and can be manipulated like any other cell.
                   7243: 
1.47      crook    7244: @cindex code field address
                   7245: @cindex CFA
1.71      anton    7246: In ANS Forth the XT is just an abstract data type (i.e., defined by the
                   7247: operations that produce or consume it).  For old hands: In Gforth, the
                   7248: XT is implemented as a code field address (CFA).
                   7249: 
                   7250: doc-execute
                   7251: doc-perform
                   7252: 
                   7253: @node Compilation token, Name token, Execution token, Tokens for Words
                   7254: @subsection Compilation token
1.47      crook    7255: 
                   7256: @cindex compilation token
1.71      anton    7257: @cindex CT (compilation token)
                   7258: Gforth represents the compilation semantics of a named word by a
1.47      crook    7259: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
                   7260: @i{xt} is an execution token. The compilation semantics represented by
                   7261: the compilation token can be performed with @code{execute}, which
                   7262: consumes the whole compilation token, with an additional stack effect
                   7263: determined by the represented compilation semantics.
                   7264: 
                   7265: At present, the @i{w} part of a compilation token is an execution token,
                   7266: and the @i{xt} part represents either @code{execute} or
                   7267: @code{compile,}@footnote{Depending upon the compilation semantics of the
                   7268: word. If the word has default compilation semantics, the @i{xt} will
                   7269: represent @code{compile,}. Otherwise (e.g., for immediate words), the
                   7270: @i{xt} will represent @code{execute}.}. However, don't rely on that
                   7271: knowledge, unless necessary; future versions of Gforth may introduce
                   7272: unusual compilation tokens (e.g., a compilation token that represents
                   7273: the compilation semantics of a literal).
                   7274: 
1.71      anton    7275: You can perform the compilation semantics represented by the compilation
                   7276: token with @code{execute}.  You can compile the compilation semantics
                   7277: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
                   7278: equivalent to @code{postpone @i{word}}.
                   7279: 
                   7280: doc-[comp']
                   7281: doc-comp'
                   7282: doc-postpone,
                   7283: 
                   7284: @node Name token,  , Compilation token, Tokens for Words
                   7285: @subsection Name token
1.47      crook    7286: 
                   7287: @cindex name token
                   7288: @cindex name field address
                   7289: @cindex NFA
1.71      anton    7290: Gforth represents named words by the @dfn{name token}, (@i{nt}). In
1.47      crook    7291: Gforth, the abstract data type @emph{name token} is implemented as a
                   7292: name field address (NFA).
                   7293: 
                   7294: doc-find-name
                   7295: doc-name>int
                   7296: doc-name?int
                   7297: doc-name>comp
                   7298: doc-name>string
                   7299: 
1.81      anton    7300: @c ----------------------------------------------------------
                   7301: @node Compiling words, The Text Interpreter, Tokens for Words, Words
                   7302: @section Compiling words
                   7303: @cindex compiling words
                   7304: @cindex macros
                   7305: 
                   7306: In contrast to most other languages, Forth has no strict boundary
1.82      anton    7307: between compilation and run-time.  E.g., you can run arbitrary code
                   7308: between defining words (or for computing data used by defining words
                   7309: like @code{constant}). Moreover, @code{Immediate} (@pxref{Interpretation
                   7310: and Compilation Semantics} and @code{[}...@code{]} (see below) allow
                   7311: running arbitrary code while compiling a colon definition (exception:
                   7312: you must not allot dictionary space).
                   7313: 
                   7314: @menu
                   7315: * Literals::                    Compiling data values
                   7316: * Macros::                      Compiling words
                   7317: @end menu
                   7318: 
                   7319: @node Literals, Macros, Compiling words, Compiling words
                   7320: @subsection Literals
                   7321: @cindex Literals
                   7322: 
                   7323: The simplest and most frequent example is to compute a literal during
                   7324: compilation.  E.g., the following definition prints an array of strings,
                   7325: one string per line:
                   7326: 
                   7327: @example
                   7328: : .strings ( addr u -- ) \ gforth
                   7329:     2* cells bounds U+DO
                   7330:        cr i 2@@ type
                   7331:     2 cells +LOOP ;  
                   7332: @end example
1.81      anton    7333: 
1.82      anton    7334: With a simple-minded compiler like Gforth's, this computes @code{2
                   7335: cells} on every loop iteration.  You can compute this value once and for
                   7336: all at compile time and compile it into the definition like this:
                   7337: 
                   7338: @example
                   7339: : .strings ( addr u -- ) \ gforth
                   7340:     2* cells bounds U+DO
                   7341:        cr i 2@@ type
                   7342:     [ 2 cells ] literal +LOOP ;  
                   7343: @end example
                   7344: 
                   7345: @code{[} switches the text interpreter to interpret state (you will get
                   7346: an @code{ok} prompt if you type this example interactively and insert a
                   7347: newline between @code{[} and @code{]}), so it performs the
                   7348: interpretation semantics of @code{2 cells}; this computes a number.
                   7349: @code{]} switches the text interpreter back into compile state.  It then
                   7350: performs @code{Literal}'s compilation semantics, which are to compile
                   7351: this number into the current word.  You can decompile the word with
                   7352: @code{see .strings} to see the effect on the compiled code.
1.81      anton    7353: 
1.82      anton    7354: You can also optimize the @code{2* cells} into @code{[ 2 cells ] literal
                   7355: *} in this way.
1.81      anton    7356: 
1.82      anton    7357: doc-[
                   7358: doc-]
1.81      anton    7359: doc-literal
                   7360: doc-]L
1.82      anton    7361: 
                   7362: There are also words for compiling other data types than single cells as
                   7363: literals:
                   7364: 
1.81      anton    7365: doc-2literal
                   7366: doc-fliteral
1.82      anton    7367: doc-sliteral
                   7368: 
                   7369: @cindex colon-sys, passing data across @code{:}
                   7370: @cindex @code{:}, passing data across
                   7371: You might be tempted to pass data from outside a colon definition to the
                   7372: inside on the data stack.  This does not work, because @code{:} puhes a
                   7373: colon-sys, making stuff below unaccessible.  E.g., this does not work:
                   7374: 
                   7375: @example
                   7376: 5 : foo literal ; \ error: "unstructured"
                   7377: @end example
                   7378: 
                   7379: Instead, you have to pass the value in some other way, e.g., through a
                   7380: variable:
                   7381: 
                   7382: @example
                   7383: variable temp
                   7384: 5 temp !
                   7385: : foo [ temp @@ ] literal ;
                   7386: @end example
                   7387: 
                   7388: 
                   7389: @node Macros,  , Literals, Compiling words
                   7390: @subsection Macros
                   7391: @cindex Macros
                   7392: @cindex compiling compilation semantics
                   7393: 
                   7394: @code{Literal} and friends compile data values into the current
                   7395: definition.  You can also write words that compile other words into the
                   7396: current definition.  E.g.,
                   7397: 
                   7398: @example
                   7399: : compile-+ ( -- ) \ compiled code: ( n1 n2 -- n )
                   7400:   POSTPONE + ;
                   7401: 
                   7402: : foo ( n1 n2 -- n )
                   7403:   [ compile-+ ] ;
                   7404: 1 2 foo .
                   7405: @end example
                   7406: 
                   7407: This is equivalent to @code{: foo + ;} (@code{see foo} to check this).
                   7408: What happens in this example?  @code{Postpone} compiles the compilation
                   7409: semantics of @code{+} into @code{compile-+}; later the text interpreter
                   7410: executes @code{compile-+} and thus the compilation semantics of +, which
                   7411: compile (the execution semantics of) @code{+} into
                   7412: @code{foo}.@footnote{A recent RFI answer requires that compiling words
                   7413: should only be executed in compile state, so this example is not
                   7414: guaranteed to work on all standard systems, but on any decent system it
                   7415: will work.}
                   7416: 
                   7417: doc-postpone
                   7418: doc-[compile]
                   7419: 
                   7420: Compiling words like @code{compile-+} are usually immediate (or similar)
                   7421: so you do not have to switch to interpret state to execute them;
                   7422: mopifying the last example accordingly produces:
                   7423: 
                   7424: @example
                   7425: : [compile-+] ( compilation: --; interpretation: -- )
                   7426:   \ compiled code: ( n1 n2 -- n )
                   7427:   POSTPONE + ; immediate
                   7428: 
                   7429: : foo ( n1 n2 -- n )
                   7430:   [compile-+] ;
                   7431: 1 2 foo .
                   7432: @end example
                   7433: 
                   7434: Immediate compiling words are similar to macros in other languages (in
                   7435: particular, Lisp).  The important differences to macros in, e.g., C are:
                   7436: 
                   7437: @itemize @bullet
                   7438: 
                   7439: @item
                   7440: You use the same language for defining and processing macros, not a
                   7441: separate preprocessing language and processor.
                   7442: 
                   7443: @item
                   7444: Consequently, the full power of Forth is available in macro definitions.
                   7445: E.g., you can perform arbitrarily complex computations, or generate
                   7446: different code conditionally or in a loop (e.g., @pxref{Advanced macros
                   7447: Tutorial}).  This power is very useful when writing a parser generators
                   7448: or other code-generating software.
                   7449: 
                   7450: @item
                   7451: Macros defined using @code{postpone} etc. deal with the language at a
                   7452: higher level than strings; name binding happens at macro definition
                   7453: time, so you can avoid the pitfalls of name collisions that can happen
                   7454: in C macros.  Of course, Forth is a liberal language and also allows to
                   7455: shoot yourself in the foot with text-interpreted macros like
                   7456: 
                   7457: @example
                   7458: : [compile-+] s" +" evaluate ; immediate
                   7459: @end example
                   7460: 
                   7461: Apart from binding the name at macro use time, using @code{evaluate}
                   7462: also makes your definition @code{state}-smart (@pxref{state-smartness}).
                   7463: @end itemize
                   7464: 
                   7465: You may want the macro to compile a number into a word.  The word to do
                   7466: it is @code{literal}, but you have to @code{postpone} it, so its
                   7467: compilation semantics take effect when the macro is executed, not when
                   7468: it is compiled:
                   7469: 
                   7470: @example
                   7471: : [compile-5] ( -- ) \ compiled code: ( -- n )
                   7472:   5 POSTPONE literal ; immediate
                   7473: 
                   7474: : foo [compile-5] ;
                   7475: foo .
                   7476: @end example
                   7477: 
                   7478: You may want to pass parameters to a macro, that the macro should
                   7479: compile into the current definition.  If the parameter is a number, then
                   7480: you can use @code{postpone literal} (similar for other values).
                   7481: 
                   7482: If you want to pass a word that is to be compiled, the usual way is to
                   7483: pass an execution token and @code{compile,} it:
                   7484: 
                   7485: @example
                   7486: : twice1 ( xt -- ) \ compiled code: ... -- ...
                   7487:   dup compile, compile, ;
                   7488: 
                   7489: : 2+ ( n1 -- n2 )
                   7490:   [ ' 1+ twice1 ] ;
                   7491: @end example
                   7492: 
                   7493: doc-compile,
                   7494: 
                   7495: An alternative available in Gforth, that allows you to pass compile-only
                   7496: words as parameters is to use the compilation token (@pxref{Compilation
                   7497: token}).  The same example in this technique:
                   7498: 
                   7499: @example
                   7500: : twice ( ... ct -- ... ) \ compiled code: ... -- ...
                   7501:   2dup 2>r execute 2r> execute ;
                   7502: 
                   7503: : 2+ ( n1 -- n2 )
                   7504:   [ comp' 1+ twice ] ;
                   7505: @end example
                   7506: 
                   7507: In the example above @code{2>r} and @code{2r>} ensure that @code{twice}
                   7508: works even if the executed compilation semantics has an effect on the
                   7509: data stack.
                   7510: 
                   7511: You can also define complete definitions with these words; this provides
                   7512: an alternative to using @code{does>} (@pxref{User-defined Defining
                   7513: Words}).  E.g., instead of
                   7514: 
                   7515: @example
                   7516: : curry+ ( n1 "name" -- )
                   7517:     CREATE ,
                   7518: DOES> ( n2 -- n1+n2 )
                   7519:     @@ + ;
                   7520: @end example
                   7521: 
                   7522: you could define
                   7523: 
                   7524: @example
                   7525: : curry+ ( n1 "name" -- )
                   7526:   \ name execution: ( n2 -- n1+n2 )
                   7527:   >r : r> POSTPONE literal POSTPONE + POSTPONE ; ;
1.81      anton    7528: 
1.82      anton    7529: -3 curry+ 3-
                   7530: see 3-
                   7531: @end example
1.81      anton    7532: 
1.82      anton    7533: The sequence @code{>r : r>} is necessary, because @code{:} puts a
                   7534: colon-sys on the data stack that makes everything below it unaccessible.
1.81      anton    7535: 
1.82      anton    7536: This way of writing defining words is sometimes more, sometimes less
                   7537: convenient than using @code{does>} (@pxref{Advanced does> usage
                   7538: example}).  One advantage of this method is that it can be optimized
                   7539: better, because the compiler knows that the value compiled with
                   7540: @code{literal} is fixed, whereas the data associated with a
                   7541: @code{create}d word can be changed.
1.47      crook    7542: 
1.26      crook    7543: @c ----------------------------------------------------------
1.81      anton    7544: @node The Text Interpreter, Word Lists, Compiling words, Words
1.26      crook    7545: @section  The Text Interpreter
                   7546: @cindex interpreter - outer
                   7547: @cindex text interpreter
                   7548: @cindex outer interpreter
1.1       anton    7549: 
1.34      anton    7550: @c Should we really describe all these ugly details?  IMO the text
                   7551: @c interpreter should be much cleaner, but that may not be possible within
                   7552: @c ANS Forth. - anton
1.44      crook    7553: @c nac-> I wanted to explain how it works to show how you can exploit
                   7554: @c it in your own programs. When I was writing a cross-compiler, figuring out
                   7555: @c some of these gory details was very helpful to me. None of the textbooks
                   7556: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
                   7557: @c seems to positively avoid going into too much detail for some of
                   7558: @c the internals.
1.34      anton    7559: 
1.71      anton    7560: @c anton: ok.  I wonder, though, if this is the right place; for some stuff
                   7561: @c it is; for the ugly details, I would prefer another place.  I wonder
                   7562: @c whether we should have a chapter before "Words" that describes some
                   7563: @c basic concepts referred to in words, and a chapter after "Words" that
                   7564: @c describes implementation details.
                   7565: 
1.29      crook    7566: The text interpreter@footnote{This is an expanded version of the
                   7567: material in @ref{Introducing the Text Interpreter}.} is an endless loop
1.34      anton    7568: that processes input from the current input device. It is also called
                   7569: the outer interpreter, in contrast to the inner interpreter
                   7570: (@pxref{Engine}) which executes the compiled Forth code on interpretive
                   7571: implementations.
1.27      crook    7572: 
1.29      crook    7573: @cindex interpret state
                   7574: @cindex compile state
                   7575: The text interpreter operates in one of two states: @dfn{interpret
                   7576: state} and @dfn{compile state}. The current state is defined by the
1.71      anton    7577: aptly-named variable @code{state}.
1.29      crook    7578: 
                   7579: This section starts by describing how the text interpreter behaves when
                   7580: it is in interpret state, processing input from the user input device --
                   7581: the keyboard. This is the mode that a Forth system is in after it starts
                   7582: up.
                   7583: 
                   7584: @cindex input buffer
                   7585: @cindex terminal input buffer
                   7586: The text interpreter works from an area of memory called the @dfn{input
                   7587: buffer}@footnote{When the text interpreter is processing input from the
                   7588: keyboard, this area of memory is called the @dfn{terminal input buffer}
                   7589: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
                   7590: @code{#TIB}.}, which stores your keyboard input when you press the
1.30      anton    7591: @key{RET} key. Starting at the beginning of the input buffer, it skips
1.29      crook    7592: leading spaces (called @dfn{delimiters}) then parses a string (a
                   7593: sequence of non-space characters) until it reaches either a space
                   7594: character or the end of the buffer. Having parsed a string, it makes two
                   7595: attempts to process it:
1.27      crook    7596: 
1.29      crook    7597: @cindex dictionary
1.27      crook    7598: @itemize @bullet
                   7599: @item
1.29      crook    7600: It looks for the string in a @dfn{dictionary} of definitions. If the
                   7601: string is found, the string names a @dfn{definition} (also known as a
                   7602: @dfn{word}) and the dictionary search returns information that allows
                   7603: the text interpreter to perform the word's @dfn{interpretation
                   7604: semantics}. In most cases, this simply means that the word will be
                   7605: executed.
1.27      crook    7606: @item
                   7607: If the string is not found in the dictionary, the text interpreter
1.29      crook    7608: attempts to treat it as a number, using the rules described in
                   7609: @ref{Number Conversion}. If the string represents a legal number in the
                   7610: current radix, the number is pushed onto a parameter stack (the data
                   7611: stack for integers, the floating-point stack for floating-point
                   7612: numbers).
                   7613: @end itemize
                   7614: 
                   7615: If both attempts fail, or if the word is found in the dictionary but has
                   7616: no interpretation semantics@footnote{This happens if the word was
                   7617: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
                   7618: remainder of the input buffer, issues an error message and waits for
                   7619: more input. If one of the attempts succeeds, the text interpreter
                   7620: repeats the parsing process until the whole of the input buffer has been
                   7621: processed, at which point it prints the status message ``@code{ ok}''
                   7622: and waits for more input.
                   7623: 
1.71      anton    7624: @c anton: this should be in the input stream subsection (or below it)
                   7625: 
1.29      crook    7626: @cindex parse area
                   7627: The text interpreter keeps track of its position in the input buffer by
                   7628: updating a variable called @code{>IN} (pronounced ``to-in''). The value
                   7629: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
                   7630: of the input buffer. The region from offset @code{>IN @@} to the end of
                   7631: the input buffer is called the @dfn{parse area}@footnote{In other words,
                   7632: the text interpreter processes the contents of the input buffer by
                   7633: parsing strings from the parse area until the parse area is empty.}.
                   7634: This example shows how @code{>IN} changes as the text interpreter parses
                   7635: the input buffer:
                   7636: 
                   7637: @example
                   7638: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
                   7639:   CR ." ->" TYPE ." <-" ; IMMEDIATE 
                   7640: 
                   7641: 1 2 3 remaining + remaining . 
                   7642: 
                   7643: : foo 1 2 3 remaining SWAP remaining ;
                   7644: @end example
                   7645: 
                   7646: @noindent
                   7647: The result is:
                   7648: 
                   7649: @example
                   7650: ->+ remaining .<-
                   7651: ->.<-5  ok
                   7652: 
                   7653: ->SWAP remaining ;-<
                   7654: ->;<-  ok
                   7655: @end example
                   7656: 
                   7657: @cindex parsing words
                   7658: The value of @code{>IN} can also be modified by a word in the input
                   7659: buffer that is executed by the text interpreter.  This means that a word
                   7660: can ``trick'' the text interpreter into either skipping a section of the
                   7661: input buffer@footnote{This is how parsing words work.} or into parsing a
                   7662: section twice. For example:
1.27      crook    7663: 
1.29      crook    7664: @example
1.71      anton    7665: : lat ." <<foo>>" ;
                   7666: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
1.29      crook    7667: @end example
                   7668: 
                   7669: @noindent
                   7670: When @code{flat} is executed, this output is produced@footnote{Exercise
                   7671: for the reader: what would happen if the @code{3} were replaced with
                   7672: @code{4}?}:
                   7673: 
                   7674: @example
1.71      anton    7675: <<bar>><<foo>>
1.29      crook    7676: @end example
                   7677: 
1.71      anton    7678: This technique can be used to work around some of the interoperability
                   7679: problems of parsing words.  Of course, it's better to avoid parsing
                   7680: words where possible.
                   7681: 
1.29      crook    7682: @noindent
                   7683: Two important notes about the behaviour of the text interpreter:
1.27      crook    7684: 
                   7685: @itemize @bullet
                   7686: @item
                   7687: It processes each input string to completion before parsing additional
1.29      crook    7688: characters from the input buffer.
                   7689: @item
                   7690: It treats the input buffer as a read-only region (and so must your code).
                   7691: @end itemize
                   7692: 
                   7693: @noindent
                   7694: When the text interpreter is in compile state, its behaviour changes in
                   7695: these ways:
                   7696: 
                   7697: @itemize @bullet
                   7698: @item
                   7699: If a parsed string is found in the dictionary, the text interpreter will
                   7700: perform the word's @dfn{compilation semantics}. In most cases, this
                   7701: simply means that the execution semantics of the word will be appended
                   7702: to the current definition.
1.27      crook    7703: @item
1.29      crook    7704: When a number is encountered, it is compiled into the current definition
                   7705: (as a literal) rather than being pushed onto a parameter stack.
                   7706: @item
                   7707: If an error occurs, @code{state} is modified to put the text interpreter
                   7708: back into interpret state.
                   7709: @item
                   7710: Each time a line is entered from the keyboard, Gforth prints
                   7711: ``@code{ compiled}'' rather than `` @code{ok}''.
                   7712: @end itemize
                   7713: 
                   7714: @cindex text interpreter - input sources
                   7715: When the text interpreter is using an input device other than the
                   7716: keyboard, its behaviour changes in these ways:
                   7717: 
                   7718: @itemize @bullet
                   7719: @item
                   7720: When the parse area is empty, the text interpreter attempts to refill
                   7721: the input buffer from the input source. When the input source is
1.71      anton    7722: exhausted, the input source is set back to the previous input source.
1.29      crook    7723: @item
                   7724: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
                   7725: time the parse area is emptied.
                   7726: @item
                   7727: If an error occurs, the input source is set back to the user input
                   7728: device.
1.27      crook    7729: @end itemize
1.21      crook    7730: 
1.49      anton    7731: You can read about this in more detail in @ref{Input Sources}.
1.44      crook    7732: 
1.26      crook    7733: doc->in
1.27      crook    7734: doc-source
                   7735: 
1.26      crook    7736: doc-tib
                   7737: doc-#tib
1.1       anton    7738: 
1.44      crook    7739: 
1.26      crook    7740: @menu
1.67      anton    7741: * Input Sources::               
                   7742: * Number Conversion::           
                   7743: * Interpret/Compile states::    
                   7744: * Interpreter Directives::      
1.26      crook    7745: @end menu
1.1       anton    7746: 
1.29      crook    7747: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
                   7748: @subsection Input Sources
                   7749: @cindex input sources
                   7750: @cindex text interpreter - input sources
                   7751: 
1.44      crook    7752: By default, the text interpreter processes input from the user input
1.29      crook    7753: device (the keyboard) when Forth starts up. The text interpreter can
                   7754: process input from any of these sources:
                   7755: 
                   7756: @itemize @bullet
                   7757: @item
                   7758: The user input device -- the keyboard.
                   7759: @item
                   7760: A file, using the words described in @ref{Forth source files}.
                   7761: @item
                   7762: A block, using the words described in @ref{Blocks}.
                   7763: @item
                   7764: A text string, using @code{evaluate}.
                   7765: @end itemize
                   7766: 
                   7767: A program can identify the current input device from the values of
                   7768: @code{source-id} and @code{blk}.
                   7769: 
1.44      crook    7770: 
1.29      crook    7771: doc-source-id
                   7772: doc-blk
                   7773: 
                   7774: doc-save-input
                   7775: doc-restore-input
                   7776: 
                   7777: doc-evaluate
1.1       anton    7778: 
1.29      crook    7779: 
1.44      crook    7780: 
1.29      crook    7781: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
1.26      crook    7782: @subsection Number Conversion
                   7783: @cindex number conversion
                   7784: @cindex double-cell numbers, input format
                   7785: @cindex input format for double-cell numbers
                   7786: @cindex single-cell numbers, input format
                   7787: @cindex input format for single-cell numbers
                   7788: @cindex floating-point numbers, input format
                   7789: @cindex input format for floating-point numbers
1.1       anton    7790: 
1.29      crook    7791: This section describes the rules that the text interpreter uses when it
                   7792: tries to convert a string into a number.
1.1       anton    7793: 
1.26      crook    7794: Let <digit> represent any character that is a legal digit in the current
1.29      crook    7795: number base@footnote{For example, 0-9 when the number base is decimal or
                   7796: 0-9, A-F when the number base is hexadecimal.}.
1.1       anton    7797: 
1.26      crook    7798: Let <decimal digit> represent any character in the range 0-9.
1.1       anton    7799: 
1.29      crook    7800: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
                   7801: in the braces (@i{a} or @i{b} or neither).
1.1       anton    7802: 
1.26      crook    7803: Let * represent any number of instances of the previous character
                   7804: (including none).
1.1       anton    7805: 
1.26      crook    7806: Let any other character represent itself.
1.1       anton    7807: 
1.29      crook    7808: @noindent
1.26      crook    7809: Now, the conversion rules are:
1.21      crook    7810: 
1.26      crook    7811: @itemize @bullet
                   7812: @item
                   7813: A string of the form <digit><digit>* is treated as a single-precision
1.29      crook    7814: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
1.26      crook    7815: @item
                   7816: A string of the form -<digit><digit>* is treated as a single-precision
1.29      crook    7817: (cell-sized) negative integer, and is represented using 2's-complement
1.26      crook    7818: arithmetic. Examples are -45 -5681 -0
                   7819: @item
                   7820: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
1.29      crook    7821: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
                   7822: (all three of these represent the same number).
1.26      crook    7823: @item
                   7824: A string of the form -<digit><digit>*.<digit>* is treated as a
1.29      crook    7825: double-precision (double-cell-sized) negative integer, and is
1.26      crook    7826: represented using 2's-complement arithmetic. Examples are -3465. -3.465
1.29      crook    7827: -34.65 (all three of these represent the same number).
1.26      crook    7828: @item
1.29      crook    7829: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
                   7830: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
1.35      anton    7831: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
1.29      crook    7832: number) +12.E-4
1.26      crook    7833: @end itemize
1.1       anton    7834: 
1.26      crook    7835: By default, the number base used for integer number conversion is given
1.35      anton    7836: by the contents of the variable @code{base}.  Note that a lot of
                   7837: confusion can result from unexpected values of @code{base}.  If you
                   7838: change @code{base} anywhere, make sure to save the old value and restore
                   7839: it afterwards.  In general I recommend keeping @code{base} decimal, and
                   7840: using the prefixes described below for the popular non-decimal bases.
1.1       anton    7841: 
1.29      crook    7842: doc-dpl
1.26      crook    7843: doc-base
                   7844: doc-hex
                   7845: doc-decimal
1.1       anton    7846: 
1.44      crook    7847: 
1.26      crook    7848: @cindex '-prefix for character strings
                   7849: @cindex &-prefix for decimal numbers
                   7850: @cindex %-prefix for binary numbers
                   7851: @cindex $-prefix for hexadecimal numbers
1.35      anton    7852: Gforth allows you to override the value of @code{base} by using a
1.29      crook    7853: prefix@footnote{Some Forth implementations provide a similar scheme by
                   7854: implementing @code{$} etc. as parsing words that process the subsequent
                   7855: number in the input stream and push it onto the stack. For example, see
                   7856: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
                   7857: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
                   7858: is required between the prefix and the number.} before the first digit
                   7859: of an (integer) number. Four prefixes are supported:
1.1       anton    7860: 
1.26      crook    7861: @itemize @bullet
                   7862: @item
1.35      anton    7863: @code{&} -- decimal
1.26      crook    7864: @item
1.35      anton    7865: @code{%} -- binary
1.26      crook    7866: @item
1.35      anton    7867: @code{$} -- hexadecimal
1.26      crook    7868: @item
1.35      anton    7869: @code{'} -- base @code{max-char+1}
1.26      crook    7870: @end itemize
1.1       anton    7871: 
1.26      crook    7872: Here are some examples, with the equivalent decimal number shown after
                   7873: in braces:
1.1       anton    7874: 
1.26      crook    7875: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
                   7876: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
                   7877: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
                   7878: &905 (905), $abc (2478), $ABC (2478).
1.1       anton    7879: 
1.26      crook    7880: @cindex number conversion - traps for the unwary
1.29      crook    7881: @noindent
1.26      crook    7882: Number conversion has a number of traps for the unwary:
1.1       anton    7883: 
1.26      crook    7884: @itemize @bullet
                   7885: @item
                   7886: You cannot determine the current number base using the code sequence
1.35      anton    7887: @code{base @@ .} -- the number base is always 10 in the current number
                   7888: base. Instead, use something like @code{base @@ dec.}
1.26      crook    7889: @item
                   7890: If the number base is set to a value greater than 14 (for example,
                   7891: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
                   7892: it to be intepreted as either a single-precision integer or a
                   7893: floating-point number (Gforth treats it as an integer). The ambiguity
                   7894: can be resolved by explicitly stating the sign of the mantissa and/or
                   7895: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
                   7896: ambiguity arises; either representation will be treated as a
                   7897: floating-point number.
                   7898: @item
1.29      crook    7899: There is a word @code{bin} but it does @i{not} set the number base!
1.26      crook    7900: It is used to specify file types.
                   7901: @item
1.72      anton    7902: ANS Forth requires the @code{.} of a double-precision number to be the
                   7903: final character in the string.  Gforth allows the @code{.} to be
                   7904: anywhere after the first digit.
1.26      crook    7905: @item
                   7906: The number conversion process does not check for overflow.
                   7907: @item
1.72      anton    7908: In an ANS Forth program @code{base} is required to be decimal when
                   7909: converting floating-point numbers.  In Gforth, number conversion to
                   7910: floating-point numbers always uses base &10, irrespective of the value
                   7911: of @code{base}.
1.26      crook    7912: @end itemize
1.1       anton    7913: 
1.49      anton    7914: You can read numbers into your programs with the words described in
                   7915: @ref{Input}.
1.1       anton    7916: 
1.82      anton    7917: @node Interpret/Compile states, Interpreter Directives, Number Conversion, The Text Interpreter
1.26      crook    7918: @subsection Interpret/Compile states
                   7919: @cindex Interpret/Compile states
1.1       anton    7920: 
1.29      crook    7921: A standard program is not permitted to change @code{state}
                   7922: explicitly. However, it can change @code{state} implicitly, using the
                   7923: words @code{[} and @code{]}. When @code{[} is executed it switches
                   7924: @code{state} to interpret state, and therefore the text interpreter
                   7925: starts interpreting. When @code{]} is executed it switches @code{state}
                   7926: to compile state and therefore the text interpreter starts
1.44      crook    7927: compiling. The most common usage for these words is for switching into
                   7928: interpret state and back from within a colon definition; this technique
1.49      anton    7929: can be used to compile a literal (for an example, @pxref{Literals}) or
                   7930: for conditional compilation (for an example, @pxref{Interpreter
                   7931: Directives}).
1.44      crook    7932: 
1.35      anton    7933: 
                   7934: @c This is a bad example: It's non-standard, and it's not necessary.
                   7935: @c However, I can't think of a good example for switching into compile
                   7936: @c state when there is no current word (@code{state}-smart words are not a
                   7937: @c good reason).  So maybe we should use an example for switching into
                   7938: @c interpret @code{state} in a colon def. - anton
1.44      crook    7939: @c nac-> I agree. I started out by putting in the example, then realised
                   7940: @c that it was non-ANS, so wrote more words around it. I hope this
                   7941: @c re-written version is acceptable to you. I do want to keep the example
                   7942: @c as it is helpful for showing what is and what is not portable, particularly
                   7943: @c where it outlaws a style in common use.
                   7944: 
1.72      anton    7945: @c anton: it's more important to show what's portable.  After we have done
1.83      anton    7946: @c that, we can also show what's not.  In any case, I have written a
                   7947: @c section Compiling Words which also deals with [ ].
1.35      anton    7948: 
1.44      crook    7949: @code{[} and @code{]} also give you the ability to switch into compile
                   7950: state and back, but we cannot think of any useful Standard application
                   7951: for this ability. Pre-ANS Forth textbooks have examples like this:
1.29      crook    7952: 
                   7953: @example
                   7954: : AA ." this is A" ;
                   7955: : BB ." this is B" ;
                   7956: : CC ." this is C" ;
                   7957: 
1.44      crook    7958: create table ] aa bb cc [
                   7959: 
1.29      crook    7960: : go ( n -- ) \ n is offset into table.. 0 for 1st entry
                   7961:   cells table + @ execute ;
                   7962: @end example
                   7963: 
1.44      crook    7964: This example builds a jump table; @code{0 go} will display ``@code{this
                   7965: is A}''. Using @code{[} and @code{]} in this example is equivalent to
                   7966: defining @code{table} like this:
1.29      crook    7967: 
                   7968: @example
1.44      crook    7969: create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
1.29      crook    7970: @end example
                   7971: 
1.44      crook    7972: The problem with this code is that the definition of @code{table} is not
                   7973: portable -- it @i{compile}s execution tokens into code space. Whilst it
                   7974: @i{may} work on systems where code space and data space co-incide, the
1.29      crook    7975: Standard only allows data space to be assigned for a @code{CREATE}d
                   7976: word. In addition, the Standard only allows @code{@@} to access data
                   7977: space, whilst this example is using it to access code space. The only
                   7978: portable, Standard way to build this table is to build it in data space,
                   7979: like this:
                   7980: 
                   7981: @example
                   7982: create table ' aa , ' bb , ' cc ,
                   7983: @end example
                   7984: 
1.26      crook    7985: doc-state
1.44      crook    7986: 
1.29      crook    7987: 
1.82      anton    7988: @node Interpreter Directives,  , Interpret/Compile states, The Text Interpreter
1.26      crook    7989: @subsection Interpreter Directives
                   7990: @cindex interpreter directives
1.72      anton    7991: @cindex conditional compilation
1.1       anton    7992: 
1.29      crook    7993: These words are usually used in interpret state; typically to control
                   7994: which parts of a source file are processed by the text
1.26      crook    7995: interpreter. There are only a few ANS Forth Standard words, but Gforth
                   7996: supplements these with a rich set of immediate control structure words
                   7997: to compensate for the fact that the non-immediate versions can only be
1.29      crook    7998: used in compile state (@pxref{Control Structures}). Typical usages:
                   7999: 
                   8000: @example
1.72      anton    8001: FALSE Constant HAVE-ASSEMBLER
1.29      crook    8002: .
                   8003: .
1.72      anton    8004: HAVE-ASSEMBLER [IF]
1.29      crook    8005: : ASSEMBLER-FEATURE
                   8006:   ...
                   8007: ;
                   8008: [ENDIF]
                   8009: .
                   8010: .
                   8011: : SEE
                   8012:   ... \ general-purpose SEE code
1.72      anton    8013:   [ HAVE-ASSEMBLER [IF] ]
1.29      crook    8014:   ... \ assembler-specific SEE code
                   8015:   [ [ENDIF] ]
                   8016: ;
                   8017: @end example
1.1       anton    8018: 
1.44      crook    8019: 
1.26      crook    8020: doc-[IF]
                   8021: doc-[ELSE]
                   8022: doc-[THEN]
                   8023: doc-[ENDIF]
1.1       anton    8024: 
1.26      crook    8025: doc-[IFDEF]
                   8026: doc-[IFUNDEF]
1.1       anton    8027: 
1.26      crook    8028: doc-[?DO]
                   8029: doc-[DO]
                   8030: doc-[FOR]
                   8031: doc-[LOOP]
                   8032: doc-[+LOOP]
                   8033: doc-[NEXT]
1.1       anton    8034: 
1.26      crook    8035: doc-[BEGIN]
                   8036: doc-[UNTIL]
                   8037: doc-[AGAIN]
                   8038: doc-[WHILE]
                   8039: doc-[REPEAT]
1.1       anton    8040: 
1.27      crook    8041: 
1.26      crook    8042: @c -------------------------------------------------------------
1.47      crook    8043: @node Word Lists, Environmental Queries, The Text Interpreter, Words
1.26      crook    8044: @section Word Lists
                   8045: @cindex word lists
1.32      anton    8046: @cindex header space
1.1       anton    8047: 
1.36      anton    8048: A wordlist is a list of named words; you can add new words and look up
                   8049: words by name (and you can remove words in a restricted way with
                   8050: markers).  Every named (and @code{reveal}ed) word is in one wordlist.
                   8051: 
                   8052: @cindex search order stack
                   8053: The text interpreter searches the wordlists present in the search order
                   8054: (a stack of wordlists), from the top to the bottom.  Within each
                   8055: wordlist, the search starts conceptually at the newest word; i.e., if
                   8056: two words in a wordlist have the same name, the newer word is found.
1.1       anton    8057: 
1.26      crook    8058: @cindex compilation word list
1.36      anton    8059: New words are added to the @dfn{compilation wordlist} (aka current
                   8060: wordlist).
1.1       anton    8061: 
1.36      anton    8062: @cindex wid
                   8063: A word list is identified by a cell-sized word list identifier (@i{wid})
                   8064: in much the same way as a file is identified by a file handle. The
                   8065: numerical value of the wid has no (portable) meaning, and might change
                   8066: from session to session.
1.1       anton    8067: 
1.29      crook    8068: The ANS Forth ``Search order'' word set is intended to provide a set of
                   8069: low-level tools that allow various different schemes to be
1.74      anton    8070: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
1.26      crook    8071: word.  @file{compat/vocabulary.fs} provides an implementation in ANS
1.45      crook    8072: Forth.
1.1       anton    8073: 
1.27      crook    8074: @comment TODO: locals section refers to here, saying that every word list (aka
                   8075: @comment vocabulary) has its own methods for searching etc. Need to document that.
1.78      anton    8076: @c anton: but better in a separate subsection on wordlist internals
1.1       anton    8077: 
1.45      crook    8078: @comment TODO: document markers, reveal, tables, mappedwordlist
                   8079: 
                   8080: @comment the gforthman- prefix is used to pick out the true definition of a
1.27      crook    8081: @comment word from the source files, rather than some alias.
1.44      crook    8082: 
1.26      crook    8083: doc-forth-wordlist
                   8084: doc-definitions
                   8085: doc-get-current
                   8086: doc-set-current
                   8087: doc-get-order
1.45      crook    8088: doc---gforthman-set-order
1.26      crook    8089: doc-wordlist
1.30      anton    8090: doc-table
1.79      anton    8091: doc->order
1.36      anton    8092: doc-previous
1.26      crook    8093: doc-also
1.45      crook    8094: doc---gforthman-forth
1.26      crook    8095: doc-only
1.45      crook    8096: doc---gforthman-order
1.15      anton    8097: 
1.26      crook    8098: doc-find
                   8099: doc-search-wordlist
1.15      anton    8100: 
1.26      crook    8101: doc-words
                   8102: doc-vlist
1.44      crook    8103: @c doc-words-deferred
1.1       anton    8104: 
1.74      anton    8105: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
1.26      crook    8106: doc-root
                   8107: doc-vocabulary
                   8108: doc-seal
                   8109: doc-vocs
                   8110: doc-current
                   8111: doc-context
1.1       anton    8112: 
1.44      crook    8113: 
1.26      crook    8114: @menu
1.75      anton    8115: * Vocabularies::                
1.67      anton    8116: * Why use word lists?::         
1.75      anton    8117: * Word list example::           
1.26      crook    8118: @end menu
                   8119: 
1.75      anton    8120: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
                   8121: @subsection Vocabularies
                   8122: @cindex Vocabularies, detailed explanation
                   8123: 
                   8124: Here is an example of creating and using a new wordlist using ANS
                   8125: Forth words:
                   8126: 
                   8127: @example
                   8128: wordlist constant my-new-words-wordlist
                   8129: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
                   8130: 
                   8131: \ add it to the search order
                   8132: also my-new-words
                   8133: 
                   8134: \ alternatively, add it to the search order and make it
                   8135: \ the compilation word list
                   8136: also my-new-words definitions
                   8137: \ type "order" to see the problem
                   8138: @end example
                   8139: 
                   8140: The problem with this example is that @code{order} has no way to
                   8141: associate the name @code{my-new-words} with the wid of the word list (in
                   8142: Gforth, @code{order} and @code{vocs} will display @code{???}  for a wid
                   8143: that has no associated name). There is no Standard way of associating a
                   8144: name with a wid.
                   8145: 
                   8146: In Gforth, this example can be re-coded using @code{vocabulary}, which
                   8147: associates a name with a wid:
                   8148: 
                   8149: @example
                   8150: vocabulary my-new-words
                   8151: 
                   8152: \ add it to the search order
                   8153: also my-new-words
                   8154: 
                   8155: \ alternatively, add it to the search order and make it
                   8156: \ the compilation word list
                   8157: my-new-words definitions
                   8158: \ type "order" to see that the problem is solved
                   8159: @end example
                   8160: 
                   8161: 
                   8162: @node Why use word lists?, Word list example, Vocabularies, Word Lists
1.26      crook    8163: @subsection Why use word lists?
                   8164: @cindex word lists - why use them?
                   8165: 
1.74      anton    8166: Here are some reasons why people use wordlists:
1.26      crook    8167: 
                   8168: @itemize @bullet
1.74      anton    8169: 
                   8170: @c anton: Gforth's hashing implementation makes the search speed
                   8171: @c independent from the number of words.  But it is linear with the number
                   8172: @c of wordlists that have to be searched, so in effect using more wordlists
                   8173: @c actually slows down compilation.
                   8174: 
                   8175: @c @item
                   8176: @c To improve compilation speed by reducing the number of header space
                   8177: @c entries that must be searched. This is achieved by creating a new
                   8178: @c word list that contains all of the definitions that are used in the
                   8179: @c definition of a Forth system but which would not usually be used by
                   8180: @c programs running on that system. That word list would be on the search
                   8181: @c list when the Forth system was compiled but would be removed from the
                   8182: @c search list for normal operation. This can be a useful technique for
                   8183: @c low-performance systems (for example, 8-bit processors in embedded
                   8184: @c systems) but is unlikely to be necessary in high-performance desktop
                   8185: @c systems.
                   8186: 
1.26      crook    8187: @item
                   8188: To prevent a set of words from being used outside the context in which
                   8189: they are valid. Two classic examples of this are an integrated editor
                   8190: (all of the edit commands are defined in a separate word list; the
                   8191: search order is set to the editor word list when the editor is invoked;
                   8192: the old search order is restored when the editor is terminated) and an
                   8193: integrated assembler (the op-codes for the machine are defined in a
                   8194: separate word list which is used when a @code{CODE} word is defined).
1.74      anton    8195: 
                   8196: @item
                   8197: To organize the words of an application or library into a user-visible
                   8198: set (in @code{forth-wordlist} or some other common wordlist) and a set
                   8199: of helper words used just for the implementation (hidden in a separate
1.75      anton    8200: wordlist).  This keeps @code{words}' output smaller, separates
                   8201: implementation and interface, and reduces the chance of name conflicts
                   8202: within the common wordlist.
1.74      anton    8203: 
1.26      crook    8204: @item
                   8205: To prevent a name-space clash between multiple definitions with the same
                   8206: name. For example, when building a cross-compiler you might have a word
                   8207: @code{IF} that generates conditional code for your target system. By
                   8208: placing this definition in a different word list you can control whether
                   8209: the host system's @code{IF} or the target system's @code{IF} get used in
                   8210: any particular context by controlling the order of the word lists on the
                   8211: search order stack.
1.74      anton    8212: 
1.26      crook    8213: @end itemize
1.1       anton    8214: 
1.74      anton    8215: The downsides of using wordlists are:
                   8216: 
                   8217: @itemize
                   8218: 
                   8219: @item
                   8220: Debugging becomes more cumbersome.
                   8221: 
                   8222: @item
                   8223: Name conflicts worked around with wordlists are still there, and you
                   8224: have to arrange the search order carefully to get the desired results;
                   8225: if you forget to do that, you get hard-to-find errors (as in any case
                   8226: where you read the code differently from the compiler; @code{see} can
1.75      anton    8227: help seeing which of several possible words the name resolves to in such
                   8228: cases).  @code{See} displays just the name of the words, not what
                   8229: wordlist they belong to, so it might be misleading.  Using unique names
                   8230: is a better approach to avoid name conflicts.
1.74      anton    8231: 
                   8232: @item
                   8233: You have to explicitly undo any changes to the search order.  In many
                   8234: cases it would be more convenient if this happened implicitly.  Gforth
                   8235: currently does not provide such a feature, but it may do so in the
                   8236: future.
                   8237: @end itemize
                   8238: 
                   8239: 
1.75      anton    8240: @node Word list example,  , Why use word lists?, Word Lists
                   8241: @subsection Word list example
                   8242: @cindex word lists - example
1.1       anton    8243: 
1.74      anton    8244: The following example is from the
                   8245: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
                   8246: garbage collector} and uses wordlists to separate public words from
                   8247: helper words:
                   8248: 
                   8249: @example
                   8250: get-current ( wid )
                   8251: vocabulary garbage-collector also garbage-collector definitions
                   8252: ... \ define helper words
                   8253: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
                   8254: ... \ define the public (i.e., API) words
                   8255:     \ they can refer to the helper words
                   8256: previous \ restore original search order (helper words become invisible)
                   8257: @end example
                   8258: 
1.26      crook    8259: @c -------------------------------------------------------------
                   8260: @node Environmental Queries, Files, Word Lists, Words
                   8261: @section Environmental Queries
                   8262: @cindex environmental queries
1.21      crook    8263: 
1.26      crook    8264: ANS Forth introduced the idea of ``environmental queries'' as a way
                   8265: for a program running on a system to determine certain characteristics of the system.
                   8266: The Standard specifies a number of strings that might be recognised by a system.
1.21      crook    8267: 
1.32      anton    8268: The Standard requires that the header space used for environmental queries
                   8269: be distinct from the header space used for definitions.
1.21      crook    8270: 
1.26      crook    8271: Typically, environmental queries are supported by creating a set of
1.29      crook    8272: definitions in a word list that is @i{only} used during environmental
1.26      crook    8273: queries; that is what Gforth does. There is no Standard way of adding
                   8274: definitions to the set of recognised environmental queries, but any
                   8275: implementation that supports the loading of optional word sets must have
                   8276: some mechanism for doing this (after loading the word set, the
                   8277: associated environmental query string must return @code{true}). In
                   8278: Gforth, the word list used to honour environmental queries can be
                   8279: manipulated just like any other word list.
1.21      crook    8280: 
1.44      crook    8281: 
1.26      crook    8282: doc-environment?
                   8283: doc-environment-wordlist
1.21      crook    8284: 
1.26      crook    8285: doc-gforth
                   8286: doc-os-class
1.21      crook    8287: 
1.44      crook    8288: 
1.26      crook    8289: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
                   8290: returning two items on the stack, querying it using @code{environment?}
                   8291: will return an additional item; the @code{true} flag that shows that the
                   8292: string was recognised.
1.21      crook    8293: 
1.26      crook    8294: @comment TODO Document the standard strings or note where they are documented herein
1.21      crook    8295: 
1.26      crook    8296: Here are some examples of using environmental queries:
1.21      crook    8297: 
1.26      crook    8298: @example
                   8299: s" address-unit-bits" environment? 0=
                   8300: [IF]
                   8301:      cr .( environmental attribute address-units-bits unknown... ) cr
1.75      anton    8302: [ELSE]
                   8303:      drop \ ensure balanced stack effect
1.26      crook    8304: [THEN]
1.21      crook    8305: 
1.75      anton    8306: \ this might occur in the prelude of a standard program that uses THROW
                   8307: s" exception" environment? [IF]
                   8308:    0= [IF]
                   8309:       : throw abort" exception thrown" ;
                   8310:    [THEN]
                   8311: [ELSE] \ we don't know, so make sure
                   8312:    : throw abort" exception thrown" ;
                   8313: [THEN]
1.21      crook    8314: 
1.26      crook    8315: s" gforth" environment? [IF] .( Gforth version ) TYPE
                   8316:                         [ELSE] .( Not Gforth..) [THEN]
1.75      anton    8317: 
                   8318: \ a program using v*
                   8319: s" gforth" environment? [IF]
                   8320:   s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
                   8321:    : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
                   8322:      >r swap 2swap swap 0e r> 0 ?DO
                   8323:        dup f@ over + 2swap dup f@ f* f+ over + 2swap
                   8324:      LOOP
                   8325:      2drop 2drop ; 
                   8326:   [THEN]
                   8327: [ELSE] \ 
                   8328:   : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
                   8329:   ...
                   8330: [THEN]
1.26      crook    8331: @end example
1.21      crook    8332: 
1.26      crook    8333: Here is an example of adding a definition to the environment word list:
1.21      crook    8334: 
1.26      crook    8335: @example
                   8336: get-current environment-wordlist set-current
                   8337: true constant block
                   8338: true constant block-ext
                   8339: set-current
                   8340: @end example
1.21      crook    8341: 
1.26      crook    8342: You can see what definitions are in the environment word list like this:
1.21      crook    8343: 
1.26      crook    8344: @example
1.79      anton    8345: environment-wordlist >order words previous
1.26      crook    8346: @end example
1.21      crook    8347: 
                   8348: 
1.26      crook    8349: @c -------------------------------------------------------------
                   8350: @node Files, Blocks, Environmental Queries, Words
                   8351: @section Files
1.28      crook    8352: @cindex files
                   8353: @cindex I/O - file-handling
1.21      crook    8354: 
1.26      crook    8355: Gforth provides facilities for accessing files that are stored in the
                   8356: host operating system's file-system. Files that are processed by Gforth
                   8357: can be divided into two categories:
1.21      crook    8358: 
1.23      crook    8359: @itemize @bullet
                   8360: @item
1.29      crook    8361: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
1.23      crook    8362: @item
1.29      crook    8363: Files that are processed by some other program (@dfn{general files}).
1.26      crook    8364: @end itemize
                   8365: 
                   8366: @menu
1.48      anton    8367: * Forth source files::          
                   8368: * General files::               
                   8369: * Search Paths::                
1.26      crook    8370: @end menu
                   8371: 
                   8372: @c -------------------------------------------------------------
                   8373: @node Forth source files, General files, Files, Files
                   8374: @subsection Forth source files
                   8375: @cindex including files
                   8376: @cindex Forth source files
1.21      crook    8377: 
1.26      crook    8378: The simplest way to interpret the contents of a file is to use one of
                   8379: these two formats:
1.21      crook    8380: 
1.26      crook    8381: @example
                   8382: include mysource.fs
                   8383: s" mysource.fs" included
                   8384: @end example
1.21      crook    8385: 
1.75      anton    8386: You usually want to include a file only if it is not included already
1.26      crook    8387: (by, say, another source file). In that case, you can use one of these
1.45      crook    8388: three formats:
1.21      crook    8389: 
1.26      crook    8390: @example
                   8391: require mysource.fs
                   8392: needs mysource.fs
                   8393: s" mysource.fs" required
                   8394: @end example
1.21      crook    8395: 
1.26      crook    8396: @cindex stack effect of included files
                   8397: @cindex including files, stack effect
1.45      crook    8398: It is good practice to write your source files such that interpreting them
                   8399: does not change the stack. Source files designed in this way can be used with
1.26      crook    8400: @code{required} and friends without complications. For example:
1.21      crook    8401: 
1.26      crook    8402: @example
1.75      anton    8403: 1024 require foo.fs drop
1.26      crook    8404: @end example
1.21      crook    8405: 
1.75      anton    8406: Here you want to pass the argument 1024 (e.g., a buffer size) to
                   8407: @file{foo.fs}.  Interpreting @file{foo.fs} has the stack effect ( n -- n
                   8408: ), which allows its use with @code{require}.  Of course with such
                   8409: parameters to required files, you have to ensure that the first
                   8410: @code{require} fits for all uses (i.e., @code{require} it early in the
                   8411: master load file).
1.44      crook    8412: 
1.26      crook    8413: doc-include-file
                   8414: doc-included
1.28      crook    8415: doc-included?
1.26      crook    8416: doc-include
                   8417: doc-required
                   8418: doc-require
                   8419: doc-needs
1.75      anton    8420: @c doc-init-included-files @c internal
                   8421: doc-sourcefilename
                   8422: doc-sourceline#
1.44      crook    8423: 
1.26      crook    8424: A definition in ANS Forth for @code{required} is provided in
                   8425: @file{compat/required.fs}.
1.21      crook    8426: 
1.26      crook    8427: @c -------------------------------------------------------------
                   8428: @node General files, Search Paths, Forth source files, Files
                   8429: @subsection General files
                   8430: @cindex general files
                   8431: @cindex file-handling
1.21      crook    8432: 
1.75      anton    8433: Files are opened/created by name and type. The following file access
                   8434: methods (FAMs) are recognised:
1.44      crook    8435: 
1.75      anton    8436: @cindex fam (file access method)
1.26      crook    8437: doc-r/o
                   8438: doc-r/w
                   8439: doc-w/o
                   8440: doc-bin
1.1       anton    8441: 
1.44      crook    8442: 
1.26      crook    8443: When a file is opened/created, it returns a file identifier,
1.29      crook    8444: @i{wfileid} that is used for all other file commands. All file
                   8445: commands also return a status value, @i{wior}, that is 0 for a
1.26      crook    8446: successful operation and an implementation-defined non-zero value in the
                   8447: case of an error.
1.21      crook    8448: 
1.44      crook    8449: 
1.26      crook    8450: doc-open-file
                   8451: doc-create-file
1.21      crook    8452: 
1.26      crook    8453: doc-close-file
                   8454: doc-delete-file
                   8455: doc-rename-file
                   8456: doc-read-file
                   8457: doc-read-line
                   8458: doc-write-file
                   8459: doc-write-line
                   8460: doc-emit-file
                   8461: doc-flush-file
1.21      crook    8462: 
1.26      crook    8463: doc-file-status
                   8464: doc-file-position
                   8465: doc-reposition-file
                   8466: doc-file-size
                   8467: doc-resize-file
1.21      crook    8468: 
1.44      crook    8469: 
1.26      crook    8470: @c ---------------------------------------------------------
1.48      anton    8471: @node Search Paths,  , General files, Files
1.26      crook    8472: @subsection Search Paths
                   8473: @cindex path for @code{included}
                   8474: @cindex file search path
                   8475: @cindex @code{include} search path
                   8476: @cindex search path for files
1.21      crook    8477: 
1.26      crook    8478: If you specify an absolute filename (i.e., a filename starting with
                   8479: @file{/} or @file{~}, or with @file{:} in the second position (as in
                   8480: @samp{C:...})) for @code{included} and friends, that file is included
                   8481: just as you would expect.
1.21      crook    8482: 
1.75      anton    8483: If the filename starts with @file{./}, this refers to the directory that
                   8484: the present file was @code{included} from.  This allows files to include
                   8485: other files relative to their own position (irrespective of the current
                   8486: working directory or the absolute position).  This feature is essential
                   8487: for libraries consisting of several files, where a file may include
                   8488: other files from the library.  It corresponds to @code{#include "..."}
                   8489: in C. If the current input source is not a file, @file{.} refers to the
                   8490: directory of the innermost file being included, or, if there is no file
                   8491: being included, to the current working directory.
                   8492: 
                   8493: For relative filenames (not starting with @file{./}), Gforth uses a
                   8494: search path similar to Forth's search order (@pxref{Word Lists}). It
                   8495: tries to find the given filename in the directories present in the path,
                   8496: and includes the first one it finds. There are separate search paths for
                   8497: Forth source files and general files.  If the search path contains the
                   8498: directory @file{.}, this refers to the directory of the current file, or
                   8499: the working directory, as if the file had been specified with @file{./}.
1.21      crook    8500: 
1.26      crook    8501: Use @file{~+} to refer to the current working directory (as in the
                   8502: @code{bash}).
1.1       anton    8503: 
1.75      anton    8504: @c anton: fold the following subsubsections into this subsection?
1.1       anton    8505: 
1.48      anton    8506: @menu
1.75      anton    8507: * Source Search Paths::         
1.48      anton    8508: * General Search Paths::        
                   8509: @end menu
                   8510: 
1.26      crook    8511: @c ---------------------------------------------------------
1.75      anton    8512: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
                   8513: @subsubsection Source Search Paths
                   8514: @cindex search path control, source files
1.5       anton    8515: 
1.26      crook    8516: The search path is initialized when you start Gforth (@pxref{Invoking
1.75      anton    8517: Gforth}). You can display it and change it using @code{fpath} in
                   8518: combination with the general path handling words.
1.5       anton    8519: 
1.75      anton    8520: doc-fpath
                   8521: @c the functionality of the following words is easily available through
                   8522: @c   fpath and the general path words.  The may go away.
                   8523: @c doc-.fpath
                   8524: @c doc-fpath+
                   8525: @c doc-fpath=
                   8526: @c doc-open-fpath-file
1.44      crook    8527: 
                   8528: @noindent
1.26      crook    8529: Here is an example of using @code{fpath} and @code{require}:
1.5       anton    8530: 
1.26      crook    8531: @example
1.75      anton    8532: fpath path= /usr/lib/forth/|./
1.26      crook    8533: require timer.fs
                   8534: @end example
1.5       anton    8535: 
1.75      anton    8536: 
1.26      crook    8537: @c ---------------------------------------------------------
1.75      anton    8538: @node General Search Paths,  , Source Search Paths, Search Paths
1.26      crook    8539: @subsubsection General Search Paths
1.75      anton    8540: @cindex search path control, source files
1.5       anton    8541: 
1.26      crook    8542: Your application may need to search files in several directories, like
                   8543: @code{included} does. To facilitate this, Gforth allows you to define
                   8544: and use your own search paths, by providing generic equivalents of the
                   8545: Forth search path words:
1.5       anton    8546: 
1.75      anton    8547: doc-open-path-file
                   8548: doc-path-allot
                   8549: doc-clear-path
                   8550: doc-also-path
1.26      crook    8551: doc-.path
                   8552: doc-path+
                   8553: doc-path=
1.5       anton    8554: 
1.75      anton    8555: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
1.44      crook    8556: 
1.75      anton    8557: Here's an example of creating an empty search path:
                   8558: @c
1.26      crook    8559: @example
1.75      anton    8560: create mypath 500 path-allot \ maximum length 500 chars (is checked)
1.26      crook    8561: @end example
1.5       anton    8562: 
1.26      crook    8563: @c -------------------------------------------------------------
                   8564: @node Blocks, Other I/O, Files, Words
                   8565: @section Blocks
1.28      crook    8566: @cindex I/O - blocks
                   8567: @cindex blocks
                   8568: 
                   8569: When you run Gforth on a modern desk-top computer, it runs under the
                   8570: control of an operating system which provides certain services.  One of
                   8571: these services is @var{file services}, which allows Forth source code
                   8572: and data to be stored in files and read into Gforth (@pxref{Files}).
                   8573: 
                   8574: Traditionally, Forth has been an important programming language on
                   8575: systems where it has interfaced directly to the underlying hardware with
                   8576: no intervening operating system. Forth provides a mechanism, called
1.29      crook    8577: @dfn{blocks}, for accessing mass storage on such systems.
1.28      crook    8578: 
                   8579: A block is a 1024-byte data area, which can be used to hold data or
                   8580: Forth source code. No structure is imposed on the contents of the
                   8581: block. A block is identified by its number; blocks are numbered
                   8582: contiguously from 1 to an implementation-defined maximum.
                   8583: 
                   8584: A typical system that used blocks but no operating system might use a
                   8585: single floppy-disk drive for mass storage, with the disks formatted to
                   8586: provide 256-byte sectors. Blocks would be implemented by assigning the
                   8587: first four sectors of the disk to block 1, the second four sectors to
                   8588: block 2 and so on, up to the limit of the capacity of the disk. The disk
                   8589: would not contain any file system information, just the set of blocks.
                   8590: 
1.29      crook    8591: @cindex blocks file
1.28      crook    8592: On systems that do provide file services, blocks are typically
1.29      crook    8593: implemented by storing a sequence of blocks within a single @dfn{blocks
1.28      crook    8594: file}.  The size of the blocks file will be an exact multiple of 1024
                   8595: bytes, corresponding to the number of blocks it contains. This is the
                   8596: mechanism that Gforth uses.
                   8597: 
1.29      crook    8598: @cindex @file{blocks.fb}
1.75      anton    8599: Only one blocks file can be open at a time. If you use block words without
1.28      crook    8600: having specified a blocks file, Gforth defaults to the blocks file
                   8601: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
1.75      anton    8602: locate a blocks file (@pxref{Source Search Paths}).
1.28      crook    8603: 
1.29      crook    8604: @cindex block buffers
1.28      crook    8605: When you read and write blocks under program control, Gforth uses a
1.29      crook    8606: number of @dfn{block buffers} as intermediate storage. These buffers are
1.28      crook    8607: not used when you use @code{load} to interpret the contents of a block.
                   8608: 
1.75      anton    8609: The behaviour of the block buffers is analagous to that of a cache.
                   8610: Each block buffer has three states:
1.28      crook    8611: 
                   8612: @itemize @bullet
                   8613: @item
                   8614: Unassigned
                   8615: @item
                   8616: Assigned-clean
                   8617: @item
                   8618: Assigned-dirty
                   8619: @end itemize
                   8620: 
1.29      crook    8621: Initially, all block buffers are @i{unassigned}. In order to access a
1.28      crook    8622: block, the block (specified by its block number) must be assigned to a
                   8623: block buffer.
                   8624: 
                   8625: The assignment of a block to a block buffer is performed by @code{block}
                   8626: or @code{buffer}. Use @code{block} when you wish to modify the existing
                   8627: contents of a block. Use @code{buffer} when you don't care about the
                   8628: existing contents of the block@footnote{The ANS Forth definition of
1.35      anton    8629: @code{buffer} is intended not to cause disk I/O; if the data associated
1.28      crook    8630: with the particular block is already stored in a block buffer due to an
                   8631: earlier @code{block} command, @code{buffer} will return that block
                   8632: buffer and the existing contents of the block will be
                   8633: available. Otherwise, @code{buffer} will simply assign a new, empty
1.29      crook    8634: block buffer for the block.}.
1.28      crook    8635: 
1.47      crook    8636: Once a block has been assigned to a block buffer using @code{block} or
1.75      anton    8637: @code{buffer}, that block buffer becomes the @i{current block
                   8638: buffer}. Data may only be manipulated (read or written) within the
                   8639: current block buffer.
1.47      crook    8640: 
                   8641: When the contents of the current block buffer has been modified it is
1.48      anton    8642: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
1.75      anton    8643: either abandon the changes (by doing nothing) or mark the block as
                   8644: changed (assigned-dirty), using @code{update}. Using @code{update} does
                   8645: not change the blocks file; it simply changes a block buffer's state to
                   8646: @i{assigned-dirty}.  The block will be written implicitly when it's
                   8647: buffer is needed for another block, or explicitly by @code{flush} or
                   8648: @code{save-buffers}.
                   8649: 
                   8650: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
                   8651: blocks file on disk. Leaving Gforth with @code{bye} also performs a
                   8652: @code{flush}.
1.28      crook    8653: 
1.29      crook    8654: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
1.28      crook    8655: algorithm to assign a block buffer to a block. That means that any
                   8656: particular block can only be assigned to one specific block buffer,
1.29      crook    8657: called (for the particular operation) the @i{victim buffer}. If the
1.47      crook    8658: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
                   8659: the new block immediately. If it is @i{assigned-dirty} its current
                   8660: contents are written back to the blocks file on disk before it is
1.28      crook    8661: allocated to the new block.
                   8662: 
                   8663: Although no structure is imposed on the contents of a block, it is
                   8664: traditional to display the contents as 16 lines each of 64 characters.  A
                   8665: block provides a single, continuous stream of input (for example, it
                   8666: acts as a single parse area) -- there are no end-of-line characters
                   8667: within a block, and no end-of-file character at the end of a
                   8668: block. There are two consequences of this:
1.26      crook    8669: 
1.28      crook    8670: @itemize @bullet
                   8671: @item
                   8672: The last character of one line wraps straight into the first character
                   8673: of the following line
                   8674: @item
                   8675: The word @code{\} -- comment to end of line -- requires special
                   8676: treatment; in the context of a block it causes all characters until the
                   8677: end of the current 64-character ``line'' to be ignored.
                   8678: @end itemize
                   8679: 
                   8680: In Gforth, when you use @code{block} with a non-existent block number,
1.45      crook    8681: the current blocks file will be extended to the appropriate size and the
1.28      crook    8682: block buffer will be initialised with spaces.
                   8683: 
1.47      crook    8684: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
                   8685: for details) but doesn't encourage the use of blocks; the mechanism is
                   8686: only provided for backward compatibility -- ANS Forth requires blocks to
                   8687: be available when files are.
1.28      crook    8688: 
                   8689: Common techniques that are used when working with blocks include:
                   8690: 
                   8691: @itemize @bullet
                   8692: @item
                   8693: A screen editor that allows you to edit blocks without leaving the Forth
                   8694: environment.
                   8695: @item
                   8696: Shadow screens; where every code block has an associated block
                   8697: containing comments (for example: code in odd block numbers, comments in
                   8698: even block numbers). Typically, the block editor provides a convenient
                   8699: mechanism to toggle between code and comments.
                   8700: @item
                   8701: Load blocks; a single block (typically block 1) contains a number of
                   8702: @code{thru} commands which @code{load} the whole of the application.
                   8703: @end itemize
1.26      crook    8704: 
1.29      crook    8705: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
                   8706: integrated into a Forth programming environment.
1.26      crook    8707: 
                   8708: @comment TODO what about errors on open-blocks?
1.44      crook    8709: 
1.26      crook    8710: doc-open-blocks
                   8711: doc-use
1.75      anton    8712: doc-block-offset
1.26      crook    8713: doc-get-block-fid
                   8714: doc-block-position
1.28      crook    8715: 
1.75      anton    8716: doc-list
1.28      crook    8717: doc-scr
                   8718: 
1.45      crook    8719: doc---gforthman-block
1.28      crook    8720: doc-buffer
                   8721: 
1.75      anton    8722: doc-empty-buffers
                   8723: doc-empty-buffer
1.26      crook    8724: doc-update
1.28      crook    8725: doc-updated?
1.26      crook    8726: doc-save-buffers
1.75      anton    8727: doc-save-buffer
1.26      crook    8728: doc-flush
1.28      crook    8729: 
1.26      crook    8730: doc-load
                   8731: doc-thru
                   8732: doc-+load
                   8733: doc-+thru
1.45      crook    8734: doc---gforthman--->
1.26      crook    8735: doc-block-included
                   8736: 
1.44      crook    8737: 
1.26      crook    8738: @c -------------------------------------------------------------
1.78      anton    8739: @node Other I/O, Locals, Blocks, Words
1.26      crook    8740: @section Other I/O
1.28      crook    8741: @cindex I/O - keyboard and display
1.26      crook    8742: 
                   8743: @menu
                   8744: * Simple numeric output::       Predefined formats
                   8745: * Formatted numeric output::    Formatted (pictured) output
                   8746: * String Formats::              How Forth stores strings in memory
1.67      anton    8747: * Displaying characters and strings::  Other stuff
1.26      crook    8748: * Input::                       Input
                   8749: @end menu
                   8750: 
                   8751: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
                   8752: @subsection Simple numeric output
1.28      crook    8753: @cindex numeric output - simple/free-format
1.5       anton    8754: 
1.26      crook    8755: The simplest output functions are those that display numbers from the
                   8756: data or floating-point stacks. Floating-point output is always displayed
                   8757: using base 10. Numbers displayed from the data stack use the value stored
                   8758: in @code{base}.
1.5       anton    8759: 
1.44      crook    8760: 
1.26      crook    8761: doc-.
                   8762: doc-dec.
                   8763: doc-hex.
                   8764: doc-u.
                   8765: doc-.r
                   8766: doc-u.r
                   8767: doc-d.
                   8768: doc-ud.
                   8769: doc-d.r
                   8770: doc-ud.r
                   8771: doc-f.
                   8772: doc-fe.
                   8773: doc-fs.
1.5       anton    8774: 
1.44      crook    8775: 
1.26      crook    8776: Examples of printing the number 1234.5678E23 in the different floating-point output
                   8777: formats are shown below:
1.5       anton    8778: 
                   8779: @example
1.26      crook    8780: f. 123456779999999000000000000.
                   8781: fe. 123.456779999999E24
                   8782: fs. 1.23456779999999E26
1.5       anton    8783: @end example
                   8784: 
                   8785: 
1.26      crook    8786: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
                   8787: @subsection Formatted numeric output
1.28      crook    8788: @cindex formatted numeric output
1.26      crook    8789: @cindex pictured numeric output
1.28      crook    8790: @cindex numeric output - formatted
1.26      crook    8791: 
1.29      crook    8792: Forth traditionally uses a technique called @dfn{pictured numeric
1.26      crook    8793: output} for formatted printing of integers.  In this technique, digits
                   8794: are extracted from the number (using the current output radix defined by
                   8795: @code{base}), converted to ASCII codes and appended to a string that is
                   8796: built in a scratch-pad area of memory (@pxref{core-idef,
                   8797: Implementation-defined options, Implementation-defined
                   8798: options}). Arbitrary characters can be appended to the string during the
                   8799: extraction process. The completed string is specified by an address
                   8800: and length and can be manipulated (@code{TYPE}ed, copied, modified)
                   8801: under program control.
1.5       anton    8802: 
1.75      anton    8803: All of the integer output words described in the previous section
                   8804: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
                   8805: numeric output.
1.5       anton    8806: 
1.47      crook    8807: Three important things to remember about pictured numeric output:
1.5       anton    8808: 
1.26      crook    8809: @itemize @bullet
                   8810: @item
1.28      crook    8811: It always operates on double-precision numbers; to display a
1.49      anton    8812: single-precision number, convert it first (for ways of doing this
                   8813: @pxref{Double precision}).
1.26      crook    8814: @item
1.28      crook    8815: It always treats the double-precision number as though it were
                   8816: unsigned. The examples below show ways of printing signed numbers.
1.26      crook    8817: @item
                   8818: The string is built up from right to left; least significant digit first.
                   8819: @end itemize
1.5       anton    8820: 
1.44      crook    8821: 
1.26      crook    8822: doc-<#
1.47      crook    8823: doc-<<#
1.26      crook    8824: doc-#
                   8825: doc-#s
                   8826: doc-hold
                   8827: doc-sign
                   8828: doc-#>
1.47      crook    8829: doc-#>>
1.5       anton    8830: 
1.26      crook    8831: doc-represent
1.5       anton    8832: 
1.44      crook    8833: 
                   8834: @noindent
1.26      crook    8835: Here are some examples of using pictured numeric output:
1.5       anton    8836: 
                   8837: @example
1.26      crook    8838: : my-u. ( u -- )
                   8839:   \ Simplest use of pns.. behaves like Standard u. 
                   8840:   0              \ convert to unsigned double
1.75      anton    8841:   <<#            \ start conversion
1.26      crook    8842:   #s             \ convert all digits
                   8843:   #>             \ complete conversion
1.75      anton    8844:   TYPE SPACE     \ display, with trailing space
                   8845:   #>> ;          \ release hold area
1.5       anton    8846: 
1.26      crook    8847: : cents-only ( u -- )
                   8848:   0              \ convert to unsigned double
1.75      anton    8849:   <<#            \ start conversion
1.26      crook    8850:   # #            \ convert two least-significant digits
                   8851:   #>             \ complete conversion, discard other digits
1.75      anton    8852:   TYPE SPACE     \ display, with trailing space
                   8853:   #>> ;          \ release hold area
1.5       anton    8854: 
1.26      crook    8855: : dollars-and-cents ( u -- )
                   8856:   0              \ convert to unsigned double
1.75      anton    8857:   <<#            \ start conversion
1.26      crook    8858:   # #            \ convert two least-significant digits
                   8859:   [char] . hold  \ insert decimal point
                   8860:   #s             \ convert remaining digits
                   8861:   [char] $ hold  \ append currency symbol
                   8862:   #>             \ complete conversion
1.75      anton    8863:   TYPE SPACE     \ display, with trailing space
                   8864:   #>> ;          \ release hold area
1.5       anton    8865: 
1.26      crook    8866: : my-. ( n -- )
                   8867:   \ handling negatives.. behaves like Standard .
                   8868:   s>d            \ convert to signed double
                   8869:   swap over dabs \ leave sign byte followed by unsigned double
1.75      anton    8870:   <<#            \ start conversion
1.26      crook    8871:   #s             \ convert all digits
                   8872:   rot sign       \ get at sign byte, append "-" if needed
                   8873:   #>             \ complete conversion
1.75      anton    8874:   TYPE SPACE     \ display, with trailing space
                   8875:   #>> ;          \ release hold area
1.5       anton    8876: 
1.26      crook    8877: : account. ( n -- )
1.75      anton    8878:   \ accountants don't like minus signs, they use parentheses
1.26      crook    8879:   \ for negative numbers
                   8880:   s>d            \ convert to signed double
                   8881:   swap over dabs \ leave sign byte followed by unsigned double
1.75      anton    8882:   <<#            \ start conversion
1.26      crook    8883:   2 pick         \ get copy of sign byte
                   8884:   0< IF [char] ) hold THEN \ right-most character of output
                   8885:   #s             \ convert all digits
                   8886:   rot            \ get at sign byte
                   8887:   0< IF [char] ( hold THEN
                   8888:   #>             \ complete conversion
1.75      anton    8889:   TYPE SPACE     \ display, with trailing space
                   8890:   #>> ;          \ release hold area
                   8891: 
1.5       anton    8892: @end example
                   8893: 
1.26      crook    8894: Here are some examples of using these words:
1.5       anton    8895: 
                   8896: @example
1.26      crook    8897: 1 my-u. 1
                   8898: hex -1 my-u. decimal FFFFFFFF
                   8899: 1 cents-only 01
                   8900: 1234 cents-only 34
                   8901: 2 dollars-and-cents $0.02
                   8902: 1234 dollars-and-cents $12.34
                   8903: 123 my-. 123
                   8904: -123 my. -123
                   8905: 123 account. 123
                   8906: -456 account. (456)
1.5       anton    8907: @end example
                   8908: 
                   8909: 
1.26      crook    8910: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
                   8911: @subsection String Formats
1.27      crook    8912: @cindex strings - see character strings
                   8913: @cindex character strings - formats
1.28      crook    8914: @cindex I/O - see character strings
1.75      anton    8915: @cindex counted strings
                   8916: 
                   8917: @c anton: this does not really belong here; maybe the memory section,
                   8918: @c  or the principles chapter
1.26      crook    8919: 
1.27      crook    8920: Forth commonly uses two different methods for representing character
                   8921: strings:
1.26      crook    8922: 
                   8923: @itemize @bullet
                   8924: @item
                   8925: @cindex address of counted string
1.45      crook    8926: @cindex counted string
1.29      crook    8927: As a @dfn{counted string}, represented by a @i{c-addr}. The char
                   8928: addressed by @i{c-addr} contains a character-count, @i{n}, of the
                   8929: string and the string occupies the subsequent @i{n} char addresses in
1.26      crook    8930: memory.
                   8931: @item
1.29      crook    8932: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
                   8933: of the string in characters, and @i{c-addr} is the address of the
1.26      crook    8934: first byte of the string.
                   8935: @end itemize
                   8936: 
                   8937: ANS Forth encourages the use of the second format when representing
1.75      anton    8938: strings.
1.26      crook    8939: 
1.44      crook    8940: 
1.26      crook    8941: doc-count
                   8942: 
1.44      crook    8943: 
1.49      anton    8944: For words that move, copy and search for strings see @ref{Memory
                   8945: Blocks}. For words that display characters and strings see
                   8946: @ref{Displaying characters and strings}.
1.26      crook    8947: 
                   8948: @node Displaying characters and strings, Input, String Formats, Other I/O
                   8949: @subsection Displaying characters and strings
1.27      crook    8950: @cindex characters - compiling and displaying
                   8951: @cindex character strings - compiling and displaying
1.26      crook    8952: 
                   8953: This section starts with a glossary of Forth words and ends with a set
                   8954: of examples.
                   8955: 
1.44      crook    8956: 
1.26      crook    8957: doc-bl
                   8958: doc-space
                   8959: doc-spaces
                   8960: doc-emit
                   8961: doc-toupper
                   8962: doc-."
                   8963: doc-.(
                   8964: doc-type
1.44      crook    8965: doc-typewhite
1.26      crook    8966: doc-cr
1.27      crook    8967: @cindex cursor control
1.26      crook    8968: doc-at-xy
                   8969: doc-page
                   8970: doc-s"
                   8971: doc-c"
                   8972: doc-char
                   8973: doc-[char]
                   8974: 
1.44      crook    8975: 
                   8976: @noindent
1.26      crook    8977: As an example, consider the following text, stored in a file @file{test.fs}:
1.5       anton    8978: 
                   8979: @example
1.26      crook    8980: .( text-1)
                   8981: : my-word
                   8982:   ." text-2" cr
                   8983:   .( text-3)
                   8984: ;
                   8985: 
                   8986: ." text-4"
                   8987: 
                   8988: : my-char
                   8989:   [char] ALPHABET emit
                   8990:   char emit
                   8991: ;
1.5       anton    8992: @end example
                   8993: 
1.26      crook    8994: When you load this code into Gforth, the following output is generated:
1.5       anton    8995: 
1.26      crook    8996: @example
1.30      anton    8997: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
1.26      crook    8998: @end example
1.5       anton    8999: 
1.26      crook    9000: @itemize @bullet
                   9001: @item
                   9002: Messages @code{text-1} and @code{text-3} are displayed because @code{.(} 
                   9003: is an immediate word; it behaves in the same way whether it is used inside
                   9004: or outside a colon definition.
                   9005: @item
                   9006: Message @code{text-4} is displayed because of Gforth's added interpretation
                   9007: semantics for @code{."}.
                   9008: @item
1.29      crook    9009: Message @code{text-2} is @i{not} displayed, because the text interpreter
1.26      crook    9010: performs the compilation semantics for @code{."} within the definition of
                   9011: @code{my-word}.
                   9012: @end itemize
1.5       anton    9013: 
1.26      crook    9014: Here are some examples of executing @code{my-word} and @code{my-char}:
1.5       anton    9015: 
1.26      crook    9016: @example
1.30      anton    9017: @kbd{my-word @key{RET}} text-2
1.26      crook    9018:  ok
1.30      anton    9019: @kbd{my-char fred @key{RET}} Af ok
                   9020: @kbd{my-char jim @key{RET}} Aj ok
1.26      crook    9021: @end example
1.5       anton    9022: 
                   9023: @itemize @bullet
                   9024: @item
1.26      crook    9025: Message @code{text-2} is displayed because of the run-time behaviour of
                   9026: @code{."}.
                   9027: @item
                   9028: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
                   9029: on the stack at run-time. @code{emit} always displays the character
                   9030: when @code{my-char} is executed.
                   9031: @item
                   9032: @code{char} parses a string at run-time and the second @code{emit} displays
                   9033: the first character of the string.
1.5       anton    9034: @item
1.26      crook    9035: If you type @code{see my-char} you can see that @code{[char]} discarded
                   9036: the text ``LPHABET'' and only compiled the display code for ``A'' into the
                   9037: definition of @code{my-char}.
1.5       anton    9038: @end itemize
                   9039: 
                   9040: 
                   9041: 
1.48      anton    9042: @node Input,  , Displaying characters and strings, Other I/O
1.26      crook    9043: @subsection Input
                   9044: @cindex input
1.28      crook    9045: @cindex I/O - see input
                   9046: @cindex parsing a string
1.5       anton    9047: 
1.49      anton    9048: For ways of storing character strings in memory see @ref{String Formats}.
1.5       anton    9049: 
1.27      crook    9050: @comment TODO examples for >number >float accept key key? pad parse word refill
1.29      crook    9051: @comment then index them
1.27      crook    9052: 
1.44      crook    9053: 
1.27      crook    9054: doc-key
                   9055: doc-key?
1.45      crook    9056: doc-ekey
                   9057: doc-ekey?
                   9058: doc-ekey>char
1.26      crook    9059: doc->number
                   9060: doc->float
                   9061: doc-accept
1.27      crook    9062: doc-pad
1.75      anton    9063: @c anton: these belong in the input stream section
1.27      crook    9064: doc-parse
                   9065: doc-word
                   9066: doc-sword
1.75      anton    9067: doc-name
1.27      crook    9068: doc-refill
                   9069: @comment obsolescent words..
                   9070: doc-convert
1.26      crook    9071: doc-query
                   9072: doc-expect
1.27      crook    9073: doc-span
1.5       anton    9074: 
                   9075: 
1.78      anton    9076: @c -------------------------------------------------------------
                   9077: @node Locals, Structures, Other I/O, Words
                   9078: @section Locals
                   9079: @cindex locals
                   9080: 
                   9081: Local variables can make Forth programming more enjoyable and Forth
                   9082: programs easier to read. Unfortunately, the locals of ANS Forth are
                   9083: laden with restrictions. Therefore, we provide not only the ANS Forth
                   9084: locals wordset, but also our own, more powerful locals wordset (we
                   9085: implemented the ANS Forth locals wordset through our locals wordset).
1.44      crook    9086: 
1.78      anton    9087: The ideas in this section have also been published in M. Anton Ertl,
                   9088: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
                   9089: Automatic Scoping of Local Variables}}, EuroForth '94.
1.12      anton    9090: 
                   9091: @menu
1.78      anton    9092: * Gforth locals::               
                   9093: * ANS Forth locals::            
1.5       anton    9094: @end menu
                   9095: 
1.78      anton    9096: @node Gforth locals, ANS Forth locals, Locals, Locals
                   9097: @subsection Gforth locals
                   9098: @cindex Gforth locals
                   9099: @cindex locals, Gforth style
1.5       anton    9100: 
1.78      anton    9101: Locals can be defined with
1.44      crook    9102: 
1.78      anton    9103: @example
                   9104: @{ local1 local2 ... -- comment @}
                   9105: @end example
                   9106: or
                   9107: @example
                   9108: @{ local1 local2 ... @}
                   9109: @end example
1.5       anton    9110: 
1.78      anton    9111: E.g.,
                   9112: @example
                   9113: : max @{ n1 n2 -- n3 @}
                   9114:  n1 n2 > if
                   9115:    n1
                   9116:  else
                   9117:    n2
                   9118:  endif ;
                   9119: @end example
1.44      crook    9120: 
1.78      anton    9121: The similarity of locals definitions with stack comments is intended. A
                   9122: locals definition often replaces the stack comment of a word. The order
                   9123: of the locals corresponds to the order in a stack comment and everything
                   9124: after the @code{--} is really a comment.
1.77      anton    9125: 
1.78      anton    9126: This similarity has one disadvantage: It is too easy to confuse locals
                   9127: declarations with stack comments, causing bugs and making them hard to
                   9128: find. However, this problem can be avoided by appropriate coding
                   9129: conventions: Do not use both notations in the same program. If you do,
                   9130: they should be distinguished using additional means, e.g. by position.
1.77      anton    9131: 
1.78      anton    9132: @cindex types of locals
                   9133: @cindex locals types
                   9134: The name of the local may be preceded by a type specifier, e.g.,
                   9135: @code{F:} for a floating point value:
1.5       anton    9136: 
1.78      anton    9137: @example
                   9138: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
                   9139: \ complex multiplication
                   9140:  Ar Br f* Ai Bi f* f-
                   9141:  Ar Bi f* Ai Br f* f+ ;
                   9142: @end example
1.44      crook    9143: 
1.78      anton    9144: @cindex flavours of locals
                   9145: @cindex locals flavours
                   9146: @cindex value-flavoured locals
                   9147: @cindex variable-flavoured locals
                   9148: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
                   9149: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
                   9150: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
                   9151: with @code{W:}, @code{D:} etc.) produces its value and can be changed
                   9152: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
                   9153: produces its address (which becomes invalid when the variable's scope is
                   9154: left). E.g., the standard word @code{emit} can be defined in terms of
                   9155: @code{type} like this:
1.5       anton    9156: 
1.78      anton    9157: @example
                   9158: : emit @{ C^ char* -- @}
                   9159:     char* 1 type ;
                   9160: @end example
1.5       anton    9161: 
1.78      anton    9162: @cindex default type of locals
                   9163: @cindex locals, default type
                   9164: A local without type specifier is a @code{W:} local. Both flavours of
                   9165: locals are initialized with values from the data or FP stack.
1.44      crook    9166: 
1.78      anton    9167: Currently there is no way to define locals with user-defined data
                   9168: structures, but we are working on it.
1.5       anton    9169: 
1.78      anton    9170: Gforth allows defining locals everywhere in a colon definition. This
                   9171: poses the following questions:
1.5       anton    9172: 
1.78      anton    9173: @menu
                   9174: * Where are locals visible by name?::  
                   9175: * How long do locals live?::    
                   9176: * Locals programming style::    
                   9177: * Locals implementation::       
                   9178: @end menu
1.44      crook    9179: 
1.78      anton    9180: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
                   9181: @subsubsection Where are locals visible by name?
                   9182: @cindex locals visibility
                   9183: @cindex visibility of locals
                   9184: @cindex scope of locals
1.5       anton    9185: 
1.78      anton    9186: Basically, the answer is that locals are visible where you would expect
                   9187: it in block-structured languages, and sometimes a little longer. If you
                   9188: want to restrict the scope of a local, enclose its definition in
                   9189: @code{SCOPE}...@code{ENDSCOPE}.
1.5       anton    9190: 
                   9191: 
1.78      anton    9192: doc-scope
                   9193: doc-endscope
1.5       anton    9194: 
                   9195: 
1.78      anton    9196: These words behave like control structure words, so you can use them
                   9197: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
                   9198: arbitrary ways.
1.77      anton    9199: 
1.78      anton    9200: If you want a more exact answer to the visibility question, here's the
                   9201: basic principle: A local is visible in all places that can only be
                   9202: reached through the definition of the local@footnote{In compiler
                   9203: construction terminology, all places dominated by the definition of the
                   9204: local.}. In other words, it is not visible in places that can be reached
                   9205: without going through the definition of the local. E.g., locals defined
                   9206: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
                   9207: defined in @code{BEGIN}...@code{UNTIL} are visible after the
                   9208: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1.77      anton    9209: 
1.78      anton    9210: The reasoning behind this solution is: We want to have the locals
                   9211: visible as long as it is meaningful. The user can always make the
                   9212: visibility shorter by using explicit scoping. In a place that can
                   9213: only be reached through the definition of a local, the meaning of a
                   9214: local name is clear. In other places it is not: How is the local
                   9215: initialized at the control flow path that does not contain the
                   9216: definition? Which local is meant, if the same name is defined twice in
                   9217: two independent control flow paths?
1.77      anton    9218: 
1.78      anton    9219: This should be enough detail for nearly all users, so you can skip the
                   9220: rest of this section. If you really must know all the gory details and
                   9221: options, read on.
1.77      anton    9222: 
1.78      anton    9223: In order to implement this rule, the compiler has to know which places
                   9224: are unreachable. It knows this automatically after @code{AHEAD},
                   9225: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
                   9226: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
                   9227: compiler that the control flow never reaches that place. If
                   9228: @code{UNREACHABLE} is not used where it could, the only consequence is
                   9229: that the visibility of some locals is more limited than the rule above
                   9230: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
                   9231: lie to the compiler), buggy code will be produced.
1.77      anton    9232: 
1.5       anton    9233: 
1.78      anton    9234: doc-unreachable
1.5       anton    9235: 
1.23      crook    9236: 
1.78      anton    9237: Another problem with this rule is that at @code{BEGIN}, the compiler
                   9238: does not know which locals will be visible on the incoming
                   9239: back-edge. All problems discussed in the following are due to this
                   9240: ignorance of the compiler (we discuss the problems using @code{BEGIN}
                   9241: loops as examples; the discussion also applies to @code{?DO} and other
                   9242: loops). Perhaps the most insidious example is:
1.26      crook    9243: @example
1.78      anton    9244: AHEAD
                   9245: BEGIN
                   9246:   x
                   9247: [ 1 CS-ROLL ] THEN
                   9248:   @{ x @}
                   9249:   ...
                   9250: UNTIL
1.26      crook    9251: @end example
1.23      crook    9252: 
1.78      anton    9253: This should be legal according to the visibility rule. The use of
                   9254: @code{x} can only be reached through the definition; but that appears
                   9255: textually below the use.
                   9256: 
                   9257: From this example it is clear that the visibility rules cannot be fully
                   9258: implemented without major headaches. Our implementation treats common
                   9259: cases as advertised and the exceptions are treated in a safe way: The
                   9260: compiler makes a reasonable guess about the locals visible after a
                   9261: @code{BEGIN}; if it is too pessimistic, the
                   9262: user will get a spurious error about the local not being defined; if the
                   9263: compiler is too optimistic, it will notice this later and issue a
                   9264: warning. In the case above the compiler would complain about @code{x}
                   9265: being undefined at its use. You can see from the obscure examples in
                   9266: this section that it takes quite unusual control structures to get the
                   9267: compiler into trouble, and even then it will often do fine.
1.23      crook    9268: 
1.78      anton    9269: If the @code{BEGIN} is reachable from above, the most optimistic guess
                   9270: is that all locals visible before the @code{BEGIN} will also be
                   9271: visible after the @code{BEGIN}. This guess is valid for all loops that
                   9272: are entered only through the @code{BEGIN}, in particular, for normal
                   9273: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
                   9274: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
                   9275: compiler. When the branch to the @code{BEGIN} is finally generated by
                   9276: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
                   9277: warns the user if it was too optimistic:
1.26      crook    9278: @example
1.78      anton    9279: IF
                   9280:   @{ x @}
                   9281: BEGIN
                   9282:   \ x ? 
                   9283: [ 1 cs-roll ] THEN
                   9284:   ...
                   9285: UNTIL
1.26      crook    9286: @end example
1.23      crook    9287: 
1.78      anton    9288: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
                   9289: optimistically assumes that it lives until the @code{THEN}. It notices
                   9290: this difference when it compiles the @code{UNTIL} and issues a
                   9291: warning. The user can avoid the warning, and make sure that @code{x}
                   9292: is not used in the wrong area by using explicit scoping:
                   9293: @example
                   9294: IF
                   9295:   SCOPE
                   9296:   @{ x @}
                   9297:   ENDSCOPE
                   9298: BEGIN
                   9299: [ 1 cs-roll ] THEN
                   9300:   ...
                   9301: UNTIL
                   9302: @end example
1.23      crook    9303: 
1.78      anton    9304: Since the guess is optimistic, there will be no spurious error messages
                   9305: about undefined locals.
1.44      crook    9306: 
1.78      anton    9307: If the @code{BEGIN} is not reachable from above (e.g., after
                   9308: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
                   9309: optimistic guess, as the locals visible after the @code{BEGIN} may be
                   9310: defined later. Therefore, the compiler assumes that no locals are
                   9311: visible after the @code{BEGIN}. However, the user can use
                   9312: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
                   9313: visible at the BEGIN as at the point where the top control-flow stack
                   9314: item was created.
1.23      crook    9315: 
1.44      crook    9316: 
1.78      anton    9317: doc-assume-live
1.26      crook    9318: 
1.23      crook    9319: 
1.78      anton    9320: @noindent
                   9321: E.g.,
                   9322: @example
                   9323: @{ x @}
                   9324: AHEAD
                   9325: ASSUME-LIVE
                   9326: BEGIN
                   9327:   x
                   9328: [ 1 CS-ROLL ] THEN
                   9329:   ...
                   9330: UNTIL
                   9331: @end example
1.44      crook    9332: 
1.78      anton    9333: Other cases where the locals are defined before the @code{BEGIN} can be
                   9334: handled by inserting an appropriate @code{CS-ROLL} before the
                   9335: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
                   9336: behind the @code{ASSUME-LIVE}).
1.23      crook    9337: 
1.78      anton    9338: Cases where locals are defined after the @code{BEGIN} (but should be
                   9339: visible immediately after the @code{BEGIN}) can only be handled by
                   9340: rearranging the loop. E.g., the ``most insidious'' example above can be
                   9341: arranged into:
                   9342: @example
                   9343: BEGIN
                   9344:   @{ x @}
                   9345:   ... 0=
                   9346: WHILE
                   9347:   x
                   9348: REPEAT
                   9349: @end example
1.44      crook    9350: 
1.78      anton    9351: @node How long do locals live?, Locals programming style, Where are locals visible by name?, Gforth locals
                   9352: @subsubsection How long do locals live?
                   9353: @cindex locals lifetime
                   9354: @cindex lifetime of locals
1.23      crook    9355: 
1.78      anton    9356: The right answer for the lifetime question would be: A local lives at
                   9357: least as long as it can be accessed. For a value-flavoured local this
                   9358: means: until the end of its visibility. However, a variable-flavoured
                   9359: local could be accessed through its address far beyond its visibility
                   9360: scope. Ultimately, this would mean that such locals would have to be
                   9361: garbage collected. Since this entails un-Forth-like implementation
                   9362: complexities, I adopted the same cowardly solution as some other
                   9363: languages (e.g., C): The local lives only as long as it is visible;
                   9364: afterwards its address is invalid (and programs that access it
                   9365: afterwards are erroneous).
1.23      crook    9366: 
1.78      anton    9367: @node Locals programming style, Locals implementation, How long do locals live?, Gforth locals
                   9368: @subsubsection Locals programming style
                   9369: @cindex locals programming style
                   9370: @cindex programming style, locals
1.23      crook    9371: 
1.78      anton    9372: The freedom to define locals anywhere has the potential to change
                   9373: programming styles dramatically. In particular, the need to use the
                   9374: return stack for intermediate storage vanishes. Moreover, all stack
                   9375: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
                   9376: determined arguments) can be eliminated: If the stack items are in the
                   9377: wrong order, just write a locals definition for all of them; then
                   9378: write the items in the order you want.
1.23      crook    9379: 
1.78      anton    9380: This seems a little far-fetched and eliminating stack manipulations is
                   9381: unlikely to become a conscious programming objective. Still, the number
                   9382: of stack manipulations will be reduced dramatically if local variables
                   9383: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
                   9384: a traditional implementation of @code{max}).
1.23      crook    9385: 
1.78      anton    9386: This shows one potential benefit of locals: making Forth programs more
                   9387: readable. Of course, this benefit will only be realized if the
                   9388: programmers continue to honour the principle of factoring instead of
                   9389: using the added latitude to make the words longer.
1.23      crook    9390: 
1.78      anton    9391: @cindex single-assignment style for locals
                   9392: Using @code{TO} can and should be avoided.  Without @code{TO},
                   9393: every value-flavoured local has only a single assignment and many
                   9394: advantages of functional languages apply to Forth. I.e., programs are
                   9395: easier to analyse, to optimize and to read: It is clear from the
                   9396: definition what the local stands for, it does not turn into something
                   9397: different later.
1.23      crook    9398: 
1.78      anton    9399: E.g., a definition using @code{TO} might look like this:
                   9400: @example
                   9401: : strcmp @{ addr1 u1 addr2 u2 -- n @}
                   9402:  u1 u2 min 0
                   9403:  ?do
                   9404:    addr1 c@@ addr2 c@@ -
                   9405:    ?dup-if
                   9406:      unloop exit
                   9407:    then
                   9408:    addr1 char+ TO addr1
                   9409:    addr2 char+ TO addr2
                   9410:  loop
                   9411:  u1 u2 - ;
1.26      crook    9412: @end example
1.78      anton    9413: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
                   9414: every loop iteration. @code{strcmp} is a typical example of the
                   9415: readability problems of using @code{TO}. When you start reading
                   9416: @code{strcmp}, you think that @code{addr1} refers to the start of the
                   9417: string. Only near the end of the loop you realize that it is something
                   9418: else.
1.23      crook    9419: 
1.78      anton    9420: This can be avoided by defining two locals at the start of the loop that
                   9421: are initialized with the right value for the current iteration.
                   9422: @example
                   9423: : strcmp @{ addr1 u1 addr2 u2 -- n @}
                   9424:  addr1 addr2
                   9425:  u1 u2 min 0 
                   9426:  ?do @{ s1 s2 @}
                   9427:    s1 c@@ s2 c@@ -
                   9428:    ?dup-if
                   9429:      unloop exit
                   9430:    then
                   9431:    s1 char+ s2 char+
                   9432:  loop
                   9433:  2drop
                   9434:  u1 u2 - ;
                   9435: @end example
                   9436: Here it is clear from the start that @code{s1} has a different value
                   9437: in every loop iteration.
1.23      crook    9438: 
1.78      anton    9439: @node Locals implementation,  , Locals programming style, Gforth locals
                   9440: @subsubsection Locals implementation
                   9441: @cindex locals implementation
                   9442: @cindex implementation of locals
1.23      crook    9443: 
1.78      anton    9444: @cindex locals stack
                   9445: Gforth uses an extra locals stack. The most compelling reason for
                   9446: this is that the return stack is not float-aligned; using an extra stack
                   9447: also eliminates the problems and restrictions of using the return stack
                   9448: as locals stack. Like the other stacks, the locals stack grows toward
                   9449: lower addresses. A few primitives allow an efficient implementation:
                   9450: 
                   9451: 
                   9452: doc-@local#
                   9453: doc-f@local#
                   9454: doc-laddr#
                   9455: doc-lp+!#
                   9456: doc-lp!
                   9457: doc->l
                   9458: doc-f>l
                   9459: 
                   9460: 
                   9461: In addition to these primitives, some specializations of these
                   9462: primitives for commonly occurring inline arguments are provided for
                   9463: efficiency reasons, e.g., @code{@@local0} as specialization of
                   9464: @code{@@local#} for the inline argument 0. The following compiling words
                   9465: compile the right specialized version, or the general version, as
                   9466: appropriate:
1.23      crook    9467: 
1.5       anton    9468: 
1.78      anton    9469: doc-compile-@local
                   9470: doc-compile-f@local
                   9471: doc-compile-lp+!
1.5       anton    9472: 
                   9473: 
1.78      anton    9474: Combinations of conditional branches and @code{lp+!#} like
                   9475: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
                   9476: is taken) are provided for efficiency and correctness in loops.
1.5       anton    9477: 
1.78      anton    9478: A special area in the dictionary space is reserved for keeping the
                   9479: local variable names. @code{@{} switches the dictionary pointer to this
                   9480: area and @code{@}} switches it back and generates the locals
                   9481: initializing code. @code{W:} etc.@ are normal defining words. This
                   9482: special area is cleared at the start of every colon definition.
1.5       anton    9483: 
1.78      anton    9484: @cindex word list for defining locals
                   9485: A special feature of Gforth's dictionary is used to implement the
                   9486: definition of locals without type specifiers: every word list (aka
                   9487: vocabulary) has its own methods for searching
                   9488: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
                   9489: with a special search method: When it is searched for a word, it
                   9490: actually creates that word using @code{W:}. @code{@{} changes the search
                   9491: order to first search the word list containing @code{@}}, @code{W:} etc.,
                   9492: and then the word list for defining locals without type specifiers.
1.5       anton    9493: 
1.78      anton    9494: The lifetime rules support a stack discipline within a colon
                   9495: definition: The lifetime of a local is either nested with other locals
                   9496: lifetimes or it does not overlap them.
1.23      crook    9497: 
1.78      anton    9498: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
                   9499: pointer manipulation is generated. Between control structure words
                   9500: locals definitions can push locals onto the locals stack. @code{AGAIN}
                   9501: is the simplest of the other three control flow words. It has to
                   9502: restore the locals stack depth of the corresponding @code{BEGIN}
                   9503: before branching. The code looks like this:
                   9504: @format
                   9505: @code{lp+!#} current-locals-size @minus{} dest-locals-size
                   9506: @code{branch} <begin>
                   9507: @end format
1.26      crook    9508: 
1.78      anton    9509: @code{UNTIL} is a little more complicated: If it branches back, it
                   9510: must adjust the stack just like @code{AGAIN}. But if it falls through,
                   9511: the locals stack must not be changed. The compiler generates the
                   9512: following code:
                   9513: @format
                   9514: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
                   9515: @end format
                   9516: The locals stack pointer is only adjusted if the branch is taken.
1.44      crook    9517: 
1.78      anton    9518: @code{THEN} can produce somewhat inefficient code:
                   9519: @format
                   9520: @code{lp+!#} current-locals-size @minus{} orig-locals-size
                   9521: <orig target>:
                   9522: @code{lp+!#} orig-locals-size @minus{} new-locals-size
                   9523: @end format
                   9524: The second @code{lp+!#} adjusts the locals stack pointer from the
                   9525: level at the @i{orig} point to the level after the @code{THEN}. The
                   9526: first @code{lp+!#} adjusts the locals stack pointer from the current
                   9527: level to the level at the orig point, so the complete effect is an
                   9528: adjustment from the current level to the right level after the
                   9529: @code{THEN}.
1.26      crook    9530: 
1.78      anton    9531: @cindex locals information on the control-flow stack
                   9532: @cindex control-flow stack items, locals information
                   9533: In a conventional Forth implementation a dest control-flow stack entry
                   9534: is just the target address and an orig entry is just the address to be
                   9535: patched. Our locals implementation adds a word list to every orig or dest
                   9536: item. It is the list of locals visible (or assumed visible) at the point
                   9537: described by the entry. Our implementation also adds a tag to identify
                   9538: the kind of entry, in particular to differentiate between live and dead
                   9539: (reachable and unreachable) orig entries.
1.26      crook    9540: 
1.78      anton    9541: A few unusual operations have to be performed on locals word lists:
1.44      crook    9542: 
1.5       anton    9543: 
1.78      anton    9544: doc-common-list
                   9545: doc-sub-list?
                   9546: doc-list-size
1.52      anton    9547: 
                   9548: 
1.78      anton    9549: Several features of our locals word list implementation make these
                   9550: operations easy to implement: The locals word lists are organised as
                   9551: linked lists; the tails of these lists are shared, if the lists
                   9552: contain some of the same locals; and the address of a name is greater
                   9553: than the address of the names behind it in the list.
1.5       anton    9554: 
1.78      anton    9555: Another important implementation detail is the variable
                   9556: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
                   9557: determine if they can be reached directly or only through the branch
                   9558: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
                   9559: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
                   9560: definition, by @code{BEGIN} and usually by @code{THEN}.
1.5       anton    9561: 
1.78      anton    9562: Counted loops are similar to other loops in most respects, but
                   9563: @code{LEAVE} requires special attention: It performs basically the same
                   9564: service as @code{AHEAD}, but it does not create a control-flow stack
                   9565: entry. Therefore the information has to be stored elsewhere;
                   9566: traditionally, the information was stored in the target fields of the
                   9567: branches created by the @code{LEAVE}s, by organizing these fields into a
                   9568: linked list. Unfortunately, this clever trick does not provide enough
                   9569: space for storing our extended control flow information. Therefore, we
                   9570: introduce another stack, the leave stack. It contains the control-flow
                   9571: stack entries for all unresolved @code{LEAVE}s.
1.44      crook    9572: 
1.78      anton    9573: Local names are kept until the end of the colon definition, even if
                   9574: they are no longer visible in any control-flow path. In a few cases
                   9575: this may lead to increased space needs for the locals name area, but
                   9576: usually less than reclaiming this space would cost in code size.
1.5       anton    9577: 
1.44      crook    9578: 
1.78      anton    9579: @node ANS Forth locals,  , Gforth locals, Locals
                   9580: @subsection ANS Forth locals
                   9581: @cindex locals, ANS Forth style
1.5       anton    9582: 
1.78      anton    9583: The ANS Forth locals wordset does not define a syntax for locals, but
                   9584: words that make it possible to define various syntaxes. One of the
                   9585: possible syntaxes is a subset of the syntax we used in the Gforth locals
                   9586: wordset, i.e.:
1.29      crook    9587: 
                   9588: @example
1.78      anton    9589: @{ local1 local2 ... -- comment @}
                   9590: @end example
                   9591: @noindent
                   9592: or
                   9593: @example
                   9594: @{ local1 local2 ... @}
1.29      crook    9595: @end example
                   9596: 
1.78      anton    9597: The order of the locals corresponds to the order in a stack comment. The
                   9598: restrictions are:
1.5       anton    9599: 
1.78      anton    9600: @itemize @bullet
                   9601: @item
                   9602: Locals can only be cell-sized values (no type specifiers are allowed).
                   9603: @item
                   9604: Locals can be defined only outside control structures.
                   9605: @item
                   9606: Locals can interfere with explicit usage of the return stack. For the
                   9607: exact (and long) rules, see the standard. If you don't use return stack
                   9608: accessing words in a definition using locals, you will be all right. The
                   9609: purpose of this rule is to make locals implementation on the return
                   9610: stack easier.
                   9611: @item
                   9612: The whole definition must be in one line.
                   9613: @end itemize
1.5       anton    9614: 
1.78      anton    9615: Locals defined in ANS Forth behave like @code{VALUE}s
                   9616: (@pxref{Values}). I.e., they are initialized from the stack. Using their
                   9617: name produces their value. Their value can be changed using @code{TO}.
1.77      anton    9618: 
1.78      anton    9619: Since the syntax above is supported by Gforth directly, you need not do
                   9620: anything to use it. If you want to port a program using this syntax to
                   9621: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
                   9622: syntax on the other system.
1.5       anton    9623: 
1.78      anton    9624: Note that a syntax shown in the standard, section A.13 looks
                   9625: similar, but is quite different in having the order of locals
                   9626: reversed. Beware!
1.5       anton    9627: 
1.78      anton    9628: The ANS Forth locals wordset itself consists of one word:
1.5       anton    9629: 
1.78      anton    9630: doc-(local)
1.5       anton    9631: 
1.78      anton    9632: The ANS Forth locals extension wordset defines a syntax using
                   9633: @code{locals|}, but it is so awful that we strongly recommend not to use
                   9634: it. We have implemented this syntax to make porting to Gforth easy, but
                   9635: do not document it here. The problem with this syntax is that the locals
                   9636: are defined in an order reversed with respect to the standard stack
                   9637: comment notation, making programs harder to read, and easier to misread
                   9638: and miswrite. The only merit of this syntax is that it is easy to
                   9639: implement using the ANS Forth locals wordset.
1.53      anton    9640: 
                   9641: 
1.78      anton    9642: @c ----------------------------------------------------------
                   9643: @node Structures, Object-oriented Forth, Locals, Words
                   9644: @section  Structures
                   9645: @cindex structures
                   9646: @cindex records
1.53      anton    9647: 
1.78      anton    9648: This section presents the structure package that comes with Gforth. A
                   9649: version of the package implemented in ANS Forth is available in
                   9650: @file{compat/struct.fs}. This package was inspired by a posting on
                   9651: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
                   9652: possibly John Hayes). A version of this section has been published in
                   9653: M. Anton Ertl,
                   9654: @uref{http://www.complang.tuwien.ac.at/forth/objects/structs.html, Yet
                   9655: Another Forth Structures Package}, Forth Dimensions 19(3), pages
                   9656: 13--16. Marcel Hendrix provided helpful comments.
1.53      anton    9657: 
1.78      anton    9658: @menu
                   9659: * Why explicit structure support?::  
                   9660: * Structure Usage::             
                   9661: * Structure Naming Convention::  
                   9662: * Structure Implementation::    
                   9663: * Structure Glossary::          
                   9664: @end menu
1.55      anton    9665: 
1.78      anton    9666: @node Why explicit structure support?, Structure Usage, Structures, Structures
                   9667: @subsection Why explicit structure support?
1.53      anton    9668: 
1.78      anton    9669: @cindex address arithmetic for structures
                   9670: @cindex structures using address arithmetic
                   9671: If we want to use a structure containing several fields, we could simply
                   9672: reserve memory for it, and access the fields using address arithmetic
                   9673: (@pxref{Address arithmetic}). As an example, consider a structure with
                   9674: the following fields
1.57      anton    9675: 
1.78      anton    9676: @table @code
                   9677: @item a
                   9678: is a float
                   9679: @item b
                   9680: is a cell
                   9681: @item c
                   9682: is a float
                   9683: @end table
1.57      anton    9684: 
1.78      anton    9685: Given the (float-aligned) base address of the structure we get the
                   9686: address of the field
1.52      anton    9687: 
1.78      anton    9688: @table @code
                   9689: @item a
                   9690: without doing anything further.
                   9691: @item b
                   9692: with @code{float+}
                   9693: @item c
                   9694: with @code{float+ cell+ faligned}
                   9695: @end table
1.52      anton    9696: 
1.78      anton    9697: It is easy to see that this can become quite tiring. 
1.52      anton    9698: 
1.78      anton    9699: Moreover, it is not very readable, because seeing a
                   9700: @code{cell+} tells us neither which kind of structure is
                   9701: accessed nor what field is accessed; we have to somehow infer the kind
                   9702: of structure, and then look up in the documentation, which field of
                   9703: that structure corresponds to that offset.
1.53      anton    9704: 
1.78      anton    9705: Finally, this kind of address arithmetic also causes maintenance
                   9706: troubles: If you add or delete a field somewhere in the middle of the
                   9707: structure, you have to find and change all computations for the fields
                   9708: afterwards.
1.52      anton    9709: 
1.78      anton    9710: So, instead of using @code{cell+} and friends directly, how
                   9711: about storing the offsets in constants:
1.52      anton    9712: 
1.78      anton    9713: @example
                   9714: 0 constant a-offset
                   9715: 0 float+ constant b-offset
                   9716: 0 float+ cell+ faligned c-offset
                   9717: @end example
1.64      pazsan   9718: 
1.78      anton    9719: Now we can get the address of field @code{x} with @code{x-offset
                   9720: +}. This is much better in all respects. Of course, you still
                   9721: have to change all later offset definitions if you add a field. You can
                   9722: fix this by declaring the offsets in the following way:
1.57      anton    9723: 
1.78      anton    9724: @example
                   9725: 0 constant a-offset
                   9726: a-offset float+ constant b-offset
                   9727: b-offset cell+ faligned constant c-offset
                   9728: @end example
1.57      anton    9729: 
1.78      anton    9730: Since we always use the offsets with @code{+}, we could use a defining
                   9731: word @code{cfield} that includes the @code{+} in the action of the
                   9732: defined word:
1.64      pazsan   9733: 
1.78      anton    9734: @example
                   9735: : cfield ( n "name" -- )
                   9736:     create ,
                   9737: does> ( name execution: addr1 -- addr2 )
                   9738:     @@ + ;
1.64      pazsan   9739: 
1.78      anton    9740: 0 cfield a
                   9741: 0 a float+ cfield b
                   9742: 0 b cell+ faligned cfield c
                   9743: @end example
1.64      pazsan   9744: 
1.78      anton    9745: Instead of @code{x-offset +}, we now simply write @code{x}.
1.64      pazsan   9746: 
1.78      anton    9747: The structure field words now can be used quite nicely. However,
                   9748: their definition is still a bit cumbersome: We have to repeat the
                   9749: name, the information about size and alignment is distributed before
                   9750: and after the field definitions etc.  The structure package presented
                   9751: here addresses these problems.
1.64      pazsan   9752: 
1.78      anton    9753: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
                   9754: @subsection Structure Usage
                   9755: @cindex structure usage
1.57      anton    9756: 
1.78      anton    9757: @cindex @code{field} usage
                   9758: @cindex @code{struct} usage
                   9759: @cindex @code{end-struct} usage
                   9760: You can define a structure for a (data-less) linked list with:
1.57      anton    9761: @example
1.78      anton    9762: struct
                   9763:     cell% field list-next
                   9764: end-struct list%
1.57      anton    9765: @end example
                   9766: 
1.78      anton    9767: With the address of the list node on the stack, you can compute the
                   9768: address of the field that contains the address of the next node with
                   9769: @code{list-next}. E.g., you can determine the length of a list
                   9770: with:
1.57      anton    9771: 
                   9772: @example
1.78      anton    9773: : list-length ( list -- n )
                   9774: \ "list" is a pointer to the first element of a linked list
                   9775: \ "n" is the length of the list
                   9776:     0 BEGIN ( list1 n1 )
                   9777:         over
                   9778:     WHILE ( list1 n1 )
                   9779:         1+ swap list-next @@ swap
                   9780:     REPEAT
                   9781:     nip ;
1.57      anton    9782: @end example
                   9783: 
1.78      anton    9784: You can reserve memory for a list node in the dictionary with
                   9785: @code{list% %allot}, which leaves the address of the list node on the
                   9786: stack. For the equivalent allocation on the heap you can use @code{list%
                   9787: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
                   9788: use @code{list% %allocate}). You can get the the size of a list
                   9789: node with @code{list% %size} and its alignment with @code{list%
                   9790: %alignment}.
                   9791: 
                   9792: Note that in ANS Forth the body of a @code{create}d word is
                   9793: @code{aligned} but not necessarily @code{faligned};
                   9794: therefore, if you do a:
1.57      anton    9795: 
                   9796: @example
1.78      anton    9797: create @emph{name} foo% %allot drop
1.57      anton    9798: @end example
                   9799: 
1.78      anton    9800: @noindent
                   9801: then the memory alloted for @code{foo%} is guaranteed to start at the
                   9802: body of @code{@emph{name}} only if @code{foo%} contains only character,
                   9803: cell and double fields.  Therefore, if your structure contains floats,
                   9804: better use
1.57      anton    9805: 
                   9806: @example
1.78      anton    9807: foo% %allot constant @emph{name}
1.57      anton    9808: @end example
                   9809: 
1.78      anton    9810: @cindex structures containing structures
                   9811: You can include a structure @code{foo%} as a field of
                   9812: another structure, like this:
1.65      anton    9813: @example
1.78      anton    9814: struct
                   9815: ...
                   9816:     foo% field ...
                   9817: ...
                   9818: end-struct ...
1.65      anton    9819: @end example
1.52      anton    9820: 
1.78      anton    9821: @cindex structure extension
                   9822: @cindex extended records
                   9823: Instead of starting with an empty structure, you can extend an
                   9824: existing structure. E.g., a plain linked list without data, as defined
                   9825: above, is hardly useful; You can extend it to a linked list of integers,
                   9826: like this:@footnote{This feature is also known as @emph{extended
                   9827: records}. It is the main innovation in the Oberon language; in other
                   9828: words, adding this feature to Modula-2 led Wirth to create a new
                   9829: language, write a new compiler etc.  Adding this feature to Forth just
                   9830: required a few lines of code.}
1.52      anton    9831: 
1.78      anton    9832: @example
                   9833: list%
                   9834:     cell% field intlist-int
                   9835: end-struct intlist%
                   9836: @end example
1.55      anton    9837: 
1.78      anton    9838: @code{intlist%} is a structure with two fields:
                   9839: @code{list-next} and @code{intlist-int}.
1.55      anton    9840: 
1.78      anton    9841: @cindex structures containing arrays
                   9842: You can specify an array type containing @emph{n} elements of
                   9843: type @code{foo%} like this:
1.55      anton    9844: 
                   9845: @example
1.78      anton    9846: foo% @emph{n} *
1.56      anton    9847: @end example
1.55      anton    9848: 
1.78      anton    9849: You can use this array type in any place where you can use a normal
                   9850: type, e.g., when defining a @code{field}, or with
                   9851: @code{%allot}.
                   9852: 
                   9853: @cindex first field optimization
                   9854: The first field is at the base address of a structure and the word for
                   9855: this field (e.g., @code{list-next}) actually does not change the address
                   9856: on the stack. You may be tempted to leave it away in the interest of
                   9857: run-time and space efficiency. This is not necessary, because the
                   9858: structure package optimizes this case: If you compile a first-field
                   9859: words, no code is generated. So, in the interest of readability and
                   9860: maintainability you should include the word for the field when accessing
                   9861: the field.
1.52      anton    9862: 
                   9863: 
1.78      anton    9864: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
                   9865: @subsection Structure Naming Convention
                   9866: @cindex structure naming convention
1.52      anton    9867: 
1.78      anton    9868: The field names that come to (my) mind are often quite generic, and,
                   9869: if used, would cause frequent name clashes. E.g., many structures
                   9870: probably contain a @code{counter} field. The structure names
                   9871: that come to (my) mind are often also the logical choice for the names
                   9872: of words that create such a structure.
1.52      anton    9873: 
1.78      anton    9874: Therefore, I have adopted the following naming conventions: 
1.52      anton    9875: 
1.78      anton    9876: @itemize @bullet
                   9877: @cindex field naming convention
                   9878: @item
                   9879: The names of fields are of the form
                   9880: @code{@emph{struct}-@emph{field}}, where
                   9881: @code{@emph{struct}} is the basic name of the structure, and
                   9882: @code{@emph{field}} is the basic name of the field. You can
                   9883: think of field words as converting the (address of the)
                   9884: structure into the (address of the) field.
1.52      anton    9885: 
1.78      anton    9886: @cindex structure naming convention
                   9887: @item
                   9888: The names of structures are of the form
                   9889: @code{@emph{struct}%}, where
                   9890: @code{@emph{struct}} is the basic name of the structure.
                   9891: @end itemize
1.52      anton    9892: 
1.78      anton    9893: This naming convention does not work that well for fields of extended
                   9894: structures; e.g., the integer list structure has a field
                   9895: @code{intlist-int}, but has @code{list-next}, not
                   9896: @code{intlist-next}.
1.53      anton    9897: 
1.78      anton    9898: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
                   9899: @subsection Structure Implementation
                   9900: @cindex structure implementation
                   9901: @cindex implementation of structures
1.52      anton    9902: 
1.78      anton    9903: The central idea in the implementation is to pass the data about the
                   9904: structure being built on the stack, not in some global
                   9905: variable. Everything else falls into place naturally once this design
                   9906: decision is made.
1.53      anton    9907: 
1.78      anton    9908: The type description on the stack is of the form @emph{align
                   9909: size}. Keeping the size on the top-of-stack makes dealing with arrays
                   9910: very simple.
1.53      anton    9911: 
1.78      anton    9912: @code{field} is a defining word that uses @code{Create}
                   9913: and @code{DOES>}. The body of the field contains the offset
                   9914: of the field, and the normal @code{DOES>} action is simply:
1.53      anton    9915: 
                   9916: @example
1.78      anton    9917: @@ +
1.53      anton    9918: @end example
                   9919: 
1.78      anton    9920: @noindent
                   9921: i.e., add the offset to the address, giving the stack effect
                   9922: @i{addr1 -- addr2} for a field.
                   9923: 
                   9924: @cindex first field optimization, implementation
                   9925: This simple structure is slightly complicated by the optimization
                   9926: for fields with offset 0, which requires a different
                   9927: @code{DOES>}-part (because we cannot rely on there being
                   9928: something on the stack if such a field is invoked during
                   9929: compilation). Therefore, we put the different @code{DOES>}-parts
                   9930: in separate words, and decide which one to invoke based on the
                   9931: offset. For a zero offset, the field is basically a noop; it is
                   9932: immediate, and therefore no code is generated when it is compiled.
1.53      anton    9933: 
1.78      anton    9934: @node Structure Glossary,  , Structure Implementation, Structures
                   9935: @subsection Structure Glossary
                   9936: @cindex structure glossary
1.53      anton    9937: 
1.5       anton    9938: 
1.78      anton    9939: doc-%align
                   9940: doc-%alignment
                   9941: doc-%alloc
                   9942: doc-%allocate
                   9943: doc-%allot
                   9944: doc-cell%
                   9945: doc-char%
                   9946: doc-dfloat%
                   9947: doc-double%
                   9948: doc-end-struct
                   9949: doc-field
                   9950: doc-float%
                   9951: doc-naligned
                   9952: doc-sfloat%
                   9953: doc-%size
                   9954: doc-struct
1.54      anton    9955: 
                   9956: 
1.26      crook    9957: @c -------------------------------------------------------------
1.78      anton    9958: @node Object-oriented Forth, Programming Tools, Structures, Words
                   9959: @section Object-oriented Forth
                   9960: 
                   9961: Gforth comes with three packages for object-oriented programming:
                   9962: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
                   9963: is preloaded, so you have to @code{include} them before use. The most
                   9964: important differences between these packages (and others) are discussed
                   9965: in @ref{Comparison with other object models}. All packages are written
                   9966: in ANS Forth and can be used with any other ANS Forth.
1.5       anton    9967: 
1.78      anton    9968: @menu
                   9969: * Why object-oriented programming?::  
                   9970: * Object-Oriented Terminology::  
                   9971: * Objects::                     
                   9972: * OOF::                         
                   9973: * Mini-OOF::                    
                   9974: * Comparison with other object models::  
                   9975: @end menu
1.5       anton    9976: 
1.78      anton    9977: @c ----------------------------------------------------------------
                   9978: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
                   9979: @subsection Why object-oriented programming?
                   9980: @cindex object-oriented programming motivation
                   9981: @cindex motivation for object-oriented programming
1.44      crook    9982: 
1.78      anton    9983: Often we have to deal with several data structures (@emph{objects}),
                   9984: that have to be treated similarly in some respects, but differently in
                   9985: others. Graphical objects are the textbook example: circles, triangles,
                   9986: dinosaurs, icons, and others, and we may want to add more during program
                   9987: development. We want to apply some operations to any graphical object,
                   9988: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
                   9989: has to do something different for every kind of object.
                   9990: @comment TODO add some other operations eg perimeter, area
                   9991: @comment and tie in to concrete examples later..
1.5       anton    9992: 
1.78      anton    9993: We could implement @code{draw} as a big @code{CASE}
                   9994: control structure that executes the appropriate code depending on the
                   9995: kind of object to be drawn. This would be not be very elegant, and,
                   9996: moreover, we would have to change @code{draw} every time we add
                   9997: a new kind of graphical object (say, a spaceship).
1.44      crook    9998: 
1.78      anton    9999: What we would rather do is: When defining spaceships, we would tell
                   10000: the system: ``Here's how you @code{draw} a spaceship; you figure
                   10001: out the rest''.
1.5       anton    10002: 
1.78      anton    10003: This is the problem that all systems solve that (rightfully) call
                   10004: themselves object-oriented; the object-oriented packages presented here
                   10005: solve this problem (and not much else).
                   10006: @comment TODO ?list properties of oo systems.. oo vs o-based?
1.44      crook    10007: 
1.78      anton    10008: @c ------------------------------------------------------------------------
                   10009: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
                   10010: @subsection Object-Oriented Terminology
                   10011: @cindex object-oriented terminology
                   10012: @cindex terminology for object-oriented programming
1.5       anton    10013: 
1.78      anton    10014: This section is mainly for reference, so you don't have to understand
                   10015: all of it right away.  The terminology is mainly Smalltalk-inspired.  In
                   10016: short:
1.44      crook    10017: 
1.78      anton    10018: @table @emph
                   10019: @cindex class
                   10020: @item class
                   10021: a data structure definition with some extras.
1.5       anton    10022: 
1.78      anton    10023: @cindex object
                   10024: @item object
                   10025: an instance of the data structure described by the class definition.
1.5       anton    10026: 
1.78      anton    10027: @cindex instance variables
                   10028: @item instance variables
                   10029: fields of the data structure.
1.5       anton    10030: 
1.78      anton    10031: @cindex selector
                   10032: @cindex method selector
                   10033: @cindex virtual function
                   10034: @item selector
                   10035: (or @emph{method selector}) a word (e.g.,
                   10036: @code{draw}) that performs an operation on a variety of data
                   10037: structures (classes). A selector describes @emph{what} operation to
                   10038: perform. In C++ terminology: a (pure) virtual function.
1.5       anton    10039: 
1.78      anton    10040: @cindex method
                   10041: @item method
                   10042: the concrete definition that performs the operation
                   10043: described by the selector for a specific class. A method specifies
                   10044: @emph{how} the operation is performed for a specific class.
1.5       anton    10045: 
1.78      anton    10046: @cindex selector invocation
                   10047: @cindex message send
                   10048: @cindex invoking a selector
                   10049: @item selector invocation
                   10050: a call of a selector. One argument of the call (the TOS (top-of-stack))
                   10051: is used for determining which method is used. In Smalltalk terminology:
                   10052: a message (consisting of the selector and the other arguments) is sent
                   10053: to the object.
1.5       anton    10054: 
1.78      anton    10055: @cindex receiving object
                   10056: @item receiving object
                   10057: the object used for determining the method executed by a selector
                   10058: invocation. In the @file{objects.fs} model, it is the object that is on
                   10059: the TOS when the selector is invoked. (@emph{Receiving} comes from
                   10060: the Smalltalk @emph{message} terminology.)
1.5       anton    10061: 
1.78      anton    10062: @cindex child class
                   10063: @cindex parent class
                   10064: @cindex inheritance
                   10065: @item child class
                   10066: a class that has (@emph{inherits}) all properties (instance variables,
                   10067: selectors, methods) from a @emph{parent class}. In Smalltalk
                   10068: terminology: The subclass inherits from the superclass. In C++
                   10069: terminology: The derived class inherits from the base class.
1.5       anton    10070: 
1.78      anton    10071: @end table
1.5       anton    10072: 
1.78      anton    10073: @c If you wonder about the message sending terminology, it comes from
                   10074: @c a time when each object had it's own task and objects communicated via
                   10075: @c message passing; eventually the Smalltalk developers realized that
                   10076: @c they can do most things through simple (indirect) calls. They kept the
                   10077: @c terminology.
1.5       anton    10078: 
1.78      anton    10079: @c --------------------------------------------------------------
                   10080: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
                   10081: @subsection The @file{objects.fs} model
                   10082: @cindex objects
                   10083: @cindex object-oriented programming
1.26      crook    10084: 
1.78      anton    10085: @cindex @file{objects.fs}
                   10086: @cindex @file{oof.fs}
1.26      crook    10087: 
1.78      anton    10088: This section describes the @file{objects.fs} package. This material also
                   10089: has been published in M. Anton Ertl,
                   10090: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
                   10091: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
                   10092: 37--43.
                   10093: @c McKewan's and Zsoter's packages
1.26      crook    10094: 
1.78      anton    10095: This section assumes that you have read @ref{Structures}.
1.5       anton    10096: 
1.78      anton    10097: The techniques on which this model is based have been used to implement
                   10098: the parser generator, Gray, and have also been used in Gforth for
                   10099: implementing the various flavours of word lists (hashed or not,
                   10100: case-sensitive or not, special-purpose word lists for locals etc.).
1.5       anton    10101: 
                   10102: 
1.26      crook    10103: @menu
1.78      anton    10104: * Properties of the Objects model::  
                   10105: * Basic Objects Usage::         
                   10106: * The Objects base class::      
                   10107: * Creating objects::            
                   10108: * Object-Oriented Programming Style::  
                   10109: * Class Binding::               
                   10110: * Method conveniences::         
                   10111: * Classes and Scoping::         
                   10112: * Dividing classes::            
                   10113: * Object Interfaces::           
                   10114: * Objects Implementation::      
                   10115: * Objects Glossary::            
1.26      crook    10116: @end menu
1.5       anton    10117: 
1.78      anton    10118: Marcel Hendrix provided helpful comments on this section.
1.5       anton    10119: 
1.78      anton    10120: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
                   10121: @subsubsection Properties of the @file{objects.fs} model
                   10122: @cindex @file{objects.fs} properties
1.5       anton    10123: 
1.78      anton    10124: @itemize @bullet
                   10125: @item
                   10126: It is straightforward to pass objects on the stack. Passing
                   10127: selectors on the stack is a little less convenient, but possible.
1.44      crook    10128: 
1.78      anton    10129: @item
                   10130: Objects are just data structures in memory, and are referenced by their
                   10131: address. You can create words for objects with normal defining words
                   10132: like @code{constant}. Likewise, there is no difference between instance
                   10133: variables that contain objects and those that contain other data.
1.5       anton    10134: 
1.78      anton    10135: @item
                   10136: Late binding is efficient and easy to use.
1.44      crook    10137: 
1.78      anton    10138: @item
                   10139: It avoids parsing, and thus avoids problems with state-smartness
                   10140: and reduced extensibility; for convenience there are a few parsing
                   10141: words, but they have non-parsing counterparts. There are also a few
                   10142: defining words that parse. This is hard to avoid, because all standard
                   10143: defining words parse (except @code{:noname}); however, such
                   10144: words are not as bad as many other parsing words, because they are not
                   10145: state-smart.
1.5       anton    10146: 
1.78      anton    10147: @item
                   10148: It does not try to incorporate everything. It does a few things and does
                   10149: them well (IMO). In particular, this model was not designed to support
                   10150: information hiding (although it has features that may help); you can use
                   10151: a separate package for achieving this.
1.5       anton    10152: 
1.78      anton    10153: @item
                   10154: It is layered; you don't have to learn and use all features to use this
                   10155: model. Only a few features are necessary (@pxref{Basic Objects Usage},
                   10156: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
                   10157: are optional and independent of each other.
1.5       anton    10158: 
1.78      anton    10159: @item
                   10160: An implementation in ANS Forth is available.
1.5       anton    10161: 
1.78      anton    10162: @end itemize
1.5       anton    10163: 
1.44      crook    10164: 
1.78      anton    10165: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
                   10166: @subsubsection Basic @file{objects.fs} Usage
                   10167: @cindex basic objects usage
                   10168: @cindex objects, basic usage
1.5       anton    10169: 
1.78      anton    10170: You can define a class for graphical objects like this:
1.44      crook    10171: 
1.78      anton    10172: @cindex @code{class} usage
                   10173: @cindex @code{end-class} usage
                   10174: @cindex @code{selector} usage
1.5       anton    10175: @example
1.78      anton    10176: object class \ "object" is the parent class
                   10177:   selector draw ( x y graphical -- )
                   10178: end-class graphical
                   10179: @end example
                   10180: 
                   10181: This code defines a class @code{graphical} with an
                   10182: operation @code{draw}.  We can perform the operation
                   10183: @code{draw} on any @code{graphical} object, e.g.:
                   10184: 
                   10185: @example
                   10186: 100 100 t-rex draw
1.26      crook    10187: @end example
1.5       anton    10188: 
1.78      anton    10189: @noindent
                   10190: where @code{t-rex} is a word (say, a constant) that produces a
                   10191: graphical object.
                   10192: 
                   10193: @comment TODO add a 2nd operation eg perimeter.. and use for
                   10194: @comment a concrete example
1.5       anton    10195: 
1.78      anton    10196: @cindex abstract class
                   10197: How do we create a graphical object? With the present definitions,
                   10198: we cannot create a useful graphical object. The class
                   10199: @code{graphical} describes graphical objects in general, but not
                   10200: any concrete graphical object type (C++ users would call it an
                   10201: @emph{abstract class}); e.g., there is no method for the selector
                   10202: @code{draw} in the class @code{graphical}.
1.5       anton    10203: 
1.78      anton    10204: For concrete graphical objects, we define child classes of the
                   10205: class @code{graphical}, e.g.:
1.5       anton    10206: 
1.78      anton    10207: @cindex @code{overrides} usage
                   10208: @cindex @code{field} usage in class definition
1.26      crook    10209: @example
1.78      anton    10210: graphical class \ "graphical" is the parent class
                   10211:   cell% field circle-radius
1.5       anton    10212: 
1.78      anton    10213: :noname ( x y circle -- )
                   10214:   circle-radius @@ draw-circle ;
                   10215: overrides draw
1.5       anton    10216: 
1.78      anton    10217: :noname ( n-radius circle -- )
                   10218:   circle-radius ! ;
                   10219: overrides construct
1.5       anton    10220: 
1.78      anton    10221: end-class circle
                   10222: @end example
1.44      crook    10223: 
1.78      anton    10224: Here we define a class @code{circle} as a child of @code{graphical},
                   10225: with field @code{circle-radius} (which behaves just like a field
                   10226: (@pxref{Structures}); it defines (using @code{overrides}) new methods
                   10227: for the selectors @code{draw} and @code{construct} (@code{construct} is
                   10228: defined in @code{object}, the parent class of @code{graphical}).
1.5       anton    10229: 
1.78      anton    10230: Now we can create a circle on the heap (i.e.,
                   10231: @code{allocate}d memory) with:
1.44      crook    10232: 
1.78      anton    10233: @cindex @code{heap-new} usage
1.5       anton    10234: @example
1.78      anton    10235: 50 circle heap-new constant my-circle
1.5       anton    10236: @end example
                   10237: 
1.78      anton    10238: @noindent
                   10239: @code{heap-new} invokes @code{construct}, thus
                   10240: initializing the field @code{circle-radius} with 50. We can draw
                   10241: this new circle at (100,100) with:
1.5       anton    10242: 
                   10243: @example
1.78      anton    10244: 100 100 my-circle draw
1.5       anton    10245: @end example
                   10246: 
1.78      anton    10247: @cindex selector invocation, restrictions
                   10248: @cindex class definition, restrictions
                   10249: Note: You can only invoke a selector if the object on the TOS
                   10250: (the receiving object) belongs to the class where the selector was
                   10251: defined or one of its descendents; e.g., you can invoke
                   10252: @code{draw} only for objects belonging to @code{graphical}
                   10253: or its descendents (e.g., @code{circle}).  Immediately before
                   10254: @code{end-class}, the search order has to be the same as
                   10255: immediately after @code{class}.
                   10256: 
                   10257: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
                   10258: @subsubsection The @file{object.fs} base class
                   10259: @cindex @code{object} class
                   10260: 
                   10261: When you define a class, you have to specify a parent class.  So how do
                   10262: you start defining classes? There is one class available from the start:
                   10263: @code{object}. It is ancestor for all classes and so is the
                   10264: only class that has no parent. It has two selectors: @code{construct}
                   10265: and @code{print}.
                   10266: 
                   10267: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
                   10268: @subsubsection Creating objects
                   10269: @cindex creating objects
                   10270: @cindex object creation
                   10271: @cindex object allocation options
                   10272: 
                   10273: @cindex @code{heap-new} discussion
                   10274: @cindex @code{dict-new} discussion
                   10275: @cindex @code{construct} discussion
                   10276: You can create and initialize an object of a class on the heap with
                   10277: @code{heap-new} ( ... class -- object ) and in the dictionary
                   10278: (allocation with @code{allot}) with @code{dict-new} (
                   10279: ... class -- object ). Both words invoke @code{construct}, which
                   10280: consumes the stack items indicated by "..." above.
                   10281: 
                   10282: @cindex @code{init-object} discussion
                   10283: @cindex @code{class-inst-size} discussion
                   10284: If you want to allocate memory for an object yourself, you can get its
                   10285: alignment and size with @code{class-inst-size 2@@} ( class --
                   10286: align size ). Once you have memory for an object, you can initialize
                   10287: it with @code{init-object} ( ... class object -- );
                   10288: @code{construct} does only a part of the necessary work.
                   10289: 
                   10290: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
                   10291: @subsubsection Object-Oriented Programming Style
                   10292: @cindex object-oriented programming style
                   10293: @cindex programming style, object-oriented
1.5       anton    10294: 
1.78      anton    10295: This section is not exhaustive.
1.5       anton    10296: 
1.78      anton    10297: @cindex stack effects of selectors
                   10298: @cindex selectors and stack effects
                   10299: In general, it is a good idea to ensure that all methods for the
                   10300: same selector have the same stack effect: when you invoke a selector,
                   10301: you often have no idea which method will be invoked, so, unless all
                   10302: methods have the same stack effect, you will not know the stack effect
                   10303: of the selector invocation.
1.5       anton    10304: 
1.78      anton    10305: One exception to this rule is methods for the selector
                   10306: @code{construct}. We know which method is invoked, because we
                   10307: specify the class to be constructed at the same place. Actually, I
                   10308: defined @code{construct} as a selector only to give the users a
                   10309: convenient way to specify initialization. The way it is used, a
                   10310: mechanism different from selector invocation would be more natural
                   10311: (but probably would take more code and more space to explain).
1.5       anton    10312: 
1.78      anton    10313: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
                   10314: @subsubsection Class Binding
                   10315: @cindex class binding
                   10316: @cindex early binding
1.5       anton    10317: 
1.78      anton    10318: @cindex late binding
                   10319: Normal selector invocations determine the method at run-time depending
                   10320: on the class of the receiving object. This run-time selection is called
                   10321: @i{late binding}.
1.5       anton    10322: 
1.78      anton    10323: Sometimes it's preferable to invoke a different method. For example,
                   10324: you might want to use the simple method for @code{print}ing
                   10325: @code{object}s instead of the possibly long-winded @code{print} method
                   10326: of the receiver class. You can achieve this by replacing the invocation
                   10327: of @code{print} with:
1.5       anton    10328: 
1.78      anton    10329: @cindex @code{[bind]} usage
1.5       anton    10330: @example
1.78      anton    10331: [bind] object print
1.5       anton    10332: @end example
                   10333: 
1.78      anton    10334: @noindent
                   10335: in compiled code or:
                   10336: 
                   10337: @cindex @code{bind} usage
1.5       anton    10338: @example
1.78      anton    10339: bind object print
1.5       anton    10340: @end example
                   10341: 
1.78      anton    10342: @cindex class binding, alternative to
                   10343: @noindent
                   10344: in interpreted code. Alternatively, you can define the method with a
                   10345: name (e.g., @code{print-object}), and then invoke it through the
                   10346: name. Class binding is just a (often more convenient) way to achieve
                   10347: the same effect; it avoids name clutter and allows you to invoke
                   10348: methods directly without naming them first.
1.5       anton    10349: 
1.78      anton    10350: @cindex superclass binding
                   10351: @cindex parent class binding
                   10352: A frequent use of class binding is this: When we define a method
                   10353: for a selector, we often want the method to do what the selector does
                   10354: in the parent class, and a little more. There is a special word for
                   10355: this purpose: @code{[parent]}; @code{[parent]
                   10356: @emph{selector}} is equivalent to @code{[bind] @emph{parent
                   10357: selector}}, where @code{@emph{parent}} is the parent
                   10358: class of the current class. E.g., a method definition might look like:
1.44      crook    10359: 
1.78      anton    10360: @cindex @code{[parent]} usage
                   10361: @example
                   10362: :noname
                   10363:   dup [parent] foo \ do parent's foo on the receiving object
                   10364:   ... \ do some more
                   10365: ; overrides foo
                   10366: @end example
1.6       pazsan   10367: 
1.78      anton    10368: @cindex class binding as optimization
                   10369: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
                   10370: March 1997), Andrew McKewan presents class binding as an optimization
                   10371: technique. I recommend not using it for this purpose unless you are in
                   10372: an emergency. Late binding is pretty fast with this model anyway, so the
                   10373: benefit of using class binding is small; the cost of using class binding
                   10374: where it is not appropriate is reduced maintainability.
1.44      crook    10375: 
1.78      anton    10376: While we are at programming style questions: You should bind
                   10377: selectors only to ancestor classes of the receiving object. E.g., say,
                   10378: you know that the receiving object is of class @code{foo} or its
                   10379: descendents; then you should bind only to @code{foo} and its
                   10380: ancestors.
1.12      anton    10381: 
1.78      anton    10382: @node Method conveniences, Classes and Scoping, Class Binding, Objects
                   10383: @subsubsection Method conveniences
                   10384: @cindex method conveniences
1.44      crook    10385: 
1.78      anton    10386: In a method you usually access the receiving object pretty often.  If
                   10387: you define the method as a plain colon definition (e.g., with
                   10388: @code{:noname}), you may have to do a lot of stack
                   10389: gymnastics. To avoid this, you can define the method with @code{m:
                   10390: ... ;m}. E.g., you could define the method for
                   10391: @code{draw}ing a @code{circle} with
1.6       pazsan   10392: 
1.78      anton    10393: @cindex @code{this} usage
                   10394: @cindex @code{m:} usage
                   10395: @cindex @code{;m} usage
                   10396: @example
                   10397: m: ( x y circle -- )
                   10398:   ( x y ) this circle-radius @@ draw-circle ;m
                   10399: @end example
1.6       pazsan   10400: 
1.78      anton    10401: @cindex @code{exit} in @code{m: ... ;m}
                   10402: @cindex @code{exitm} discussion
                   10403: @cindex @code{catch} in @code{m: ... ;m}
                   10404: When this method is executed, the receiver object is removed from the
                   10405: stack; you can access it with @code{this} (admittedly, in this
                   10406: example the use of @code{m: ... ;m} offers no advantage). Note
                   10407: that I specify the stack effect for the whole method (i.e. including
                   10408: the receiver object), not just for the code between @code{m:}
                   10409: and @code{;m}. You cannot use @code{exit} in
                   10410: @code{m:...;m}; instead, use
                   10411: @code{exitm}.@footnote{Moreover, for any word that calls
                   10412: @code{catch} and was defined before loading
                   10413: @code{objects.fs}, you have to redefine it like I redefined
                   10414: @code{catch}: @code{: catch this >r catch r> to-this ;}}
1.12      anton    10415: 
1.78      anton    10416: @cindex @code{inst-var} usage
                   10417: You will frequently use sequences of the form @code{this
                   10418: @emph{field}} (in the example above: @code{this
                   10419: circle-radius}). If you use the field only in this way, you can
                   10420: define it with @code{inst-var} and eliminate the
                   10421: @code{this} before the field name. E.g., the @code{circle}
                   10422: class above could also be defined with:
1.6       pazsan   10423: 
1.78      anton    10424: @example
                   10425: graphical class
                   10426:   cell% inst-var radius
1.6       pazsan   10427: 
1.78      anton    10428: m: ( x y circle -- )
                   10429:   radius @@ draw-circle ;m
                   10430: overrides draw
1.6       pazsan   10431: 
1.78      anton    10432: m: ( n-radius circle -- )
                   10433:   radius ! ;m
                   10434: overrides construct
1.6       pazsan   10435: 
1.78      anton    10436: end-class circle
                   10437: @end example
1.6       pazsan   10438: 
1.78      anton    10439: @code{radius} can only be used in @code{circle} and its
                   10440: descendent classes and inside @code{m:...;m}.
1.6       pazsan   10441: 
1.78      anton    10442: @cindex @code{inst-value} usage
                   10443: You can also define fields with @code{inst-value}, which is
                   10444: to @code{inst-var} what @code{value} is to
                   10445: @code{variable}.  You can change the value of such a field with
                   10446: @code{[to-inst]}.  E.g., we could also define the class
                   10447: @code{circle} like this:
1.44      crook    10448: 
1.78      anton    10449: @example
                   10450: graphical class
                   10451:   inst-value radius
1.6       pazsan   10452: 
1.78      anton    10453: m: ( x y circle -- )
                   10454:   radius draw-circle ;m
                   10455: overrides draw
1.44      crook    10456: 
1.78      anton    10457: m: ( n-radius circle -- )
                   10458:   [to-inst] radius ;m
                   10459: overrides construct
1.6       pazsan   10460: 
1.78      anton    10461: end-class circle
                   10462: @end example
1.6       pazsan   10463: 
1.78      anton    10464: @c !! :m is easy to confuse with m:.  Another name would be better.
1.6       pazsan   10465: 
1.78      anton    10466: @c Finally, you can define named methods with @code{:m}.  One use of this
                   10467: @c feature is the definition of words that occur only in one class and are
                   10468: @c not intended to be overridden, but which still need method context
                   10469: @c (e.g., for accessing @code{inst-var}s).  Another use is for methods that
                   10470: @c would be bound frequently, if defined anonymously.
1.6       pazsan   10471: 
                   10472: 
1.78      anton    10473: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
                   10474: @subsubsection Classes and Scoping
                   10475: @cindex classes and scoping
                   10476: @cindex scoping and classes
1.6       pazsan   10477: 
1.78      anton    10478: Inheritance is frequent, unlike structure extension. This exacerbates
                   10479: the problem with the field name convention (@pxref{Structure Naming
                   10480: Convention}): One always has to remember in which class the field was
                   10481: originally defined; changing a part of the class structure would require
                   10482: changes for renaming in otherwise unaffected code.
1.6       pazsan   10483: 
1.78      anton    10484: @cindex @code{inst-var} visibility
                   10485: @cindex @code{inst-value} visibility
                   10486: To solve this problem, I added a scoping mechanism (which was not in my
                   10487: original charter): A field defined with @code{inst-var} (or
                   10488: @code{inst-value}) is visible only in the class where it is defined and in
                   10489: the descendent classes of this class.  Using such fields only makes
                   10490: sense in @code{m:}-defined methods in these classes anyway.
1.6       pazsan   10491: 
1.78      anton    10492: This scoping mechanism allows us to use the unadorned field name,
                   10493: because name clashes with unrelated words become much less likely.
1.6       pazsan   10494: 
1.78      anton    10495: @cindex @code{protected} discussion
                   10496: @cindex @code{private} discussion
                   10497: Once we have this mechanism, we can also use it for controlling the
                   10498: visibility of other words: All words defined after
                   10499: @code{protected} are visible only in the current class and its
                   10500: descendents. @code{public} restores the compilation
                   10501: (i.e. @code{current}) word list that was in effect before. If you
                   10502: have several @code{protected}s without an intervening
                   10503: @code{public} or @code{set-current}, @code{public}
                   10504: will restore the compilation word list in effect before the first of
                   10505: these @code{protected}s.
1.6       pazsan   10506: 
1.78      anton    10507: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
                   10508: @subsubsection Dividing classes
                   10509: @cindex Dividing classes
                   10510: @cindex @code{methods}...@code{end-methods}
1.6       pazsan   10511: 
1.78      anton    10512: You may want to do the definition of methods separate from the
                   10513: definition of the class, its selectors, fields, and instance variables,
                   10514: i.e., separate the implementation from the definition.  You can do this
                   10515: in the following way:
1.6       pazsan   10516: 
1.78      anton    10517: @example
                   10518: graphical class
                   10519:   inst-value radius
                   10520: end-class circle
1.6       pazsan   10521: 
1.78      anton    10522: ... \ do some other stuff
1.6       pazsan   10523: 
1.78      anton    10524: circle methods \ now we are ready
1.44      crook    10525: 
1.78      anton    10526: m: ( x y circle -- )
                   10527:   radius draw-circle ;m
                   10528: overrides draw
1.6       pazsan   10529: 
1.78      anton    10530: m: ( n-radius circle -- )
                   10531:   [to-inst] radius ;m
                   10532: overrides construct
1.44      crook    10533: 
1.78      anton    10534: end-methods
                   10535: @end example
1.7       pazsan   10536: 
1.78      anton    10537: You can use several @code{methods}...@code{end-methods} sections.  The
                   10538: only things you can do to the class in these sections are: defining
                   10539: methods, and overriding the class's selectors.  You must not define new
                   10540: selectors or fields.
1.7       pazsan   10541: 
1.78      anton    10542: Note that you often have to override a selector before using it.  In
                   10543: particular, you usually have to override @code{construct} with a new
                   10544: method before you can invoke @code{heap-new} and friends.  E.g., you
                   10545: must not create a circle before the @code{overrides construct} sequence
                   10546: in the example above.
1.7       pazsan   10547: 
1.78      anton    10548: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
                   10549: @subsubsection Object Interfaces
                   10550: @cindex object interfaces
                   10551: @cindex interfaces for objects
1.7       pazsan   10552: 
1.78      anton    10553: In this model you can only call selectors defined in the class of the
                   10554: receiving objects or in one of its ancestors. If you call a selector
                   10555: with a receiving object that is not in one of these classes, the
                   10556: result is undefined; if you are lucky, the program crashes
                   10557: immediately.
1.7       pazsan   10558: 
1.78      anton    10559: @cindex selectors common to hardly-related classes
                   10560: Now consider the case when you want to have a selector (or several)
                   10561: available in two classes: You would have to add the selector to a
                   10562: common ancestor class, in the worst case to @code{object}. You
                   10563: may not want to do this, e.g., because someone else is responsible for
                   10564: this ancestor class.
1.7       pazsan   10565: 
1.78      anton    10566: The solution for this problem is interfaces. An interface is a
                   10567: collection of selectors. If a class implements an interface, the
                   10568: selectors become available to the class and its descendents. A class
                   10569: can implement an unlimited number of interfaces. For the problem
                   10570: discussed above, we would define an interface for the selector(s), and
                   10571: both classes would implement the interface.
1.7       pazsan   10572: 
1.78      anton    10573: As an example, consider an interface @code{storage} for
                   10574: writing objects to disk and getting them back, and a class
                   10575: @code{foo} that implements it. The code would look like this:
1.7       pazsan   10576: 
1.78      anton    10577: @cindex @code{interface} usage
                   10578: @cindex @code{end-interface} usage
                   10579: @cindex @code{implementation} usage
                   10580: @example
                   10581: interface
                   10582:   selector write ( file object -- )
                   10583:   selector read1 ( file object -- )
                   10584: end-interface storage
1.13      pazsan   10585: 
1.78      anton    10586: bar class
                   10587:   storage implementation
1.13      pazsan   10588: 
1.78      anton    10589: ... overrides write
                   10590: ... overrides read1
                   10591: ...
                   10592: end-class foo
                   10593: @end example
1.13      pazsan   10594: 
1.78      anton    10595: @noindent
                   10596: (I would add a word @code{read} @i{( file -- object )} that uses
                   10597: @code{read1} internally, but that's beyond the point illustrated
                   10598: here.)
1.13      pazsan   10599: 
1.78      anton    10600: Note that you cannot use @code{protected} in an interface; and
                   10601: of course you cannot define fields.
1.13      pazsan   10602: 
1.78      anton    10603: In the Neon model, all selectors are available for all classes;
                   10604: therefore it does not need interfaces. The price you pay in this model
                   10605: is slower late binding, and therefore, added complexity to avoid late
                   10606: binding.
1.13      pazsan   10607: 
1.78      anton    10608: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
                   10609: @subsubsection @file{objects.fs} Implementation
                   10610: @cindex @file{objects.fs} implementation
1.13      pazsan   10611: 
1.78      anton    10612: @cindex @code{object-map} discussion
                   10613: An object is a piece of memory, like one of the data structures
                   10614: described with @code{struct...end-struct}. It has a field
                   10615: @code{object-map} that points to the method map for the object's
                   10616: class.
1.13      pazsan   10617: 
1.78      anton    10618: @cindex method map
                   10619: @cindex virtual function table
                   10620: The @emph{method map}@footnote{This is Self terminology; in C++
                   10621: terminology: virtual function table.} is an array that contains the
                   10622: execution tokens (@i{xt}s) of the methods for the object's class. Each
                   10623: selector contains an offset into a method map.
1.13      pazsan   10624: 
1.78      anton    10625: @cindex @code{selector} implementation, class
                   10626: @code{selector} is a defining word that uses
                   10627: @code{CREATE} and @code{DOES>}. The body of the
                   10628: selector contains the offset; the @code{DOES>} action for a
                   10629: class selector is, basically:
1.8       pazsan   10630: 
                   10631: @example
1.78      anton    10632: ( object addr ) @@ over object-map @@ + @@ execute
1.13      pazsan   10633: @end example
                   10634: 
1.78      anton    10635: Since @code{object-map} is the first field of the object, it
                   10636: does not generate any code. As you can see, calling a selector has a
                   10637: small, constant cost.
1.26      crook    10638: 
1.78      anton    10639: @cindex @code{current-interface} discussion
                   10640: @cindex class implementation and representation
                   10641: A class is basically a @code{struct} combined with a method
                   10642: map. During the class definition the alignment and size of the class
                   10643: are passed on the stack, just as with @code{struct}s, so
                   10644: @code{field} can also be used for defining class
                   10645: fields. However, passing more items on the stack would be
                   10646: inconvenient, so @code{class} builds a data structure in memory,
                   10647: which is accessed through the variable
                   10648: @code{current-interface}. After its definition is complete, the
                   10649: class is represented on the stack by a pointer (e.g., as parameter for
                   10650: a child class definition).
1.26      crook    10651: 
1.78      anton    10652: A new class starts off with the alignment and size of its parent,
                   10653: and a copy of the parent's method map. Defining new fields extends the
                   10654: size and alignment; likewise, defining new selectors extends the
                   10655: method map. @code{overrides} just stores a new @i{xt} in the method
                   10656: map at the offset given by the selector.
1.13      pazsan   10657: 
1.78      anton    10658: @cindex class binding, implementation
                   10659: Class binding just gets the @i{xt} at the offset given by the selector
                   10660: from the class's method map and @code{compile,}s (in the case of
                   10661: @code{[bind]}) it.
1.13      pazsan   10662: 
1.78      anton    10663: @cindex @code{this} implementation
                   10664: @cindex @code{catch} and @code{this}
                   10665: @cindex @code{this} and @code{catch}
                   10666: I implemented @code{this} as a @code{value}. At the
                   10667: start of an @code{m:...;m} method the old @code{this} is
                   10668: stored to the return stack and restored at the end; and the object on
                   10669: the TOS is stored @code{TO this}. This technique has one
                   10670: disadvantage: If the user does not leave the method via
                   10671: @code{;m}, but via @code{throw} or @code{exit},
                   10672: @code{this} is not restored (and @code{exit} may
                   10673: crash). To deal with the @code{throw} problem, I have redefined
                   10674: @code{catch} to save and restore @code{this}; the same
                   10675: should be done with any word that can catch an exception. As for
                   10676: @code{exit}, I simply forbid it (as a replacement, there is
                   10677: @code{exitm}).
1.13      pazsan   10678: 
1.78      anton    10679: @cindex @code{inst-var} implementation
                   10680: @code{inst-var} is just the same as @code{field}, with
                   10681: a different @code{DOES>} action:
1.13      pazsan   10682: @example
1.78      anton    10683: @@ this +
1.8       pazsan   10684: @end example
1.78      anton    10685: Similar for @code{inst-value}.
1.8       pazsan   10686: 
1.78      anton    10687: @cindex class scoping implementation
                   10688: Each class also has a word list that contains the words defined with
                   10689: @code{inst-var} and @code{inst-value}, and its protected
                   10690: words. It also has a pointer to its parent. @code{class} pushes
                   10691: the word lists of the class and all its ancestors onto the search order stack,
                   10692: and @code{end-class} drops them.
1.20      pazsan   10693: 
1.78      anton    10694: @cindex interface implementation
                   10695: An interface is like a class without fields, parent and protected
                   10696: words; i.e., it just has a method map. If a class implements an
                   10697: interface, its method map contains a pointer to the method map of the
                   10698: interface. The positive offsets in the map are reserved for class
                   10699: methods, therefore interface map pointers have negative
                   10700: offsets. Interfaces have offsets that are unique throughout the
                   10701: system, unlike class selectors, whose offsets are only unique for the
                   10702: classes where the selector is available (invokable).
1.20      pazsan   10703: 
1.78      anton    10704: This structure means that interface selectors have to perform one
                   10705: indirection more than class selectors to find their method. Their body
                   10706: contains the interface map pointer offset in the class method map, and
                   10707: the method offset in the interface method map. The
                   10708: @code{does>} action for an interface selector is, basically:
1.20      pazsan   10709: 
                   10710: @example
1.78      anton    10711: ( object selector-body )
                   10712: 2dup selector-interface @@ ( object selector-body object interface-offset )
                   10713: swap object-map @@ + @@ ( object selector-body map )
                   10714: swap selector-offset @@ + @@ execute
1.20      pazsan   10715: @end example
                   10716: 
1.78      anton    10717: where @code{object-map} and @code{selector-offset} are
                   10718: first fields and generate no code.
1.20      pazsan   10719: 
1.78      anton    10720: As a concrete example, consider the following code:
1.20      pazsan   10721: 
                   10722: @example
1.78      anton    10723: interface
                   10724:   selector if1sel1
                   10725:   selector if1sel2
                   10726: end-interface if1
1.20      pazsan   10727: 
1.78      anton    10728: object class
                   10729:   if1 implementation
                   10730:   selector cl1sel1
                   10731:   cell% inst-var cl1iv1
1.20      pazsan   10732: 
1.78      anton    10733: ' m1 overrides construct
                   10734: ' m2 overrides if1sel1
                   10735: ' m3 overrides if1sel2
                   10736: ' m4 overrides cl1sel2
                   10737: end-class cl1
1.20      pazsan   10738: 
1.78      anton    10739: create obj1 object dict-new drop
                   10740: create obj2 cl1    dict-new drop
                   10741: @end example
1.20      pazsan   10742: 
1.78      anton    10743: The data structure created by this code (including the data structure
                   10744: for @code{object}) is shown in the
                   10745: @uref{objects-implementation.eps,figure}, assuming a cell size of 4.
                   10746: @comment TODO add this diagram..
1.20      pazsan   10747: 
1.78      anton    10748: @node Objects Glossary,  , Objects Implementation, Objects
                   10749: @subsubsection @file{objects.fs} Glossary
                   10750: @cindex @file{objects.fs} Glossary
1.20      pazsan   10751: 
                   10752: 
1.78      anton    10753: doc---objects-bind
                   10754: doc---objects-<bind>
                   10755: doc---objects-bind'
                   10756: doc---objects-[bind]
                   10757: doc---objects-class
                   10758: doc---objects-class->map
                   10759: doc---objects-class-inst-size
                   10760: doc---objects-class-override!
1.79      anton    10761: doc---objects-class-previous
                   10762: doc---objects-class>order
1.78      anton    10763: doc---objects-construct
                   10764: doc---objects-current'
                   10765: doc---objects-[current]
                   10766: doc---objects-current-interface
                   10767: doc---objects-dict-new
                   10768: doc---objects-end-class
                   10769: doc---objects-end-class-noname
                   10770: doc---objects-end-interface
                   10771: doc---objects-end-interface-noname
                   10772: doc---objects-end-methods
                   10773: doc---objects-exitm
                   10774: doc---objects-heap-new
                   10775: doc---objects-implementation
                   10776: doc---objects-init-object
                   10777: doc---objects-inst-value
                   10778: doc---objects-inst-var
                   10779: doc---objects-interface
                   10780: doc---objects-m:
                   10781: doc---objects-:m
                   10782: doc---objects-;m
                   10783: doc---objects-method
                   10784: doc---objects-methods
                   10785: doc---objects-object
                   10786: doc---objects-overrides
                   10787: doc---objects-[parent]
                   10788: doc---objects-print
                   10789: doc---objects-protected
                   10790: doc---objects-public
                   10791: doc---objects-selector
                   10792: doc---objects-this
                   10793: doc---objects-<to-inst>
                   10794: doc---objects-[to-inst]
                   10795: doc---objects-to-this
                   10796: doc---objects-xt-new
1.20      pazsan   10797: 
                   10798: 
1.78      anton    10799: @c -------------------------------------------------------------
                   10800: @node OOF, Mini-OOF, Objects, Object-oriented Forth
                   10801: @subsection The @file{oof.fs} model
                   10802: @cindex oof
                   10803: @cindex object-oriented programming
1.20      pazsan   10804: 
1.78      anton    10805: @cindex @file{objects.fs}
                   10806: @cindex @file{oof.fs}
1.20      pazsan   10807: 
1.78      anton    10808: This section describes the @file{oof.fs} package.
1.20      pazsan   10809: 
1.78      anton    10810: The package described in this section has been used in bigFORTH since 1991, and
                   10811: used for two large applications: a chromatographic system used to
                   10812: create new medicaments, and a graphic user interface library (MINOS).
1.20      pazsan   10813: 
1.78      anton    10814: You can find a description (in German) of @file{oof.fs} in @cite{Object
                   10815: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
                   10816: 10(2), 1994.
1.20      pazsan   10817: 
1.78      anton    10818: @menu
                   10819: * Properties of the OOF model::  
                   10820: * Basic OOF Usage::             
                   10821: * The OOF base class::          
                   10822: * Class Declaration::           
                   10823: * Class Implementation::        
                   10824: @end menu
1.20      pazsan   10825: 
1.78      anton    10826: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
                   10827: @subsubsection Properties of the @file{oof.fs} model
                   10828: @cindex @file{oof.fs} properties
1.20      pazsan   10829: 
1.78      anton    10830: @itemize @bullet
                   10831: @item
                   10832: This model combines object oriented programming with information
                   10833: hiding. It helps you writing large application, where scoping is
                   10834: necessary, because it provides class-oriented scoping.
1.20      pazsan   10835: 
1.78      anton    10836: @item
                   10837: Named objects, object pointers, and object arrays can be created,
                   10838: selector invocation uses the ``object selector'' syntax. Selector invocation
                   10839: to objects and/or selectors on the stack is a bit less convenient, but
                   10840: possible.
1.44      crook    10841: 
1.78      anton    10842: @item
                   10843: Selector invocation and instance variable usage of the active object is
                   10844: straightforward, since both make use of the active object.
1.44      crook    10845: 
1.78      anton    10846: @item
                   10847: Late binding is efficient and easy to use.
1.20      pazsan   10848: 
1.78      anton    10849: @item
                   10850: State-smart objects parse selectors. However, extensibility is provided
                   10851: using a (parsing) selector @code{postpone} and a selector @code{'}.
1.20      pazsan   10852: 
1.78      anton    10853: @item
                   10854: An implementation in ANS Forth is available.
1.20      pazsan   10855: 
1.78      anton    10856: @end itemize
1.23      crook    10857: 
                   10858: 
1.78      anton    10859: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
                   10860: @subsubsection Basic @file{oof.fs} Usage
                   10861: @cindex @file{oof.fs} usage
1.23      crook    10862: 
1.78      anton    10863: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
1.23      crook    10864: 
1.78      anton    10865: You can define a class for graphical objects like this:
1.23      crook    10866: 
1.78      anton    10867: @cindex @code{class} usage
                   10868: @cindex @code{class;} usage
                   10869: @cindex @code{method} usage
                   10870: @example
                   10871: object class graphical \ "object" is the parent class
                   10872:   method draw ( x y graphical -- )
                   10873: class;
                   10874: @end example
1.23      crook    10875: 
1.78      anton    10876: This code defines a class @code{graphical} with an
                   10877: operation @code{draw}.  We can perform the operation
                   10878: @code{draw} on any @code{graphical} object, e.g.:
1.23      crook    10879: 
1.78      anton    10880: @example
                   10881: 100 100 t-rex draw
                   10882: @end example
1.23      crook    10883: 
1.78      anton    10884: @noindent
                   10885: where @code{t-rex} is an object or object pointer, created with e.g.
                   10886: @code{graphical : t-rex}.
1.23      crook    10887: 
1.78      anton    10888: @cindex abstract class
                   10889: How do we create a graphical object? With the present definitions,
                   10890: we cannot create a useful graphical object. The class
                   10891: @code{graphical} describes graphical objects in general, but not
                   10892: any concrete graphical object type (C++ users would call it an
                   10893: @emph{abstract class}); e.g., there is no method for the selector
                   10894: @code{draw} in the class @code{graphical}.
1.23      crook    10895: 
1.78      anton    10896: For concrete graphical objects, we define child classes of the
                   10897: class @code{graphical}, e.g.:
1.23      crook    10898: 
1.78      anton    10899: @example
                   10900: graphical class circle \ "graphical" is the parent class
                   10901:   cell var circle-radius
                   10902: how:
                   10903:   : draw ( x y -- )
                   10904:     circle-radius @@ draw-circle ;
1.23      crook    10905: 
1.78      anton    10906:   : init ( n-radius -- (
                   10907:     circle-radius ! ;
                   10908: class;
                   10909: @end example
1.1       anton    10910: 
1.78      anton    10911: Here we define a class @code{circle} as a child of @code{graphical},
                   10912: with a field @code{circle-radius}; it defines new methods for the
                   10913: selectors @code{draw} and @code{init} (@code{init} is defined in
                   10914: @code{object}, the parent class of @code{graphical}).
1.1       anton    10915: 
1.78      anton    10916: Now we can create a circle in the dictionary with:
1.1       anton    10917: 
1.78      anton    10918: @example
                   10919: 50 circle : my-circle
                   10920: @end example
1.21      crook    10921: 
1.78      anton    10922: @noindent
                   10923: @code{:} invokes @code{init}, thus initializing the field
                   10924: @code{circle-radius} with 50. We can draw this new circle at (100,100)
                   10925: with:
1.1       anton    10926: 
1.78      anton    10927: @example
                   10928: 100 100 my-circle draw
                   10929: @end example
1.1       anton    10930: 
1.78      anton    10931: @cindex selector invocation, restrictions
                   10932: @cindex class definition, restrictions
                   10933: Note: You can only invoke a selector if the receiving object belongs to
                   10934: the class where the selector was defined or one of its descendents;
                   10935: e.g., you can invoke @code{draw} only for objects belonging to
                   10936: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
                   10937: mechanism will check if you try to invoke a selector that is not
                   10938: defined in this class hierarchy, so you'll get an error at compilation
                   10939: time.
1.1       anton    10940: 
                   10941: 
1.78      anton    10942: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
                   10943: @subsubsection The @file{oof.fs} base class
                   10944: @cindex @file{oof.fs} base class
1.1       anton    10945: 
1.78      anton    10946: When you define a class, you have to specify a parent class.  So how do
                   10947: you start defining classes? There is one class available from the start:
                   10948: @code{object}. You have to use it as ancestor for all classes. It is the
                   10949: only class that has no parent. Classes are also objects, except that
                   10950: they don't have instance variables; class manipulation such as
                   10951: inheritance or changing definitions of a class is handled through
                   10952: selectors of the class @code{object}.
1.1       anton    10953: 
1.78      anton    10954: @code{object} provides a number of selectors:
1.1       anton    10955: 
1.78      anton    10956: @itemize @bullet
                   10957: @item
                   10958: @code{class} for subclassing, @code{definitions} to add definitions
                   10959: later on, and @code{class?} to get type informations (is the class a
                   10960: subclass of the class passed on the stack?).
1.1       anton    10961: 
1.78      anton    10962: doc---object-class
                   10963: doc---object-definitions
                   10964: doc---object-class?
1.1       anton    10965: 
                   10966: 
1.26      crook    10967: @item
1.78      anton    10968: @code{init} and @code{dispose} as constructor and destructor of the
                   10969: object. @code{init} is invocated after the object's memory is allocated,
                   10970: while @code{dispose} also handles deallocation. Thus if you redefine
                   10971: @code{dispose}, you have to call the parent's dispose with @code{super
                   10972: dispose}, too.
                   10973: 
                   10974: doc---object-init
                   10975: doc---object-dispose
                   10976: 
1.1       anton    10977: 
1.26      crook    10978: @item
1.78      anton    10979: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
                   10980: @code{[]} to create named and unnamed objects and object arrays or
                   10981: object pointers.
                   10982: 
                   10983: doc---object-new
                   10984: doc---object-new[]
                   10985: doc---object-:
                   10986: doc---object-ptr
                   10987: doc---object-asptr
                   10988: doc---object-[]
                   10989: 
1.1       anton    10990: 
1.26      crook    10991: @item
1.78      anton    10992: @code{::} and @code{super} for explicit scoping. You should use explicit
                   10993: scoping only for super classes or classes with the same set of instance
                   10994: variables. Explicitly-scoped selectors use early binding.
1.21      crook    10995: 
1.78      anton    10996: doc---object-::
                   10997: doc---object-super
1.21      crook    10998: 
                   10999: 
1.26      crook    11000: @item
1.78      anton    11001: @code{self} to get the address of the object
1.21      crook    11002: 
1.78      anton    11003: doc---object-self
1.21      crook    11004: 
                   11005: 
1.78      anton    11006: @item
                   11007: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
                   11008: pointers and instance defers.
1.21      crook    11009: 
1.78      anton    11010: doc---object-bind
                   11011: doc---object-bound
                   11012: doc---object-link
                   11013: doc---object-is
1.21      crook    11014: 
                   11015: 
1.78      anton    11016: @item
                   11017: @code{'} to obtain selector tokens, @code{send} to invocate selectors
                   11018: form the stack, and @code{postpone} to generate selector invocation code.
1.21      crook    11019: 
1.78      anton    11020: doc---object-'
                   11021: doc---object-postpone
1.21      crook    11022: 
                   11023: 
1.78      anton    11024: @item
                   11025: @code{with} and @code{endwith} to select the active object from the
                   11026: stack, and enable its scope. Using @code{with} and @code{endwith}
                   11027: also allows you to create code using selector @code{postpone} without being
                   11028: trapped by the state-smart objects.
1.21      crook    11029: 
1.78      anton    11030: doc---object-with
                   11031: doc---object-endwith
1.21      crook    11032: 
                   11033: 
1.78      anton    11034: @end itemize
1.21      crook    11035: 
1.78      anton    11036: @node Class Declaration, Class Implementation, The OOF base class, OOF
                   11037: @subsubsection Class Declaration
                   11038: @cindex class declaration
1.21      crook    11039: 
1.78      anton    11040: @itemize @bullet
                   11041: @item
                   11042: Instance variables
1.21      crook    11043: 
1.78      anton    11044: doc---oof-var
1.21      crook    11045: 
                   11046: 
1.78      anton    11047: @item
                   11048: Object pointers
1.21      crook    11049: 
1.78      anton    11050: doc---oof-ptr
                   11051: doc---oof-asptr
1.21      crook    11052: 
                   11053: 
1.78      anton    11054: @item
                   11055: Instance defers
1.21      crook    11056: 
1.78      anton    11057: doc---oof-defer
1.21      crook    11058: 
                   11059: 
1.78      anton    11060: @item
                   11061: Method selectors
1.21      crook    11062: 
1.78      anton    11063: doc---oof-early
                   11064: doc---oof-method
1.21      crook    11065: 
                   11066: 
1.78      anton    11067: @item
                   11068: Class-wide variables
1.21      crook    11069: 
1.78      anton    11070: doc---oof-static
1.21      crook    11071: 
                   11072: 
1.78      anton    11073: @item
                   11074: End declaration
1.1       anton    11075: 
1.78      anton    11076: doc---oof-how:
                   11077: doc---oof-class;
1.21      crook    11078: 
                   11079: 
1.78      anton    11080: @end itemize
1.21      crook    11081: 
1.78      anton    11082: @c -------------------------------------------------------------
                   11083: @node Class Implementation,  , Class Declaration, OOF
                   11084: @subsubsection Class Implementation
                   11085: @cindex class implementation
1.21      crook    11086: 
1.78      anton    11087: @c -------------------------------------------------------------
                   11088: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
                   11089: @subsection The @file{mini-oof.fs} model
                   11090: @cindex mini-oof
1.21      crook    11091: 
1.78      anton    11092: Gforth's third object oriented Forth package is a 12-liner. It uses a
1.79      anton    11093: mixture of the @file{objects.fs} and the @file{oof.fs} syntax,
1.78      anton    11094: and reduces to the bare minimum of features. This is based on a posting
                   11095: of Bernd Paysan in comp.lang.forth.
1.21      crook    11096: 
1.78      anton    11097: @menu
                   11098: * Basic Mini-OOF Usage::        
                   11099: * Mini-OOF Example::            
                   11100: * Mini-OOF Implementation::     
                   11101: @end menu
1.21      crook    11102: 
1.78      anton    11103: @c -------------------------------------------------------------
                   11104: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
                   11105: @subsubsection Basic @file{mini-oof.fs} Usage
                   11106: @cindex mini-oof usage
1.21      crook    11107: 
1.78      anton    11108: There is a base class (@code{class}, which allocates one cell for the
                   11109: object pointer) plus seven other words: to define a method, a variable,
                   11110: a class; to end a class, to resolve binding, to allocate an object and
                   11111: to compile a class method.
                   11112: @comment TODO better description of the last one
1.26      crook    11113: 
1.21      crook    11114: 
1.78      anton    11115: doc-object
                   11116: doc-method
                   11117: doc-var
                   11118: doc-class
                   11119: doc-end-class
                   11120: doc-defines
                   11121: doc-new
                   11122: doc-::
1.21      crook    11123: 
                   11124: 
                   11125: 
1.78      anton    11126: @c -------------------------------------------------------------
                   11127: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
                   11128: @subsubsection Mini-OOF Example
                   11129: @cindex mini-oof example
1.1       anton    11130: 
1.78      anton    11131: A short example shows how to use this package. This example, in slightly
                   11132: extended form, is supplied as @file{moof-exm.fs}
                   11133: @comment TODO could flesh this out with some comments from the Forthwrite article
1.20      pazsan   11134: 
1.26      crook    11135: @example
1.78      anton    11136: object class
                   11137:   method init
                   11138:   method draw
                   11139: end-class graphical
1.26      crook    11140: @end example
1.20      pazsan   11141: 
1.78      anton    11142: This code defines a class @code{graphical} with an
                   11143: operation @code{draw}.  We can perform the operation
                   11144: @code{draw} on any @code{graphical} object, e.g.:
1.20      pazsan   11145: 
1.26      crook    11146: @example
1.78      anton    11147: 100 100 t-rex draw
1.26      crook    11148: @end example
1.12      anton    11149: 
1.78      anton    11150: where @code{t-rex} is an object or object pointer, created with e.g.
                   11151: @code{graphical new Constant t-rex}.
1.12      anton    11152: 
1.78      anton    11153: For concrete graphical objects, we define child classes of the
                   11154: class @code{graphical}, e.g.:
1.12      anton    11155: 
1.26      crook    11156: @example
                   11157: graphical class
1.78      anton    11158:   cell var circle-radius
                   11159: end-class circle \ "graphical" is the parent class
1.12      anton    11160: 
1.78      anton    11161: :noname ( x y -- )
                   11162:   circle-radius @@ draw-circle ; circle defines draw
                   11163: :noname ( r -- )
                   11164:   circle-radius ! ; circle defines init
                   11165: @end example
1.12      anton    11166: 
1.78      anton    11167: There is no implicit init method, so we have to define one. The creation
                   11168: code of the object now has to call init explicitely.
1.21      crook    11169: 
1.78      anton    11170: @example
                   11171: circle new Constant my-circle
                   11172: 50 my-circle init
1.12      anton    11173: @end example
                   11174: 
1.78      anton    11175: It is also possible to add a function to create named objects with
                   11176: automatic call of @code{init}, given that all objects have @code{init}
                   11177: on the same place:
1.38      anton    11178: 
1.78      anton    11179: @example
                   11180: : new: ( .. o "name" -- )
                   11181:     new dup Constant init ;
                   11182: 80 circle new: large-circle
                   11183: @end example
1.12      anton    11184: 
1.78      anton    11185: We can draw this new circle at (100,100) with:
1.12      anton    11186: 
1.78      anton    11187: @example
                   11188: 100 100 my-circle draw
                   11189: @end example
1.12      anton    11190: 
1.78      anton    11191: @node Mini-OOF Implementation,  , Mini-OOF Example, Mini-OOF
                   11192: @subsubsection @file{mini-oof.fs} Implementation
1.12      anton    11193: 
1.78      anton    11194: Object-oriented systems with late binding typically use a
                   11195: ``vtable''-approach: the first variable in each object is a pointer to a
                   11196: table, which contains the methods as function pointers. The vtable
                   11197: may also contain other information.
1.12      anton    11198: 
1.79      anton    11199: So first, let's declare selectors:
1.37      anton    11200: 
                   11201: @example
1.79      anton    11202: : method ( m v "name" -- m' v ) Create  over , swap cell+ swap
1.78      anton    11203:   DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
                   11204: @end example
1.37      anton    11205: 
1.79      anton    11206: During selector declaration, the number of selectors and instance
                   11207: variables is on the stack (in address units). @code{method} creates one
                   11208: selector and increments the selector number. To execute a selector, it
1.78      anton    11209: takes the object, fetches the vtable pointer, adds the offset, and
1.79      anton    11210: executes the method @i{xt} stored there. Each selector takes the object
                   11211: it is invoked with as top of stack parameter; it passes the parameters
                   11212: (including the object) unchanged to the appropriate method which should
1.78      anton    11213: consume that object.
1.37      anton    11214: 
1.78      anton    11215: Now, we also have to declare instance variables
1.37      anton    11216: 
1.78      anton    11217: @example
1.79      anton    11218: : var ( m v size "name" -- m v' ) Create  over , +
1.78      anton    11219:   DOES> ( o -- addr ) @@ + ;
1.37      anton    11220: @end example
                   11221: 
1.78      anton    11222: As before, a word is created with the current offset. Instance
                   11223: variables can have different sizes (cells, floats, doubles, chars), so
                   11224: all we do is take the size and add it to the offset. If your machine
                   11225: has alignment restrictions, put the proper @code{aligned} or
                   11226: @code{faligned} before the variable, to adjust the variable
                   11227: offset. That's why it is on the top of stack.
1.37      anton    11228: 
1.78      anton    11229: We need a starting point (the base object) and some syntactic sugar:
1.37      anton    11230: 
1.78      anton    11231: @example
                   11232: Create object  1 cells , 2 cells ,
1.79      anton    11233: : class ( class -- class selectors vars ) dup 2@@ ;
1.78      anton    11234: @end example
1.12      anton    11235: 
1.78      anton    11236: For inheritance, the vtable of the parent object has to be
                   11237: copied when a new, derived class is declared. This gives all the
                   11238: methods of the parent class, which can be overridden, though.
1.12      anton    11239: 
1.78      anton    11240: @example
1.79      anton    11241: : end-class  ( class selectors vars "name" -- )
1.78      anton    11242:   Create  here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
                   11243:   cell+ dup cell+ r> rot @@ 2 cells /string move ;
                   11244: @end example
1.12      anton    11245: 
1.78      anton    11246: The first line creates the vtable, initialized with
                   11247: @code{noop}s. The second line is the inheritance mechanism, it
                   11248: copies the xts from the parent vtable.
1.12      anton    11249: 
1.78      anton    11250: We still have no way to define new methods, let's do that now:
1.12      anton    11251: 
1.26      crook    11252: @example
1.79      anton    11253: : defines ( xt class "name" -- ) ' >body @@ + ! ;
1.78      anton    11254: @end example
1.12      anton    11255: 
1.78      anton    11256: To allocate a new object, we need a word, too:
1.12      anton    11257: 
1.78      anton    11258: @example
                   11259: : new ( class -- o )  here over @@ allot swap over ! ;
1.12      anton    11260: @end example
                   11261: 
1.78      anton    11262: Sometimes derived classes want to access the method of the
                   11263: parent object. There are two ways to achieve this with Mini-OOF:
                   11264: first, you could use named words, and second, you could look up the
                   11265: vtable of the parent object.
1.12      anton    11266: 
1.78      anton    11267: @example
                   11268: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
                   11269: @end example
1.12      anton    11270: 
                   11271: 
1.78      anton    11272: Nothing can be more confusing than a good example, so here is
                   11273: one. First let's declare a text object (called
                   11274: @code{button}), that stores text and position:
1.12      anton    11275: 
1.78      anton    11276: @example
                   11277: object class
                   11278:   cell var text
                   11279:   cell var len
                   11280:   cell var x
                   11281:   cell var y
                   11282:   method init
                   11283:   method draw
                   11284: end-class button
                   11285: @end example
1.12      anton    11286: 
1.78      anton    11287: @noindent
                   11288: Now, implement the two methods, @code{draw} and @code{init}:
1.21      crook    11289: 
1.26      crook    11290: @example
1.78      anton    11291: :noname ( o -- )
                   11292:  >r r@@ x @@ r@@ y @@ at-xy  r@@ text @@ r> len @@ type ;
                   11293:  button defines draw
                   11294: :noname ( addr u o -- )
                   11295:  >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
                   11296:  button defines init
1.26      crook    11297: @end example
1.12      anton    11298: 
1.78      anton    11299: @noindent
                   11300: To demonstrate inheritance, we define a class @code{bold-button}, with no
1.79      anton    11301: new data and no new selectors:
1.78      anton    11302: 
                   11303: @example
                   11304: button class
                   11305: end-class bold-button
1.12      anton    11306: 
1.78      anton    11307: : bold   27 emit ." [1m" ;
                   11308: : normal 27 emit ." [0m" ;
                   11309: @end example
1.1       anton    11310: 
1.78      anton    11311: @noindent
                   11312: The class @code{bold-button} has a different draw method to
                   11313: @code{button}, but the new method is defined in terms of the draw method
                   11314: for @code{button}:
1.20      pazsan   11315: 
1.78      anton    11316: @example
                   11317: :noname bold [ button :: draw ] normal ; bold-button defines draw
                   11318: @end example
1.21      crook    11319: 
1.78      anton    11320: @noindent
1.79      anton    11321: Finally, create two objects and apply selectors:
1.21      crook    11322: 
1.26      crook    11323: @example
1.78      anton    11324: button new Constant foo
                   11325: s" thin foo" foo init
                   11326: page
                   11327: foo draw
                   11328: bold-button new Constant bar
                   11329: s" fat bar" bar init
                   11330: 1 bar y !
                   11331: bar draw
1.26      crook    11332: @end example
1.21      crook    11333: 
                   11334: 
1.78      anton    11335: @node Comparison with other object models,  , Mini-OOF, Object-oriented Forth
                   11336: @subsection Comparison with other object models
                   11337: @cindex comparison of object models
                   11338: @cindex object models, comparison
                   11339: 
                   11340: Many object-oriented Forth extensions have been proposed (@cite{A survey
                   11341: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
                   11342: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
                   11343: relation of the object models described here to two well-known and two
                   11344: closely-related (by the use of method maps) models.  Andras Zsoter
                   11345: helped us with this section.
                   11346: 
                   11347: @cindex Neon model
                   11348: The most popular model currently seems to be the Neon model (see
                   11349: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
                   11350: 1997) by Andrew McKewan) but this model has a number of limitations
                   11351: @footnote{A longer version of this critique can be
                   11352: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
                   11353: Dimensions, May 1997) by Anton Ertl.}:
                   11354: 
                   11355: @itemize @bullet
                   11356: @item
                   11357: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
                   11358: to pass objects on the stack.
1.21      crook    11359: 
1.78      anton    11360: @item
                   11361: It requires that the selector parses the input stream (at
1.79      anton    11362: compile time); this leads to reduced extensibility and to bugs that are
1.78      anton    11363: hard to find.
1.21      crook    11364: 
1.78      anton    11365: @item
1.79      anton    11366: It allows using every selector on every object; this eliminates the
                   11367: need for interfaces, but makes it harder to create efficient
                   11368: implementations.
1.78      anton    11369: @end itemize
1.21      crook    11370: 
1.78      anton    11371: @cindex Pountain's object-oriented model
                   11372: Another well-known publication is @cite{Object-Oriented Forth} (Academic
                   11373: Press, London, 1987) by Dick Pountain. However, it is not really about
                   11374: object-oriented programming, because it hardly deals with late
                   11375: binding. Instead, it focuses on features like information hiding and
                   11376: overloading that are characteristic of modular languages like Ada (83).
1.26      crook    11377: 
1.78      anton    11378: @cindex Zsoter's object-oriented model
1.79      anton    11379: In @uref{http://www.forth.org/oopf.html, Does late binding have to be
                   11380: slow?} (Forth Dimensions 18(1) 1996, pages 31-35) Andras Zsoter
                   11381: describes a model that makes heavy use of an active object (like
                   11382: @code{this} in @file{objects.fs}): The active object is not only used
                   11383: for accessing all fields, but also specifies the receiving object of
                   11384: every selector invocation; you have to change the active object
                   11385: explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
                   11386: changes more or less implicitly at @code{m: ... ;m}. Such a change at
                   11387: the method entry point is unnecessary with Zsoter's model, because the
                   11388: receiving object is the active object already. On the other hand, the
                   11389: explicit change is absolutely necessary in that model, because otherwise
                   11390: no one could ever change the active object. An ANS Forth implementation
                   11391: of this model is available through
                   11392: @uref{http://www.forth.org/oopf.html}.
1.21      crook    11393: 
1.78      anton    11394: @cindex @file{oof.fs}, differences to other models
                   11395: The @file{oof.fs} model combines information hiding and overloading
                   11396: resolution (by keeping names in various word lists) with object-oriented
                   11397: programming. It sets the active object implicitly on method entry, but
                   11398: also allows explicit changing (with @code{>o...o>} or with
                   11399: @code{with...endwith}). It uses parsing and state-smart objects and
                   11400: classes for resolving overloading and for early binding: the object or
                   11401: class parses the selector and determines the method from this. If the
                   11402: selector is not parsed by an object or class, it performs a call to the
                   11403: selector for the active object (late binding), like Zsoter's model.
                   11404: Fields are always accessed through the active object. The big
                   11405: disadvantage of this model is the parsing and the state-smartness, which
                   11406: reduces extensibility and increases the opportunities for subtle bugs;
                   11407: essentially, you are only safe if you never tick or @code{postpone} an
                   11408: object or class (Bernd disagrees, but I (Anton) am not convinced).
1.21      crook    11409: 
1.78      anton    11410: @cindex @file{mini-oof.fs}, differences to other models
                   11411: The @file{mini-oof.fs} model is quite similar to a very stripped-down
                   11412: version of the @file{objects.fs} model, but syntactically it is a
                   11413: mixture of the @file{objects.fs} and @file{oof.fs} models.
1.21      crook    11414: 
                   11415: 
1.78      anton    11416: @c -------------------------------------------------------------
                   11417: @node Programming Tools, Assembler and Code Words, Object-oriented Forth, Words
                   11418: @section Programming Tools
                   11419: @cindex programming tools
1.21      crook    11420: 
1.78      anton    11421: @c !! move this and assembler down below OO stuff.
1.21      crook    11422: 
1.78      anton    11423: @menu
                   11424: * Examining::                   
                   11425: * Forgetting words::            
                   11426: * Debugging::                   Simple and quick.
                   11427: * Assertions::                  Making your programs self-checking.
                   11428: * Singlestep Debugger::         Executing your program word by word.
                   11429: @end menu
1.21      crook    11430: 
1.78      anton    11431: @node Examining, Forgetting words, Programming Tools, Programming Tools
                   11432: @subsection Examining data and code
                   11433: @cindex examining data and code
                   11434: @cindex data examination
                   11435: @cindex code examination
1.44      crook    11436: 
1.78      anton    11437: The following words inspect the stack non-destructively:
1.21      crook    11438: 
1.78      anton    11439: doc-.s
                   11440: doc-f.s
1.44      crook    11441: 
1.78      anton    11442: There is a word @code{.r} but it does @i{not} display the return stack!
                   11443: It is used for formatted numeric output (@pxref{Simple numeric output}).
1.21      crook    11444: 
1.78      anton    11445: doc-depth
                   11446: doc-fdepth
                   11447: doc-clearstack
1.21      crook    11448: 
1.78      anton    11449: The following words inspect memory.
1.21      crook    11450: 
1.78      anton    11451: doc-?
                   11452: doc-dump
1.21      crook    11453: 
1.78      anton    11454: And finally, @code{see} allows to inspect code:
1.21      crook    11455: 
1.78      anton    11456: doc-see
                   11457: doc-xt-see
1.21      crook    11458: 
1.78      anton    11459: @node Forgetting words, Debugging, Examining, Programming Tools
                   11460: @subsection Forgetting words
                   11461: @cindex words, forgetting
                   11462: @cindex forgeting words
1.21      crook    11463: 
1.78      anton    11464: @c  anton: other, maybe better places for this subsection: Defining Words;
                   11465: @c  Dictionary allocation.  At least a reference should be there.
1.21      crook    11466: 
1.78      anton    11467: Forth allows you to forget words (and everything that was alloted in the
                   11468: dictonary after them) in a LIFO manner.
1.21      crook    11469: 
1.78      anton    11470: doc-marker
1.21      crook    11471: 
1.78      anton    11472: The most common use of this feature is during progam development: when
                   11473: you change a source file, forget all the words it defined and load it
                   11474: again (since you also forget everything defined after the source file
                   11475: was loaded, you have to reload that, too).  Note that effects like
                   11476: storing to variables and destroyed system words are not undone when you
                   11477: forget words.  With a system like Gforth, that is fast enough at
                   11478: starting up and compiling, I find it more convenient to exit and restart
                   11479: Gforth, as this gives me a clean slate.
1.21      crook    11480: 
1.78      anton    11481: Here's an example of using @code{marker} at the start of a source file
                   11482: that you are debugging; it ensures that you only ever have one copy of
                   11483: the file's definitions compiled at any time:
1.21      crook    11484: 
1.78      anton    11485: @example
                   11486: [IFDEF] my-code
                   11487:     my-code
                   11488: [ENDIF]
1.26      crook    11489: 
1.78      anton    11490: marker my-code
                   11491: init-included-files
1.21      crook    11492: 
1.78      anton    11493: \ .. definitions start here
                   11494: \ .
                   11495: \ .
                   11496: \ end
                   11497: @end example
1.21      crook    11498: 
1.26      crook    11499: 
1.78      anton    11500: @node Debugging, Assertions, Forgetting words, Programming Tools
                   11501: @subsection Debugging
                   11502: @cindex debugging
1.21      crook    11503: 
1.78      anton    11504: Languages with a slow edit/compile/link/test development loop tend to
                   11505: require sophisticated tracing/stepping debuggers to facilate debugging.
1.21      crook    11506: 
1.78      anton    11507: A much better (faster) way in fast-compiling languages is to add
                   11508: printing code at well-selected places, let the program run, look at
                   11509: the output, see where things went wrong, add more printing code, etc.,
                   11510: until the bug is found.
1.21      crook    11511: 
1.78      anton    11512: The simple debugging aids provided in @file{debugs.fs}
                   11513: are meant to support this style of debugging.
1.21      crook    11514: 
1.78      anton    11515: The word @code{~~} prints debugging information (by default the source
                   11516: location and the stack contents). It is easy to insert. If you use Emacs
                   11517: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
                   11518: query-replace them with nothing). The deferred words
                   11519: @code{printdebugdata} and @code{printdebugline} control the output of
                   11520: @code{~~}. The default source location output format works well with
                   11521: Emacs' compilation mode, so you can step through the program at the
                   11522: source level using @kbd{C-x `} (the advantage over a stepping debugger
                   11523: is that you can step in any direction and you know where the crash has
                   11524: happened or where the strange data has occurred).
1.21      crook    11525: 
1.78      anton    11526: doc-~~
                   11527: doc-printdebugdata
                   11528: doc-printdebugline
1.21      crook    11529: 
1.78      anton    11530: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
                   11531: @subsection Assertions
                   11532: @cindex assertions
1.21      crook    11533: 
1.78      anton    11534: It is a good idea to make your programs self-checking, especially if you
                   11535: make an assumption that may become invalid during maintenance (for
                   11536: example, that a certain field of a data structure is never zero). Gforth
                   11537: supports @dfn{assertions} for this purpose. They are used like this:
1.21      crook    11538: 
                   11539: @example
1.78      anton    11540: assert( @i{flag} )
1.26      crook    11541: @end example
                   11542: 
1.78      anton    11543: The code between @code{assert(} and @code{)} should compute a flag, that
                   11544: should be true if everything is alright and false otherwise. It should
                   11545: not change anything else on the stack. The overall stack effect of the
                   11546: assertion is @code{( -- )}. E.g.
1.21      crook    11547: 
1.26      crook    11548: @example
1.78      anton    11549: assert( 1 1 + 2 = ) \ what we learn in school
                   11550: assert( dup 0<> ) \ assert that the top of stack is not zero
                   11551: assert( false ) \ this code should not be reached
1.21      crook    11552: @end example
                   11553: 
1.78      anton    11554: The need for assertions is different at different times. During
                   11555: debugging, we want more checking, in production we sometimes care more
                   11556: for speed. Therefore, assertions can be turned off, i.e., the assertion
                   11557: becomes a comment. Depending on the importance of an assertion and the
                   11558: time it takes to check it, you may want to turn off some assertions and
                   11559: keep others turned on. Gforth provides several levels of assertions for
                   11560: this purpose:
                   11561: 
                   11562: 
                   11563: doc-assert0(
                   11564: doc-assert1(
                   11565: doc-assert2(
                   11566: doc-assert3(
                   11567: doc-assert(
                   11568: doc-)
1.21      crook    11569: 
                   11570: 
1.78      anton    11571: The variable @code{assert-level} specifies the highest assertions that
                   11572: are turned on. I.e., at the default @code{assert-level} of one,
                   11573: @code{assert0(} and @code{assert1(} assertions perform checking, while
                   11574: @code{assert2(} and @code{assert3(} assertions are treated as comments.
1.26      crook    11575: 
1.78      anton    11576: The value of @code{assert-level} is evaluated at compile-time, not at
                   11577: run-time. Therefore you cannot turn assertions on or off at run-time;
                   11578: you have to set the @code{assert-level} appropriately before compiling a
                   11579: piece of code. You can compile different pieces of code at different
                   11580: @code{assert-level}s (e.g., a trusted library at level 1 and
                   11581: newly-written code at level 3).
1.26      crook    11582: 
                   11583: 
1.78      anton    11584: doc-assert-level
1.26      crook    11585: 
                   11586: 
1.78      anton    11587: If an assertion fails, a message compatible with Emacs' compilation mode
                   11588: is produced and the execution is aborted (currently with @code{ABORT"}.
                   11589: If there is interest, we will introduce a special throw code. But if you
                   11590: intend to @code{catch} a specific condition, using @code{throw} is
                   11591: probably more appropriate than an assertion).
1.44      crook    11592: 
1.78      anton    11593: Definitions in ANS Forth for these assertion words are provided
                   11594: in @file{compat/assert.fs}.
1.26      crook    11595: 
1.44      crook    11596: 
1.78      anton    11597: @node Singlestep Debugger,  , Assertions, Programming Tools
                   11598: @subsection Singlestep Debugger
                   11599: @cindex singlestep Debugger
                   11600: @cindex debugging Singlestep
1.44      crook    11601: 
1.78      anton    11602: When you create a new word there's often the need to check whether it
                   11603: behaves correctly or not. You can do this by typing @code{dbg
                   11604: badword}. A debug session might look like this:
1.26      crook    11605: 
1.78      anton    11606: @example
                   11607: : badword 0 DO i . LOOP ;  ok
                   11608: 2 dbg badword 
                   11609: : badword  
                   11610: Scanning code...
1.44      crook    11611: 
1.78      anton    11612: Nesting debugger ready!
1.44      crook    11613: 
1.78      anton    11614: 400D4738  8049BC4 0              -> [ 2 ] 00002 00000 
                   11615: 400D4740  8049F68 DO             -> [ 0 ] 
                   11616: 400D4744  804A0C8 i              -> [ 1 ] 00000 
                   11617: 400D4748 400C5E60 .              -> 0 [ 0 ] 
                   11618: 400D474C  8049D0C LOOP           -> [ 0 ] 
                   11619: 400D4744  804A0C8 i              -> [ 1 ] 00001 
                   11620: 400D4748 400C5E60 .              -> 1 [ 0 ] 
                   11621: 400D474C  8049D0C LOOP           -> [ 0 ] 
                   11622: 400D4758  804B384 ;              ->  ok
                   11623: @end example
1.21      crook    11624: 
1.78      anton    11625: Each line displayed is one step. You always have to hit return to
                   11626: execute the next word that is displayed. If you don't want to execute
                   11627: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
                   11628: an overview what keys are available:
1.44      crook    11629: 
1.78      anton    11630: @table @i
1.44      crook    11631: 
1.78      anton    11632: @item @key{RET}
                   11633: Next; Execute the next word.
1.21      crook    11634: 
1.78      anton    11635: @item n
                   11636: Nest; Single step through next word.
1.44      crook    11637: 
1.78      anton    11638: @item u
                   11639: Unnest; Stop debugging and execute rest of word. If we got to this word
                   11640: with nest, continue debugging with the calling word.
1.44      crook    11641: 
1.78      anton    11642: @item d
                   11643: Done; Stop debugging and execute rest.
1.21      crook    11644: 
1.78      anton    11645: @item s
                   11646: Stop; Abort immediately.
1.44      crook    11647: 
1.78      anton    11648: @end table
1.44      crook    11649: 
1.78      anton    11650: Debugging large application with this mechanism is very difficult, because
                   11651: you have to nest very deeply into the program before the interesting part
                   11652: begins. This takes a lot of time. 
1.26      crook    11653: 
1.78      anton    11654: To do it more directly put a @code{BREAK:} command into your source code.
                   11655: When program execution reaches @code{BREAK:} the single step debugger is
                   11656: invoked and you have all the features described above.
1.44      crook    11657: 
1.78      anton    11658: If you have more than one part to debug it is useful to know where the
                   11659: program has stopped at the moment. You can do this by the 
                   11660: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
                   11661: string is typed out when the ``breakpoint'' is reached.
1.44      crook    11662: 
1.26      crook    11663: 
1.78      anton    11664: doc-dbg
                   11665: doc-break:
                   11666: doc-break"
1.44      crook    11667: 
                   11668: 
1.26      crook    11669: 
1.78      anton    11670: @c -------------------------------------------------------------
                   11671: @node Assembler and Code Words, Threading Words, Programming Tools, Words
                   11672: @section Assembler and Code Words
                   11673: @cindex assembler
                   11674: @cindex code words
1.44      crook    11675: 
1.78      anton    11676: @menu
                   11677: * Code and ;code::              
                   11678: * Common Assembler::            Assembler Syntax
                   11679: * Common Disassembler::         
                   11680: * 386 Assembler::               Deviations and special cases
                   11681: * Alpha Assembler::             Deviations and special cases
                   11682: * MIPS assembler::              Deviations and special cases
                   11683: * Other assemblers::            How to write them
                   11684: @end menu
1.21      crook    11685: 
1.78      anton    11686: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
                   11687: @subsection @code{Code} and @code{;code}
1.26      crook    11688: 
1.78      anton    11689: Gforth provides some words for defining primitives (words written in
                   11690: machine code), and for defining the machine-code equivalent of
                   11691: @code{DOES>}-based defining words. However, the machine-independent
                   11692: nature of Gforth poses a few problems: First of all, Gforth runs on
                   11693: several architectures, so it can provide no standard assembler. What's
                   11694: worse is that the register allocation not only depends on the processor,
                   11695: but also on the @code{gcc} version and options used.
1.44      crook    11696: 
1.78      anton    11697: The words that Gforth offers encapsulate some system dependences (e.g.,
                   11698: the header structure), so a system-independent assembler may be used in
                   11699: Gforth. If you do not have an assembler, you can compile machine code
                   11700: directly with @code{,} and @code{c,}@footnote{This isn't portable,
                   11701: because these words emit stuff in @i{data} space; it works because
                   11702: Gforth has unified code/data spaces. Assembler isn't likely to be
                   11703: portable anyway.}.
1.21      crook    11704: 
1.44      crook    11705: 
1.78      anton    11706: doc-assembler
                   11707: doc-init-asm
                   11708: doc-code
                   11709: doc-end-code
                   11710: doc-;code
                   11711: doc-flush-icache
1.44      crook    11712: 
1.21      crook    11713: 
1.78      anton    11714: If @code{flush-icache} does not work correctly, @code{code} words
                   11715: etc. will not work (reliably), either.
1.44      crook    11716: 
1.78      anton    11717: The typical usage of these @code{code} words can be shown most easily by
                   11718: analogy to the equivalent high-level defining words:
1.44      crook    11719: 
1.78      anton    11720: @example
                   11721: : foo                              code foo
                   11722:    <high-level Forth words>              <assembler>
                   11723: ;                                  end-code
                   11724:                                 
                   11725: : bar                              : bar
                   11726:    <high-level Forth words>           <high-level Forth words>
                   11727:    CREATE                             CREATE
                   11728:       <high-level Forth words>           <high-level Forth words>
                   11729:    DOES>                              ;code
                   11730:       <high-level Forth words>           <assembler>
                   11731: ;                                  end-code
                   11732: @end example
1.21      crook    11733: 
1.78      anton    11734: @c anton: the following stuff is also in "Common Assembler", in less detail.
1.44      crook    11735: 
1.78      anton    11736: @cindex registers of the inner interpreter
                   11737: In the assembly code you will want to refer to the inner interpreter's
                   11738: registers (e.g., the data stack pointer) and you may want to use other
                   11739: registers for temporary storage. Unfortunately, the register allocation
                   11740: is installation-dependent.
1.44      crook    11741: 
1.78      anton    11742: In particular, @code{ip} (Forth instruction pointer) and @code{rp}
                   11743: (return stack pointer) are in different places in @code{gforth} and
                   11744: @code{gforth-fast}.  This means that you cannot write a @code{NEXT}
                   11745: routine that works on both versions; so for doing @code{NEXT}, I
                   11746: recomment jumping to @code{' noop >code-address}, which contains nothing
                   11747: but a @code{NEXT}.
1.21      crook    11748: 
1.78      anton    11749: For general accesses to the inner interpreter's registers, the easiest
                   11750: solution is to use explicit register declarations (@pxref{Explicit Reg
                   11751: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) for
                   11752: all of the inner interpreter's registers: You have to compile Gforth
                   11753: with @code{-DFORCE_REG} (configure option @code{--enable-force-reg}) and
                   11754: the appropriate declarations must be present in the @code{machine.h}
                   11755: file (see @code{mips.h} for an example; you can find a full list of all
                   11756: declarable register symbols with @code{grep register engine.c}). If you
                   11757: give explicit registers to all variables that are declared at the
                   11758: beginning of @code{engine()}, you should be able to use the other
                   11759: caller-saved registers for temporary storage. Alternatively, you can use
                   11760: the @code{gcc} option @code{-ffixed-REG} (@pxref{Code Gen Options, ,
                   11761: Options for Code Generation Conventions, gcc.info, GNU C Manual}) to
                   11762: reserve a register (however, this restriction on register allocation may
                   11763: slow Gforth significantly).
1.44      crook    11764: 
1.78      anton    11765: If this solution is not viable (e.g., because @code{gcc} does not allow
                   11766: you to explicitly declare all the registers you need), you have to find
                   11767: out by looking at the code where the inner interpreter's registers
                   11768: reside and which registers can be used for temporary storage. You can
                   11769: get an assembly listing of the engine's code with @code{make engine.s}.
1.44      crook    11770: 
1.78      anton    11771: In any case, it is good practice to abstract your assembly code from the
                   11772: actual register allocation. E.g., if the data stack pointer resides in
                   11773: register @code{$17}, create an alias for this register called @code{sp},
                   11774: and use that in your assembly code.
1.21      crook    11775: 
1.78      anton    11776: @cindex code words, portable
                   11777: Another option for implementing normal and defining words efficiently
                   11778: is to add the desired functionality to the source of Gforth. For normal
                   11779: words you just have to edit @file{primitives} (@pxref{Automatic
                   11780: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
                   11781: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
                   11782: @file{prims2x.fs}, and possibly @file{cross.fs}.
1.44      crook    11783: 
1.78      anton    11784: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
                   11785: @subsection Common Assembler
1.44      crook    11786: 
1.78      anton    11787: The assemblers in Gforth generally use a postfix syntax, i.e., the
                   11788: instruction name follows the operands.
1.21      crook    11789: 
1.78      anton    11790: The operands are passed in the usual order (the same that is used in the
                   11791: manual of the architecture).  Since they all are Forth words, they have
                   11792: to be separated by spaces; you can also use Forth words to compute the
                   11793: operands.
1.44      crook    11794: 
1.78      anton    11795: The instruction names usually end with a @code{,}.  This makes it easier
                   11796: to visually separate instructions if you put several of them on one
                   11797: line; it also avoids shadowing other Forth words (e.g., @code{and}).
1.21      crook    11798: 
1.78      anton    11799: Registers are usually specified by number; e.g., (decimal) @code{11}
                   11800: specifies registers R11 and F11 on the Alpha architecture (which one,
                   11801: depends on the instruction).  The usual names are also available, e.g.,
                   11802: @code{s2} for R11 on Alpha.
1.21      crook    11803: 
1.78      anton    11804: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
                   11805: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
                   11806: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
                   11807: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}.  The
                   11808: conditions are specified in a way specific to each assembler.
1.1       anton    11809: 
1.78      anton    11810: Note that the register assignments of the Gforth engine can change
                   11811: between Gforth versions, or even between different compilations of the
                   11812: same Gforth version (e.g., if you use a different GCC version).  So if
                   11813: you want to refer to Gforth's registers (e.g., the stack pointer or
                   11814: TOS), I recommend defining your own words for refering to these
                   11815: registers, and using them later on; then you can easily adapt to a
                   11816: changed register assignment.  The stability of the register assignment
                   11817: is usually better if you build Gforth with @code{--enable-force-reg}.
1.1       anton    11818: 
1.78      anton    11819: In particular, the return stack pointer and the instruction pointer are
                   11820: in memory in @code{gforth}, and usually in registers in
                   11821: @code{gforth-fast}.  The most common use of these registers is to
                   11822: dispatch to the next word (the @code{next} routine).  A portable way to
                   11823: do this is to jump to @code{' noop >code-address} (of course, this is
                   11824: less efficient than integrating the @code{next} code and scheduling it
                   11825: well).
1.1       anton    11826: 
1.78      anton    11827: @node  Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
                   11828: @subsection Common Disassembler
1.1       anton    11829: 
1.78      anton    11830: You can disassemble a @code{code} word with @code{see}
                   11831: (@pxref{Debugging}).  You can disassemble a section of memory with
1.1       anton    11832: 
1.78      anton    11833: doc-disasm
1.44      crook    11834: 
1.78      anton    11835: The disassembler generally produces output that can be fed into the
                   11836: assembler (i.e., same syntax, etc.).  It also includes additional
                   11837: information in comments.  In particular, the address of the instruction
                   11838: is given in a comment before the instruction.
1.1       anton    11839: 
1.78      anton    11840: @code{See} may display more or less than the actual code of the word,
                   11841: because the recognition of the end of the code is unreliable.  You can
                   11842: use @code{disasm} if it did not display enough.  It may display more, if
                   11843: the code word is not immediately followed by a named word.  If you have
                   11844: something else there, you can follow the word with @code{align last @ ,}
                   11845: to ensure that the end is recognized.
1.21      crook    11846: 
1.78      anton    11847: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
                   11848: @subsection 386 Assembler
1.44      crook    11849: 
1.78      anton    11850: The 386 assembler included in Gforth was written by Bernd Paysan, it's
                   11851: available under GPL, and originally part of bigFORTH.
1.21      crook    11852: 
1.78      anton    11853: The 386 disassembler included in Gforth was written by Andrew McKewan
                   11854: and is in the public domain.
1.21      crook    11855: 
1.91    ! anton    11856: The disassembler displays code in an Intel-like prefix syntax.
1.21      crook    11857: 
1.78      anton    11858: The assembler uses a postfix syntax with reversed parameters.
1.1       anton    11859: 
1.78      anton    11860: The assembler includes all instruction of the Athlon, i.e. 486 core
                   11861: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
                   11862: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
                   11863: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
1.1       anton    11864: 
1.78      anton    11865: There are several prefixes to switch between different operation sizes,
                   11866: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
                   11867: double-word accesses. Addressing modes can be switched with @code{.wa}
                   11868: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
                   11869: need a prefix for byte register names (@code{AL} et al).
1.1       anton    11870: 
1.78      anton    11871: For floating point operations, the prefixes are @code{.fs} (IEEE
                   11872: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
                   11873: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
1.21      crook    11874: 
1.78      anton    11875: The MMX opcodes don't have size prefixes, they are spelled out like in
                   11876: the Intel assembler. Instead of move from and to memory, there are
                   11877: PLDQ/PLDD and PSTQ/PSTD.
1.21      crook    11878: 
1.78      anton    11879: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
                   11880: ax.  Immediate values are indicated by postfixing them with @code{#},
1.91    ! anton    11881: e.g., @code{3 #}.  Here are some examples of addressing modes in various
        !          11882: syntaxes:
1.21      crook    11883: 
1.26      crook    11884: @example
1.91    ! anton    11885: Gforth          Intel (NASM)   AT&T (gas)      Name
        !          11886: .w ax           ax             %ax             register (16 bit)
        !          11887: ax              eax            %eax            register (32 bit)
        !          11888: 3 #             offset 3       $3              immediate
        !          11889: 1000 #)         byte ptr 1000  1000            displacement
        !          11890: bx )            [ebx]          (%ebx)          base
        !          11891: 100 di d)       100[edi]       100(%edi)       base+displacement
        !          11892: 20 ax *4 i#)    20[eax*4]      20(,%eax,4)     (index*scale)+displacement
        !          11893: di ax *4 i)     [edi][eax*4]   (%edi,%eax,4)   base+(index*scale)
        !          11894: 4 bx cx di)     4[ebx][ecx]    4(%ebx,%ecx)    base+index+displacement
        !          11895: 12 sp ax *2 di) 12[esp][eax*2] 12(%esp,%eax,2) base+(index*scale)+displacement
        !          11896: @end example
        !          11897: 
        !          11898: You can use @code{L)} and @code{LI)} instead of @code{D)} and
        !          11899: @code{DI)} to enforce 32-bit displacement fields (useful for
        !          11900: later patching).
1.21      crook    11901: 
1.78      anton    11902: Some example of instructions are:
1.1       anton    11903: 
                   11904: @example
1.78      anton    11905: ax bx mov             \ move ebx,eax
                   11906: 3 # ax mov            \ mov eax,3
                   11907: 100 di ) ax mov       \ mov eax,100[edi]
                   11908: 4 bx cx di) ax mov    \ mov eax,4[ebx][ecx]
                   11909: .w ax bx mov          \ mov bx,ax
1.1       anton    11910: @end example
                   11911: 
1.78      anton    11912: The following forms are supported for binary instructions:
1.1       anton    11913: 
                   11914: @example
1.78      anton    11915: <reg> <reg> <inst>
                   11916: <n> # <reg> <inst>
                   11917: <mem> <reg> <inst>
                   11918: <reg> <mem> <inst>
1.1       anton    11919: @end example
                   11920: 
1.78      anton    11921: Immediate to memory is not supported.  The shift/rotate syntax is:
1.1       anton    11922: 
1.26      crook    11923: @example
1.78      anton    11924: <reg/mem> 1 # shl \ shortens to shift without immediate
                   11925: <reg/mem> 4 # shl
                   11926: <reg/mem> cl shl
1.26      crook    11927: @end example
1.1       anton    11928: 
1.78      anton    11929: Precede string instructions (@code{movs} etc.) with @code{.b} to get
                   11930: the byte version.
1.1       anton    11931: 
1.78      anton    11932: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
                   11933: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
                   11934: pc < >= <= >}. (Note that most of these words shadow some Forth words
                   11935: when @code{assembler} is in front of @code{forth} in the search path,
                   11936: e.g., in @code{code} words).  Currently the control structure words use
                   11937: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
                   11938: to shuffle them (you can also use @code{swap} etc.).
1.21      crook    11939: 
1.78      anton    11940: Here is an example of a @code{code} word (assumes that the stack pointer
                   11941: is in esi and the TOS is in ebx):
1.21      crook    11942: 
1.26      crook    11943: @example
1.78      anton    11944: code my+ ( n1 n2 -- n )
                   11945:     4 si D) bx add
                   11946:     4 # si add
                   11947:     Next
                   11948: end-code
1.26      crook    11949: @end example
1.21      crook    11950: 
1.78      anton    11951: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
                   11952: @subsection Alpha Assembler
1.21      crook    11953: 
1.78      anton    11954: The Alpha assembler and disassembler were originally written by Bernd
                   11955: Thallner.
1.26      crook    11956: 
1.78      anton    11957: The register names @code{a0}--@code{a5} are not available to avoid
                   11958: shadowing hex numbers.
1.2       jwilke   11959: 
1.78      anton    11960: Immediate forms of arithmetic instructions are distinguished by a
                   11961: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
                   11962: does not count as arithmetic instruction).
1.2       jwilke   11963: 
1.78      anton    11964: You have to specify all operands to an instruction, even those that
                   11965: other assemblers consider optional, e.g., the destination register for
                   11966: @code{br,}, or the destination register and hint for @code{jmp,}.
1.2       jwilke   11967: 
1.78      anton    11968: You can specify conditions for @code{if,} by removing the first @code{b}
                   11969: and the trailing @code{,} from a branch with a corresponding name; e.g.,
1.2       jwilke   11970: 
1.26      crook    11971: @example
1.78      anton    11972: 11 fgt if, \ if F11>0e
                   11973:   ...
                   11974: endif,
1.26      crook    11975: @end example
1.2       jwilke   11976: 
1.78      anton    11977: @code{fbgt,} gives @code{fgt}.  
                   11978: 
                   11979: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
                   11980: @subsection MIPS assembler
1.2       jwilke   11981: 
1.78      anton    11982: The MIPS assembler was originally written by Christian Pirker.
1.2       jwilke   11983: 
1.78      anton    11984: Currently the assembler and disassembler only cover the MIPS-I
                   11985: architecture (R3000), and don't support FP instructions.
1.2       jwilke   11986: 
1.78      anton    11987: The register names @code{$a0}--@code{$a3} are not available to avoid
                   11988: shadowing hex numbers.
1.2       jwilke   11989: 
1.78      anton    11990: Because there is no way to distinguish registers from immediate values,
                   11991: you have to explicitly use the immediate forms of instructions, i.e.,
                   11992: @code{addiu,}, not just @code{addu,} (@command{as} does this
                   11993: implicitly).
1.2       jwilke   11994: 
1.78      anton    11995: If the architecture manual specifies several formats for the instruction
                   11996: (e.g., for @code{jalr,}), you usually have to use the one with more
                   11997: arguments (i.e., two for @code{jalr,}).  When in doubt, see
                   11998: @code{arch/mips/testasm.fs} for an example of correct use.
1.2       jwilke   11999: 
1.78      anton    12000: Branches and jumps in the MIPS architecture have a delay slot.  You have
                   12001: to fill it yourself (the simplest way is to use @code{nop,}), the
                   12002: assembler does not do it for you (unlike @command{as}).  Even
                   12003: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
                   12004: @code{else,} and @code{repeat,} need a delay slot.  Since @code{begin,}
                   12005: and @code{then,} just specify branch targets, they are not affected.
1.2       jwilke   12006: 
1.78      anton    12007: Note that you must not put branches, jumps, or @code{li,} into the delay
                   12008: slot: @code{li,} may expand to several instructions, and control flow
                   12009: instructions may not be put into the branch delay slot in any case.
1.2       jwilke   12010: 
1.78      anton    12011: For branches the argument specifying the target is a relative address;
                   12012: You have to add the address of the delay slot to get the absolute
                   12013: address.
1.1       anton    12014: 
1.78      anton    12015: The MIPS architecture also has load delay slots and restrictions on
                   12016: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
                   12017: yourself to satisfy these restrictions, the assembler does not do it for
                   12018: you.
1.1       anton    12019: 
1.78      anton    12020: You can specify the conditions for @code{if,} etc. by taking a
                   12021: conditional branch and leaving away the @code{b} at the start and the
                   12022: @code{,} at the end.  E.g.,
1.1       anton    12023: 
1.26      crook    12024: @example
1.78      anton    12025: 4 5 eq if,
                   12026:   ... \ do something if $4 equals $5
                   12027: then,
1.26      crook    12028: @end example
1.1       anton    12029: 
1.78      anton    12030: @node Other assemblers,  , MIPS assembler, Assembler and Code Words
                   12031: @subsection Other assemblers
                   12032: 
                   12033: If you want to contribute another assembler/disassembler, please contact
                   12034: us (@email{bug-gforth@@gnu.org}) to check if we have such an assembler
                   12035: already.  If you are writing them from scratch, please use a similar
                   12036: syntax style as the one we use (i.e., postfix, commas at the end of the
                   12037: instruction names, @pxref{Common Assembler}); make the output of the
                   12038: disassembler be valid input for the assembler, and keep the style
                   12039: similar to the style we used.
                   12040: 
                   12041: Hints on implementation: The most important part is to have a good test
                   12042: suite that contains all instructions.  Once you have that, the rest is
                   12043: easy.  For actual coding you can take a look at
                   12044: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
                   12045: the assembler and disassembler, avoiding redundancy and some potential
                   12046: bugs.  You can also look at that file (and @pxref{Advanced does> usage
                   12047: example}) to get ideas how to factor a disassembler.
                   12048: 
                   12049: Start with the disassembler, because it's easier to reuse data from the
                   12050: disassembler for the assembler than the other way round.
1.1       anton    12051: 
1.78      anton    12052: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
                   12053: how simple it can be.
1.1       anton    12054: 
1.78      anton    12055: @c -------------------------------------------------------------
                   12056: @node Threading Words, Passing Commands to the OS, Assembler and Code Words, Words
                   12057: @section Threading Words
                   12058: @cindex threading words
1.1       anton    12059: 
1.78      anton    12060: @cindex code address
                   12061: These words provide access to code addresses and other threading stuff
                   12062: in Gforth (and, possibly, other interpretive Forths). It more or less
                   12063: abstracts away the differences between direct and indirect threading
                   12064: (and, for direct threading, the machine dependences). However, at
                   12065: present this wordset is still incomplete. It is also pretty low-level;
                   12066: some day it will hopefully be made unnecessary by an internals wordset
                   12067: that abstracts implementation details away completely.
1.1       anton    12068: 
1.78      anton    12069: The terminology used here stems from indirect threaded Forth systems; in
                   12070: such a system, the XT of a word is represented by the CFA (code field
                   12071: address) of a word; the CFA points to a cell that contains the code
                   12072: address.  The code address is the address of some machine code that
                   12073: performs the run-time action of invoking the word (e.g., the
                   12074: @code{dovar:} routine pushes the address of the body of the word (a
                   12075: variable) on the stack
                   12076: ).
1.1       anton    12077: 
1.78      anton    12078: @cindex code address
                   12079: @cindex code field address
                   12080: In an indirect threaded Forth, you can get the code address of @i{name}
                   12081: with @code{' @i{name} @@}; in Gforth you can get it with @code{' @i{name}
                   12082: >code-address}, independent of the threading method.
1.1       anton    12083: 
1.78      anton    12084: doc-threading-method
                   12085: doc->code-address
                   12086: doc-code-address!
1.1       anton    12087: 
1.78      anton    12088: @cindex @code{does>}-handler
                   12089: @cindex @code{does>}-code
                   12090: For a word defined with @code{DOES>}, the code address usually points to
                   12091: a jump instruction (the @dfn{does-handler}) that jumps to the dodoes
                   12092: routine (in Gforth on some platforms, it can also point to the dodoes
                   12093: routine itself).  What you are typically interested in, though, is
                   12094: whether a word is a @code{DOES>}-defined word, and what Forth code it
                   12095: executes; @code{>does-code} tells you that.
1.1       anton    12096: 
1.78      anton    12097: doc->does-code
1.1       anton    12098: 
1.78      anton    12099: To create a @code{DOES>}-defined word with the following basic words,
                   12100: you have to set up a @code{DOES>}-handler with @code{does-handler!};
                   12101: @code{/does-handler} aus behind you have to place your executable Forth
                   12102: code.  Finally you have to create a word and modify its behaviour with
                   12103: @code{does-handler!}.
1.1       anton    12104: 
1.78      anton    12105: doc-does-code!
                   12106: doc-does-handler!
                   12107: doc-/does-handler
1.1       anton    12108: 
1.78      anton    12109: The code addresses produced by various defining words are produced by
                   12110: the following words:
1.1       anton    12111: 
1.78      anton    12112: doc-docol:
                   12113: doc-docon:
                   12114: doc-dovar:
                   12115: doc-douser:
                   12116: doc-dodefer:
                   12117: doc-dofield:
1.1       anton    12118: 
1.26      crook    12119: @c -------------------------------------------------------------
1.78      anton    12120: @node Passing Commands to the OS, Keeping track of Time, Threading Words, Words
1.21      crook    12121: @section Passing Commands to the Operating System
                   12122: @cindex operating system - passing commands
                   12123: @cindex shell commands
                   12124: 
                   12125: Gforth allows you to pass an arbitrary string to the host operating
                   12126: system shell (if such a thing exists) for execution.
                   12127: 
1.44      crook    12128: 
1.21      crook    12129: doc-sh
                   12130: doc-system
                   12131: doc-$?
1.23      crook    12132: doc-getenv
1.21      crook    12133: 
1.44      crook    12134: 
1.26      crook    12135: @c -------------------------------------------------------------
1.47      crook    12136: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
                   12137: @section Keeping track of Time
                   12138: @cindex time-related words
                   12139: 
                   12140: doc-ms
                   12141: doc-time&date
1.79      anton    12142: doc-utime
                   12143: doc-cputime
1.47      crook    12144: 
                   12145: 
                   12146: @c -------------------------------------------------------------
                   12147: @node Miscellaneous Words,  , Keeping track of Time, Words
1.21      crook    12148: @section Miscellaneous Words
                   12149: @cindex miscellaneous words
                   12150: 
1.29      crook    12151: @comment TODO find homes for these
                   12152: 
1.26      crook    12153: These section lists the ANS Forth words that are not documented
1.21      crook    12154: elsewhere in this manual. Ultimately, they all need proper homes.
                   12155: 
1.68      anton    12156: doc-quit
1.44      crook    12157: 
1.26      crook    12158: The following ANS Forth words are not currently supported by Gforth 
1.27      crook    12159: (@pxref{ANS conformance}):
1.21      crook    12160: 
                   12161: @code{EDITOR} 
                   12162: @code{EMIT?} 
                   12163: @code{FORGET} 
                   12164: 
1.24      anton    12165: @c ******************************************************************
                   12166: @node Error messages, Tools, Words, Top
                   12167: @chapter Error messages
                   12168: @cindex error messages
                   12169: @cindex backtrace
                   12170: 
                   12171: A typical Gforth error message looks like this:
                   12172: 
                   12173: @example
1.86      anton    12174: in file included from \evaluated string/:-1
1.24      anton    12175: in file included from ./yyy.fs:1
                   12176: ./xxx.fs:4: Invalid memory address
                   12177: bar
                   12178: ^^^
1.79      anton    12179: Backtrace:
1.25      anton    12180: $400E664C @@
                   12181: $400E6664 foo
1.24      anton    12182: @end example
                   12183: 
                   12184: The message identifying the error is @code{Invalid memory address}.  The
                   12185: error happened when text-interpreting line 4 of the file
                   12186: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
                   12187: word on the line where the error happened, is pointed out (with
                   12188: @code{^^^}).
                   12189: 
                   12190: The file containing the error was included in line 1 of @file{./yyy.fs},
                   12191: and @file{yyy.fs} was included from a non-file (in this case, by giving
                   12192: @file{yyy.fs} as command-line parameter to Gforth).
                   12193: 
                   12194: At the end of the error message you find a return stack dump that can be
                   12195: interpreted as a backtrace (possibly empty). On top you find the top of
                   12196: the return stack when the @code{throw} happened, and at the bottom you
                   12197: find the return stack entry just above the return stack of the topmost
                   12198: text interpreter.
                   12199: 
                   12200: To the right of most return stack entries you see a guess for the word
                   12201: that pushed that return stack entry as its return address. This gives a
                   12202: backtrace. In our case we see that @code{bar} called @code{foo}, and
                   12203: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
                   12204: address} exception).
                   12205: 
                   12206: Note that the backtrace is not perfect: We don't know which return stack
                   12207: entries are return addresses (so we may get false positives); and in
                   12208: some cases (e.g., for @code{abort"}) we cannot determine from the return
                   12209: address the word that pushed the return address, so for some return
                   12210: addresses you see no names in the return stack dump.
1.25      anton    12211: 
                   12212: @cindex @code{catch} and backtraces
                   12213: The return stack dump represents the return stack at the time when a
                   12214: specific @code{throw} was executed.  In programs that make use of
                   12215: @code{catch}, it is not necessarily clear which @code{throw} should be
                   12216: used for the return stack dump (e.g., consider one @code{throw} that
                   12217: indicates an error, which is caught, and during recovery another error
1.42      anton    12218: happens; which @code{throw} should be used for the stack dump?).  Gforth
1.25      anton    12219: presents the return stack dump for the first @code{throw} after the last
                   12220: executed (not returned-to) @code{catch}; this works well in the usual
                   12221: case.
                   12222: 
                   12223: @cindex @code{gforth-fast} and backtraces
                   12224: @cindex @code{gforth-fast}, difference from @code{gforth}
                   12225: @cindex backtraces with @code{gforth-fast}
                   12226: @cindex return stack dump with @code{gforth-fast}
1.79      anton    12227: @code{Gforth} is able to do a return stack dump for throws generated
1.25      anton    12228: from primitives (e.g., invalid memory address, stack empty etc.);
                   12229: @code{gforth-fast} is only able to do a return stack dump from a
                   12230: directly called @code{throw} (including @code{abort} etc.).  This is the
1.30      anton    12231: only difference (apart from a speed factor of between 1.15 (K6-2) and
1.78      anton    12232: 2 (21264)) between @code{gforth} and @code{gforth-fast}.  Given an
1.30      anton    12233: exception caused by a primitive in @code{gforth-fast}, you will
                   12234: typically see no return stack dump at all; however, if the exception is
                   12235: caught by @code{catch} (e.g., for restoring some state), and then
                   12236: @code{throw}n again, the return stack dump will be for the first such
                   12237: @code{throw}.
1.2       jwilke   12238: 
1.5       anton    12239: @c ******************************************************************
1.24      anton    12240: @node Tools, ANS conformance, Error messages, Top
1.1       anton    12241: @chapter Tools
                   12242: 
                   12243: @menu
                   12244: * ANS Report::                  Report the words used, sorted by wordset.
                   12245: @end menu
                   12246: 
                   12247: See also @ref{Emacs and Gforth}.
                   12248: 
                   12249: @node ANS Report,  , Tools, Tools
                   12250: @section @file{ans-report.fs}: Report the words used, sorted by wordset
                   12251: @cindex @file{ans-report.fs}
                   12252: @cindex report the words used in your program
                   12253: @cindex words used in your program
                   12254: 
                   12255: If you want to label a Forth program as ANS Forth Program, you must
                   12256: document which wordsets the program uses; for extension wordsets, it is
                   12257: helpful to list the words the program requires from these wordsets
                   12258: (because Forth systems are allowed to provide only some words of them).
                   12259: 
                   12260: The @file{ans-report.fs} tool makes it easy for you to determine which
                   12261: words from which wordset and which non-ANS words your application
                   12262: uses. You simply have to include @file{ans-report.fs} before loading the
                   12263: program you want to check. After loading your program, you can get the
                   12264: report with @code{print-ans-report}. A typical use is to run this as
                   12265: batch job like this:
                   12266: @example
                   12267: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
                   12268: @end example
                   12269: 
                   12270: The output looks like this (for @file{compat/control.fs}):
                   12271: @example
                   12272: The program uses the following words
                   12273: from CORE :
                   12274: : POSTPONE THEN ; immediate ?dup IF 0= 
                   12275: from BLOCK-EXT :
                   12276: \ 
                   12277: from FILE :
                   12278: ( 
                   12279: @end example
                   12280: 
                   12281: @subsection Caveats
                   12282: 
                   12283: Note that @file{ans-report.fs} just checks which words are used, not whether
                   12284: they are used in an ANS Forth conforming way!
                   12285: 
                   12286: Some words are defined in several wordsets in the
                   12287: standard. @file{ans-report.fs} reports them for only one of the
                   12288: wordsets, and not necessarily the one you expect. It depends on usage
                   12289: which wordset is the right one to specify. E.g., if you only use the
                   12290: compilation semantics of @code{S"}, it is a Core word; if you also use
                   12291: its interpretation semantics, it is a File word.
                   12292: 
                   12293: @c ******************************************************************
1.65      anton    12294: @node ANS conformance, Standard vs Extensions, Tools, Top
1.1       anton    12295: @chapter ANS conformance
                   12296: @cindex ANS conformance of Gforth
                   12297: 
                   12298: To the best of our knowledge, Gforth is an
                   12299: 
                   12300: ANS Forth System
                   12301: @itemize @bullet
                   12302: @item providing the Core Extensions word set
                   12303: @item providing the Block word set
                   12304: @item providing the Block Extensions word set
                   12305: @item providing the Double-Number word set
                   12306: @item providing the Double-Number Extensions word set
                   12307: @item providing the Exception word set
                   12308: @item providing the Exception Extensions word set
                   12309: @item providing the Facility word set
1.40      anton    12310: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
1.1       anton    12311: @item providing the File Access word set
                   12312: @item providing the File Access Extensions word set
                   12313: @item providing the Floating-Point word set
                   12314: @item providing the Floating-Point Extensions word set
                   12315: @item providing the Locals word set
                   12316: @item providing the Locals Extensions word set
                   12317: @item providing the Memory-Allocation word set
                   12318: @item providing the Memory-Allocation Extensions word set (that one's easy)
                   12319: @item providing the Programming-Tools word set
                   12320: @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
                   12321: @item providing the Search-Order word set
                   12322: @item providing the Search-Order Extensions word set
                   12323: @item providing the String word set
                   12324: @item providing the String Extensions word set (another easy one)
                   12325: @end itemize
                   12326: 
                   12327: @cindex system documentation
                   12328: In addition, ANS Forth systems are required to document certain
                   12329: implementation choices. This chapter tries to meet these
                   12330: requirements. In many cases it gives a way to ask the system for the
                   12331: information instead of providing the information directly, in
                   12332: particular, if the information depends on the processor, the operating
                   12333: system or the installation options chosen, or if they are likely to
                   12334: change during the maintenance of Gforth.
                   12335: 
                   12336: @comment The framework for the rest has been taken from pfe.
                   12337: 
                   12338: @menu
                   12339: * The Core Words::              
                   12340: * The optional Block word set::  
                   12341: * The optional Double Number word set::  
                   12342: * The optional Exception word set::  
                   12343: * The optional Facility word set::  
                   12344: * The optional File-Access word set::  
                   12345: * The optional Floating-Point word set::  
                   12346: * The optional Locals word set::  
                   12347: * The optional Memory-Allocation word set::  
                   12348: * The optional Programming-Tools word set::  
                   12349: * The optional Search-Order word set::  
                   12350: @end menu
                   12351: 
                   12352: 
                   12353: @c =====================================================================
                   12354: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
                   12355: @comment  node-name,  next,  previous,  up
                   12356: @section The Core Words
                   12357: @c =====================================================================
                   12358: @cindex core words, system documentation
                   12359: @cindex system documentation, core words
                   12360: 
                   12361: @menu
                   12362: * core-idef::                   Implementation Defined Options                   
                   12363: * core-ambcond::                Ambiguous Conditions                
                   12364: * core-other::                  Other System Documentation                  
                   12365: @end menu
                   12366: 
                   12367: @c ---------------------------------------------------------------------
                   12368: @node core-idef, core-ambcond, The Core Words, The Core Words
                   12369: @subsection Implementation Defined Options
                   12370: @c ---------------------------------------------------------------------
                   12371: @cindex core words, implementation-defined options
                   12372: @cindex implementation-defined options, core words
                   12373: 
                   12374: 
                   12375: @table @i
                   12376: @item (Cell) aligned addresses:
                   12377: @cindex cell-aligned addresses
                   12378: @cindex aligned addresses
                   12379: processor-dependent. Gforth's alignment words perform natural alignment
                   12380: (e.g., an address aligned for a datum of size 8 is divisible by
                   12381: 8). Unaligned accesses usually result in a @code{-23 THROW}.
                   12382: 
                   12383: @item @code{EMIT} and non-graphic characters:
                   12384: @cindex @code{EMIT} and non-graphic characters
                   12385: @cindex non-graphic characters and @code{EMIT}
                   12386: The character is output using the C library function (actually, macro)
                   12387: @code{putc}.
                   12388: 
                   12389: @item character editing of @code{ACCEPT} and @code{EXPECT}:
                   12390: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
                   12391: @cindex editing in @code{ACCEPT} and @code{EXPECT}
                   12392: @cindex @code{ACCEPT}, editing
                   12393: @cindex @code{EXPECT}, editing
                   12394: This is modeled on the GNU readline library (@pxref{Readline
                   12395: Interaction, , Command Line Editing, readline, The GNU Readline
                   12396: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
                   12397: producing a full word completion every time you type it (instead of
1.28      crook    12398: producing the common prefix of all completions). @xref{Command-line editing}.
1.1       anton    12399: 
                   12400: @item character set:
                   12401: @cindex character set
                   12402: The character set of your computer and display device. Gforth is
                   12403: 8-bit-clean (but some other component in your system may make trouble).
                   12404: 
                   12405: @item Character-aligned address requirements:
                   12406: @cindex character-aligned address requirements
                   12407: installation-dependent. Currently a character is represented by a C
                   12408: @code{unsigned char}; in the future we might switch to @code{wchar_t}
                   12409: (Comments on that requested).
                   12410: 
                   12411: @item character-set extensions and matching of names:
                   12412: @cindex character-set extensions and matching of names
1.26      crook    12413: @cindex case-sensitivity for name lookup
                   12414: @cindex name lookup, case-sensitivity
                   12415: @cindex locale and case-sensitivity
1.21      crook    12416: Any character except the ASCII NUL character can be used in a
1.1       anton    12417: name. Matching is case-insensitive (except in @code{TABLE}s). The
1.47      crook    12418: matching is performed using the C library function @code{strncasecmp}, whose
1.1       anton    12419: function is probably influenced by the locale. E.g., the @code{C} locale
                   12420: does not know about accents and umlauts, so they are matched
                   12421: case-sensitively in that locale. For portability reasons it is best to
                   12422: write programs such that they work in the @code{C} locale. Then one can
                   12423: use libraries written by a Polish programmer (who might use words
                   12424: containing ISO Latin-2 encoded characters) and by a French programmer
                   12425: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
                   12426: funny results for some of the words (which ones, depends on the font you
                   12427: are using)). Also, the locale you prefer may not be available in other
                   12428: operating systems. Hopefully, Unicode will solve these problems one day.
                   12429: 
                   12430: @item conditions under which control characters match a space delimiter:
                   12431: @cindex space delimiters
                   12432: @cindex control characters as delimiters
                   12433: If @code{WORD} is called with the space character as a delimiter, all
                   12434: white-space characters (as identified by the C macro @code{isspace()})
                   12435: are delimiters. @code{PARSE}, on the other hand, treats space like other
1.44      crook    12436: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
1.79      anton    12437: like @code{PARSE} otherwise. @code{Name}, which is used by the outer
1.1       anton    12438: interpreter (aka text interpreter) by default, treats all white-space
                   12439: characters as delimiters.
                   12440: 
1.26      crook    12441: @item format of the control-flow stack:
                   12442: @cindex control-flow stack, format
                   12443: The data stack is used as control-flow stack. The size of a control-flow
1.1       anton    12444: stack item in cells is given by the constant @code{cs-item-size}. At the
                   12445: time of this writing, an item consists of a (pointer to a) locals list
                   12446: (third), an address in the code (second), and a tag for identifying the
                   12447: item (TOS). The following tags are used: @code{defstart},
                   12448: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
                   12449: @code{scopestart}.
                   12450: 
                   12451: @item conversion of digits > 35
                   12452: @cindex digits > 35
                   12453: The characters @code{[\]^_'} are the digits with the decimal value
                   12454: 36@minus{}41. There is no way to input many of the larger digits.
                   12455: 
                   12456: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
                   12457: @cindex @code{EXPECT}, display after end of input
                   12458: @cindex @code{ACCEPT}, display after end of input
                   12459: The cursor is moved to the end of the entered string. If the input is
                   12460: terminated using the @kbd{Return} key, a space is typed.
                   12461: 
                   12462: @item exception abort sequence of @code{ABORT"}:
                   12463: @cindex exception abort sequence of @code{ABORT"}
                   12464: @cindex @code{ABORT"}, exception abort sequence
                   12465: The error string is stored into the variable @code{"error} and a
                   12466: @code{-2 throw} is performed.
                   12467: 
                   12468: @item input line terminator:
                   12469: @cindex input line terminator
                   12470: @cindex line terminator on input
1.26      crook    12471: @cindex newline character on input
1.1       anton    12472: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
                   12473: lines. One of these characters is typically produced when you type the
                   12474: @kbd{Enter} or @kbd{Return} key.
                   12475: 
                   12476: @item maximum size of a counted string:
                   12477: @cindex maximum size of a counted string
                   12478: @cindex counted string, maximum size
                   12479: @code{s" /counted-string" environment? drop .}. Currently 255 characters
1.79      anton    12480: on all platforms, but this may change.
1.1       anton    12481: 
                   12482: @item maximum size of a parsed string:
                   12483: @cindex maximum size of a parsed string
                   12484: @cindex parsed string, maximum size
                   12485: Given by the constant @code{/line}. Currently 255 characters.
                   12486: 
                   12487: @item maximum size of a definition name, in characters:
                   12488: @cindex maximum size of a definition name, in characters
                   12489: @cindex name, maximum length
                   12490: 31
                   12491: 
                   12492: @item maximum string length for @code{ENVIRONMENT?}, in characters:
                   12493: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
                   12494: @cindex @code{ENVIRONMENT?} string length, maximum
                   12495: 31
                   12496: 
                   12497: @item method of selecting the user input device:
                   12498: @cindex user input device, method of selecting
                   12499: The user input device is the standard input. There is currently no way to
                   12500: change it from within Gforth. However, the input can typically be
                   12501: redirected in the command line that starts Gforth.
                   12502: 
                   12503: @item method of selecting the user output device:
                   12504: @cindex user output device, method of selecting
                   12505: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
1.10      anton    12506: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
                   12507: output when the user output device is a terminal, otherwise the output
                   12508: is buffered.
1.1       anton    12509: 
                   12510: @item methods of dictionary compilation:
                   12511: What are we expected to document here?
                   12512: 
                   12513: @item number of bits in one address unit:
                   12514: @cindex number of bits in one address unit
                   12515: @cindex address unit, size in bits
                   12516: @code{s" address-units-bits" environment? drop .}. 8 in all current
1.79      anton    12517: platforms.
1.1       anton    12518: 
                   12519: @item number representation and arithmetic:
                   12520: @cindex number representation and arithmetic
1.79      anton    12521: Processor-dependent. Binary two's complement on all current platforms.
1.1       anton    12522: 
                   12523: @item ranges for integer types:
                   12524: @cindex ranges for integer types
                   12525: @cindex integer types, ranges
                   12526: Installation-dependent. Make environmental queries for @code{MAX-N},
                   12527: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
                   12528: unsigned (and positive) types is 0. The lower bound for signed types on
                   12529: two's complement and one's complement machines machines can be computed
                   12530: by adding 1 to the upper bound.
                   12531: 
                   12532: @item read-only data space regions:
                   12533: @cindex read-only data space regions
                   12534: @cindex data-space, read-only regions
                   12535: The whole Forth data space is writable.
                   12536: 
                   12537: @item size of buffer at @code{WORD}:
                   12538: @cindex size of buffer at @code{WORD}
                   12539: @cindex @code{WORD} buffer size
                   12540: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
                   12541: shared with the pictured numeric output string. If overwriting
                   12542: @code{PAD} is acceptable, it is as large as the remaining dictionary
                   12543: space, although only as much can be sensibly used as fits in a counted
                   12544: string.
                   12545: 
                   12546: @item size of one cell in address units:
                   12547: @cindex cell size
                   12548: @code{1 cells .}.
                   12549: 
                   12550: @item size of one character in address units:
                   12551: @cindex char size
1.79      anton    12552: @code{1 chars .}. 1 on all current platforms.
1.1       anton    12553: 
                   12554: @item size of the keyboard terminal buffer:
                   12555: @cindex size of the keyboard terminal buffer
                   12556: @cindex terminal buffer, size
                   12557: Varies. You can determine the size at a specific time using @code{lp@@
                   12558: tib - .}. It is shared with the locals stack and TIBs of files that
                   12559: include the current file. You can change the amount of space for TIBs
                   12560: and locals stack at Gforth startup with the command line option
                   12561: @code{-l}.
                   12562: 
                   12563: @item size of the pictured numeric output buffer:
                   12564: @cindex size of the pictured numeric output buffer
                   12565: @cindex pictured numeric output buffer, size
                   12566: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
                   12567: shared with @code{WORD}.
                   12568: 
                   12569: @item size of the scratch area returned by @code{PAD}:
                   12570: @cindex size of the scratch area returned by @code{PAD}
                   12571: @cindex @code{PAD} size
                   12572: The remainder of dictionary space. @code{unused pad here - - .}.
                   12573: 
                   12574: @item system case-sensitivity characteristics:
                   12575: @cindex case-sensitivity characteristics
1.26      crook    12576: Dictionary searches are case-insensitive (except in
1.1       anton    12577: @code{TABLE}s). However, as explained above under @i{character-set
                   12578: extensions}, the matching for non-ASCII characters is determined by the
                   12579: locale you are using. In the default @code{C} locale all non-ASCII
                   12580: characters are matched case-sensitively.
                   12581: 
                   12582: @item system prompt:
                   12583: @cindex system prompt
                   12584: @cindex prompt
                   12585: @code{ ok} in interpret state, @code{ compiled} in compile state.
                   12586: 
                   12587: @item division rounding:
                   12588: @cindex division rounding
                   12589: installation dependent. @code{s" floored" environment? drop .}. We leave
                   12590: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
                   12591: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
                   12592: 
                   12593: @item values of @code{STATE} when true:
                   12594: @cindex @code{STATE} values
                   12595: -1.
                   12596: 
                   12597: @item values returned after arithmetic overflow:
                   12598: On two's complement machines, arithmetic is performed modulo
                   12599: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
                   12600: arithmetic (with appropriate mapping for signed types). Division by zero
                   12601: typically results in a @code{-55 throw} (Floating-point unidentified
1.80      anton    12602: fault) or @code{-10 throw} (divide by zero).
1.1       anton    12603: 
                   12604: @item whether the current definition can be found after @t{DOES>}:
                   12605: @cindex @t{DOES>}, visibility of current definition
                   12606: No.
                   12607: 
                   12608: @end table
                   12609: 
                   12610: @c ---------------------------------------------------------------------
                   12611: @node core-ambcond, core-other, core-idef, The Core Words
                   12612: @subsection Ambiguous conditions
                   12613: @c ---------------------------------------------------------------------
                   12614: @cindex core words, ambiguous conditions
                   12615: @cindex ambiguous conditions, core words
                   12616: 
                   12617: @table @i
                   12618: 
                   12619: @item a name is neither a word nor a number:
                   12620: @cindex name not found
1.26      crook    12621: @cindex undefined word
1.80      anton    12622: @code{-13 throw} (Undefined word).
1.1       anton    12623: 
                   12624: @item a definition name exceeds the maximum length allowed:
1.26      crook    12625: @cindex word name too long
1.1       anton    12626: @code{-19 throw} (Word name too long)
                   12627: 
                   12628: @item addressing a region not inside the various data spaces of the forth system:
                   12629: @cindex Invalid memory address
1.32      anton    12630: The stacks, code space and header space are accessible. Machine code space is
1.1       anton    12631: typically readable. Accessing other addresses gives results dependent on
                   12632: the operating system. On decent systems: @code{-9 throw} (Invalid memory
                   12633: address).
                   12634: 
                   12635: @item argument type incompatible with parameter:
1.26      crook    12636: @cindex argument type mismatch
1.1       anton    12637: This is usually not caught. Some words perform checks, e.g., the control
                   12638: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
                   12639: mismatch).
                   12640: 
                   12641: @item attempting to obtain the execution token of a word with undefined execution semantics:
                   12642: @cindex Interpreting a compile-only word, for @code{'} etc.
                   12643: @cindex execution token of words with undefined execution semantics
                   12644: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
                   12645: get an execution token for @code{compile-only-error} (which performs a
                   12646: @code{-14 throw} when executed).
                   12647: 
                   12648: @item dividing by zero:
                   12649: @cindex dividing by zero
                   12650: @cindex floating point unidentified fault, integer division
1.80      anton    12651: On some platforms, this produces a @code{-10 throw} (Division by
1.24      anton    12652: zero); on other systems, this typically results in a @code{-55 throw}
                   12653: (Floating-point unidentified fault).
1.1       anton    12654: 
                   12655: @item insufficient data stack or return stack space:
                   12656: @cindex insufficient data stack or return stack space
                   12657: @cindex stack overflow
1.26      crook    12658: @cindex address alignment exception, stack overflow
1.1       anton    12659: @cindex Invalid memory address, stack overflow
                   12660: Depending on the operating system, the installation, and the invocation
                   12661: of Gforth, this is either checked by the memory management hardware, or
1.24      anton    12662: it is not checked. If it is checked, you typically get a @code{-3 throw}
                   12663: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
                   12664: throw} (Invalid memory address) (depending on the platform and how you
                   12665: achieved the overflow) as soon as the overflow happens. If it is not
                   12666: checked, overflows typically result in mysterious illegal memory
                   12667: accesses, producing @code{-9 throw} (Invalid memory address) or
                   12668: @code{-23 throw} (Address alignment exception); they might also destroy
                   12669: the internal data structure of @code{ALLOCATE} and friends, resulting in
                   12670: various errors in these words.
1.1       anton    12671: 
                   12672: @item insufficient space for loop control parameters:
                   12673: @cindex insufficient space for loop control parameters
1.80      anton    12674: Like other return stack overflows.
1.1       anton    12675: 
                   12676: @item insufficient space in the dictionary:
                   12677: @cindex insufficient space in the dictionary
                   12678: @cindex dictionary overflow
1.12      anton    12679: If you try to allot (either directly with @code{allot}, or indirectly
                   12680: with @code{,}, @code{create} etc.) more memory than available in the
                   12681: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
                   12682: to access memory beyond the end of the dictionary, the results are
                   12683: similar to stack overflows.
1.1       anton    12684: 
                   12685: @item interpreting a word with undefined interpretation semantics:
                   12686: @cindex interpreting a word with undefined interpretation semantics
                   12687: @cindex Interpreting a compile-only word
                   12688: For some words, we have defined interpretation semantics. For the
                   12689: others: @code{-14 throw} (Interpreting a compile-only word).
                   12690: 
                   12691: @item modifying the contents of the input buffer or a string literal:
                   12692: @cindex modifying the contents of the input buffer or a string literal
                   12693: These are located in writable memory and can be modified.
                   12694: 
                   12695: @item overflow of the pictured numeric output string:
                   12696: @cindex overflow of the pictured numeric output string
                   12697: @cindex pictured numeric output string, overflow
1.24      anton    12698: @code{-17 throw} (Pictured numeric ouput string overflow).
1.1       anton    12699: 
                   12700: @item parsed string overflow:
                   12701: @cindex parsed string overflow
                   12702: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
                   12703: 
                   12704: @item producing a result out of range:
                   12705: @cindex result out of range
                   12706: On two's complement machines, arithmetic is performed modulo
                   12707: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
                   12708: arithmetic (with appropriate mapping for signed types). Division by zero
1.24      anton    12709: typically results in a @code{-10 throw} (divide by zero) or @code{-55
                   12710: throw} (floating point unidentified fault). @code{convert} and
                   12711: @code{>number} currently overflow silently.
1.1       anton    12712: 
                   12713: @item reading from an empty data or return stack:
                   12714: @cindex stack empty
                   12715: @cindex stack underflow
1.24      anton    12716: @cindex return stack underflow
1.1       anton    12717: The data stack is checked by the outer (aka text) interpreter after
                   12718: every word executed. If it has underflowed, a @code{-4 throw} (Stack
                   12719: underflow) is performed. Apart from that, stacks may be checked or not,
1.24      anton    12720: depending on operating system, installation, and invocation. If they are
                   12721: caught by a check, they typically result in @code{-4 throw} (Stack
                   12722: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
                   12723: (Invalid memory address), depending on the platform and which stack
                   12724: underflows and by how much. Note that even if the system uses checking
                   12725: (through the MMU), your program may have to underflow by a significant
                   12726: number of stack items to trigger the reaction (the reason for this is
                   12727: that the MMU, and therefore the checking, works with a page-size
                   12728: granularity).  If there is no checking, the symptoms resulting from an
                   12729: underflow are similar to those from an overflow.  Unbalanced return
1.80      anton    12730: stack errors can result in a variety of symptoms, including @code{-9 throw}
1.24      anton    12731: (Invalid memory address) and Illegal Instruction (typically @code{-260
                   12732: throw}).
1.1       anton    12733: 
                   12734: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
                   12735: @cindex unexpected end of the input buffer
                   12736: @cindex zero-length string as a name
                   12737: @cindex Attempt to use zero-length string as a name
                   12738: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
                   12739: use zero-length string as a name). Words like @code{'} probably will not
                   12740: find what they search. Note that it is possible to create zero-length
                   12741: names with @code{nextname} (should it not?).
                   12742: 
                   12743: @item @code{>IN} greater than input buffer:
                   12744: @cindex @code{>IN} greater than input buffer
                   12745: The next invocation of a parsing word returns a string with length 0.
                   12746: 
                   12747: @item @code{RECURSE} appears after @code{DOES>}:
                   12748: @cindex @code{RECURSE} appears after @code{DOES>}
                   12749: Compiles a recursive call to the defining word, not to the defined word.
                   12750: 
                   12751: @item argument input source different than current input source for @code{RESTORE-INPUT}:
                   12752: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
1.26      crook    12753: @cindex argument type mismatch, @code{RESTORE-INPUT}
1.1       anton    12754: @cindex @code{RESTORE-INPUT}, Argument type mismatch
                   12755: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
                   12756: the end of the file was reached), its source-id may be
                   12757: reused. Therefore, restoring an input source specification referencing a
                   12758: closed file may lead to unpredictable results instead of a @code{-12
                   12759: THROW}.
                   12760: 
                   12761: In the future, Gforth may be able to restore input source specifications
                   12762: from other than the current input source.
                   12763: 
                   12764: @item data space containing definitions gets de-allocated:
                   12765: @cindex data space containing definitions gets de-allocated
                   12766: Deallocation with @code{allot} is not checked. This typically results in
                   12767: memory access faults or execution of illegal instructions.
                   12768: 
                   12769: @item data space read/write with incorrect alignment:
                   12770: @cindex data space read/write with incorrect alignment
                   12771: @cindex alignment faults
1.26      crook    12772: @cindex address alignment exception
1.1       anton    12773: Processor-dependent. Typically results in a @code{-23 throw} (Address
1.12      anton    12774: alignment exception). Under Linux-Intel on a 486 or later processor with
1.1       anton    12775: alignment turned on, incorrect alignment results in a @code{-9 throw}
                   12776: (Invalid memory address). There are reportedly some processors with
1.12      anton    12777: alignment restrictions that do not report violations.
1.1       anton    12778: 
                   12779: @item data space pointer not properly aligned, @code{,}, @code{C,}:
                   12780: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
                   12781: Like other alignment errors.
                   12782: 
                   12783: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
                   12784: Like other stack underflows.
                   12785: 
                   12786: @item loop control parameters not available:
                   12787: @cindex loop control parameters not available
                   12788: Not checked. The counted loop words simply assume that the top of return
                   12789: stack items are loop control parameters and behave accordingly.
                   12790: 
                   12791: @item most recent definition does not have a name (@code{IMMEDIATE}):
                   12792: @cindex most recent definition does not have a name (@code{IMMEDIATE})
                   12793: @cindex last word was headerless
                   12794: @code{abort" last word was headerless"}.
                   12795: 
                   12796: @item name not defined by @code{VALUE} used by @code{TO}:
                   12797: @cindex name not defined by @code{VALUE} used by @code{TO}
                   12798: @cindex @code{TO} on non-@code{VALUE}s
                   12799: @cindex Invalid name argument, @code{TO}
                   12800: @code{-32 throw} (Invalid name argument) (unless name is a local or was
                   12801: defined by @code{CONSTANT}; in the latter case it just changes the constant).
                   12802: 
                   12803: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
                   12804: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
1.26      crook    12805: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
1.1       anton    12806: @code{-13 throw} (Undefined word)
                   12807: 
                   12808: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
                   12809: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
                   12810: Gforth behaves as if they were of the same type. I.e., you can predict
                   12811: the behaviour by interpreting all parameters as, e.g., signed.
                   12812: 
                   12813: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
                   12814: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
                   12815: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
                   12816: compilation semantics of @code{TO}.
                   12817: 
                   12818: @item String longer than a counted string returned by @code{WORD}:
1.26      crook    12819: @cindex string longer than a counted string returned by @code{WORD}
1.1       anton    12820: @cindex @code{WORD}, string overflow
                   12821: Not checked. The string will be ok, but the count will, of course,
                   12822: contain only the least significant bits of the length.
                   12823: 
                   12824: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
                   12825: @cindex @code{LSHIFT}, large shift counts
                   12826: @cindex @code{RSHIFT}, large shift counts
                   12827: Processor-dependent. Typical behaviours are returning 0 and using only
                   12828: the low bits of the shift count.
                   12829: 
                   12830: @item word not defined via @code{CREATE}:
                   12831: @cindex @code{>BODY} of non-@code{CREATE}d words
                   12832: @code{>BODY} produces the PFA of the word no matter how it was defined.
                   12833: 
                   12834: @cindex @code{DOES>} of non-@code{CREATE}d words
                   12835: @code{DOES>} changes the execution semantics of the last defined word no
                   12836: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
                   12837: @code{CREATE , DOES>}.
                   12838: 
                   12839: @item words improperly used outside @code{<#} and @code{#>}:
                   12840: Not checked. As usual, you can expect memory faults.
                   12841: 
                   12842: @end table
                   12843: 
                   12844: 
                   12845: @c ---------------------------------------------------------------------
                   12846: @node core-other,  , core-ambcond, The Core Words
                   12847: @subsection Other system documentation
                   12848: @c ---------------------------------------------------------------------
                   12849: @cindex other system documentation, core words
                   12850: @cindex core words, other system documentation
                   12851: 
                   12852: @table @i
                   12853: @item nonstandard words using @code{PAD}:
                   12854: @cindex @code{PAD} use by nonstandard words
                   12855: None.
                   12856: 
                   12857: @item operator's terminal facilities available:
                   12858: @cindex operator's terminal facilities available
1.80      anton    12859: After processing the OS's command line, Gforth goes into interactive mode,
1.1       anton    12860: and you can give commands to Gforth interactively. The actual facilities
                   12861: available depend on how you invoke Gforth.
                   12862: 
                   12863: @item program data space available:
                   12864: @cindex program data space available
                   12865: @cindex data space available
                   12866: @code{UNUSED .} gives the remaining dictionary space. The total
                   12867: dictionary space can be specified with the @code{-m} switch
                   12868: (@pxref{Invoking Gforth}) when Gforth starts up.
                   12869: 
                   12870: @item return stack space available:
                   12871: @cindex return stack space available
                   12872: You can compute the total return stack space in cells with
                   12873: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
                   12874: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
                   12875: 
                   12876: @item stack space available:
                   12877: @cindex stack space available
                   12878: You can compute the total data stack space in cells with
                   12879: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
                   12880: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
                   12881: 
                   12882: @item system dictionary space required, in address units:
                   12883: @cindex system dictionary space required, in address units
                   12884: Type @code{here forthstart - .} after startup. At the time of this
                   12885: writing, this gives 80080 (bytes) on a 32-bit system.
                   12886: @end table
                   12887: 
                   12888: 
                   12889: @c =====================================================================
                   12890: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
                   12891: @section The optional Block word set
                   12892: @c =====================================================================
                   12893: @cindex system documentation, block words
                   12894: @cindex block words, system documentation
                   12895: 
                   12896: @menu
                   12897: * block-idef::                  Implementation Defined Options
                   12898: * block-ambcond::               Ambiguous Conditions               
                   12899: * block-other::                 Other System Documentation                 
                   12900: @end menu
                   12901: 
                   12902: 
                   12903: @c ---------------------------------------------------------------------
                   12904: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
                   12905: @subsection Implementation Defined Options
                   12906: @c ---------------------------------------------------------------------
                   12907: @cindex implementation-defined options, block words
                   12908: @cindex block words, implementation-defined options
                   12909: 
                   12910: @table @i
                   12911: @item the format for display by @code{LIST}:
                   12912: @cindex @code{LIST} display format
                   12913: First the screen number is displayed, then 16 lines of 64 characters,
                   12914: each line preceded by the line number.
                   12915: 
                   12916: @item the length of a line affected by @code{\}:
                   12917: @cindex length of a line affected by @code{\}
                   12918: @cindex @code{\}, line length in blocks
                   12919: 64 characters.
                   12920: @end table
                   12921: 
                   12922: 
                   12923: @c ---------------------------------------------------------------------
                   12924: @node block-ambcond, block-other, block-idef, The optional Block word set
                   12925: @subsection Ambiguous conditions
                   12926: @c ---------------------------------------------------------------------
                   12927: @cindex block words, ambiguous conditions
                   12928: @cindex ambiguous conditions, block words
                   12929: 
                   12930: @table @i
                   12931: @item correct block read was not possible:
                   12932: @cindex block read not possible
                   12933: Typically results in a @code{throw} of some OS-derived value (between
                   12934: -512 and -2048). If the blocks file was just not long enough, blanks are
                   12935: supplied for the missing portion.
                   12936: 
                   12937: @item I/O exception in block transfer:
                   12938: @cindex I/O exception in block transfer
                   12939: @cindex block transfer, I/O exception
                   12940: Typically results in a @code{throw} of some OS-derived value (between
                   12941: -512 and -2048).
                   12942: 
                   12943: @item invalid block number:
                   12944: @cindex invalid block number
                   12945: @cindex block number invalid
                   12946: @code{-35 throw} (Invalid block number)
                   12947: 
                   12948: @item a program directly alters the contents of @code{BLK}:
                   12949: @cindex @code{BLK}, altering @code{BLK}
                   12950: The input stream is switched to that other block, at the same
                   12951: position. If the storing to @code{BLK} happens when interpreting
                   12952: non-block input, the system will get quite confused when the block ends.
                   12953: 
                   12954: @item no current block buffer for @code{UPDATE}:
                   12955: @cindex @code{UPDATE}, no current block buffer
                   12956: @code{UPDATE} has no effect.
                   12957: 
                   12958: @end table
                   12959: 
                   12960: @c ---------------------------------------------------------------------
                   12961: @node block-other,  , block-ambcond, The optional Block word set
                   12962: @subsection Other system documentation
                   12963: @c ---------------------------------------------------------------------
                   12964: @cindex other system documentation, block words
                   12965: @cindex block words, other system documentation
                   12966: 
                   12967: @table @i
                   12968: @item any restrictions a multiprogramming system places on the use of buffer addresses:
                   12969: No restrictions (yet).
                   12970: 
                   12971: @item the number of blocks available for source and data:
                   12972: depends on your disk space.
                   12973: 
                   12974: @end table
                   12975: 
                   12976: 
                   12977: @c =====================================================================
                   12978: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
                   12979: @section The optional Double Number word set
                   12980: @c =====================================================================
                   12981: @cindex system documentation, double words
                   12982: @cindex double words, system documentation
                   12983: 
                   12984: @menu
                   12985: * double-ambcond::              Ambiguous Conditions              
                   12986: @end menu
                   12987: 
                   12988: 
                   12989: @c ---------------------------------------------------------------------
                   12990: @node double-ambcond,  , The optional Double Number word set, The optional Double Number word set
                   12991: @subsection Ambiguous conditions
                   12992: @c ---------------------------------------------------------------------
                   12993: @cindex double words, ambiguous conditions
                   12994: @cindex ambiguous conditions, double words
                   12995: 
                   12996: @table @i
1.29      crook    12997: @item @i{d} outside of range of @i{n} in @code{D>S}:
                   12998: @cindex @code{D>S}, @i{d} out of range of @i{n} 
                   12999: The least significant cell of @i{d} is produced.
1.1       anton    13000: 
                   13001: @end table
                   13002: 
                   13003: 
                   13004: @c =====================================================================
                   13005: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
                   13006: @section The optional Exception word set
                   13007: @c =====================================================================
                   13008: @cindex system documentation, exception words
                   13009: @cindex exception words, system documentation
                   13010: 
                   13011: @menu
                   13012: * exception-idef::              Implementation Defined Options              
                   13013: @end menu
                   13014: 
                   13015: 
                   13016: @c ---------------------------------------------------------------------
                   13017: @node exception-idef,  , The optional Exception word set, The optional Exception word set
                   13018: @subsection Implementation Defined Options
                   13019: @c ---------------------------------------------------------------------
                   13020: @cindex implementation-defined options, exception words
                   13021: @cindex exception words, implementation-defined options
                   13022: 
                   13023: @table @i
                   13024: @item @code{THROW}-codes used in the system:
                   13025: @cindex @code{THROW}-codes used in the system
                   13026: The codes -256@minus{}-511 are used for reporting signals. The mapping
1.29      crook    13027: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
1.1       anton    13028: codes -512@minus{}-2047 are used for OS errors (for file and memory
                   13029: allocation operations). The mapping from OS error numbers to throw codes
                   13030: is -512@minus{}@code{errno}. One side effect of this mapping is that
                   13031: undefined OS errors produce a message with a strange number; e.g.,
                   13032: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
                   13033: @end table
                   13034: 
                   13035: @c =====================================================================
                   13036: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
                   13037: @section The optional Facility word set
                   13038: @c =====================================================================
                   13039: @cindex system documentation, facility words
                   13040: @cindex facility words, system documentation
                   13041: 
                   13042: @menu
                   13043: * facility-idef::               Implementation Defined Options               
                   13044: * facility-ambcond::            Ambiguous Conditions            
                   13045: @end menu
                   13046: 
                   13047: 
                   13048: @c ---------------------------------------------------------------------
                   13049: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
                   13050: @subsection Implementation Defined Options
                   13051: @c ---------------------------------------------------------------------
                   13052: @cindex implementation-defined options, facility words
                   13053: @cindex facility words, implementation-defined options
                   13054: 
                   13055: @table @i
                   13056: @item encoding of keyboard events (@code{EKEY}):
                   13057: @cindex keyboard events, encoding in @code{EKEY}
                   13058: @cindex @code{EKEY}, encoding of keyboard events
1.40      anton    13059: Keys corresponding to ASCII characters are encoded as ASCII characters.
1.41      anton    13060: Other keys are encoded with the constants @code{k-left}, @code{k-right},
                   13061: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
                   13062: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
                   13063: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
1.40      anton    13064: 
1.1       anton    13065: 
                   13066: @item duration of a system clock tick:
                   13067: @cindex duration of a system clock tick
                   13068: @cindex clock tick duration
                   13069: System dependent. With respect to @code{MS}, the time is specified in
                   13070: microseconds. How well the OS and the hardware implement this, is
                   13071: another question.
                   13072: 
                   13073: @item repeatability to be expected from the execution of @code{MS}:
                   13074: @cindex repeatability to be expected from the execution of @code{MS}
                   13075: @cindex @code{MS}, repeatability to be expected
                   13076: System dependent. On Unix, a lot depends on load. If the system is
                   13077: lightly loaded, and the delay is short enough that Gforth does not get
                   13078: swapped out, the performance should be acceptable. Under MS-DOS and
                   13079: other single-tasking systems, it should be good.
                   13080: 
                   13081: @end table
                   13082: 
                   13083: 
                   13084: @c ---------------------------------------------------------------------
                   13085: @node facility-ambcond,  , facility-idef, The optional Facility word set
                   13086: @subsection Ambiguous conditions
                   13087: @c ---------------------------------------------------------------------
                   13088: @cindex facility words, ambiguous conditions
                   13089: @cindex ambiguous conditions, facility words
                   13090: 
                   13091: @table @i
                   13092: @item @code{AT-XY} can't be performed on user output device:
                   13093: @cindex @code{AT-XY} can't be performed on user output device
                   13094: Largely terminal dependent. No range checks are done on the arguments.
                   13095: No errors are reported. You may see some garbage appearing, you may see
                   13096: simply nothing happen.
                   13097: 
                   13098: @end table
                   13099: 
                   13100: 
                   13101: @c =====================================================================
                   13102: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
                   13103: @section The optional File-Access word set
                   13104: @c =====================================================================
                   13105: @cindex system documentation, file words
                   13106: @cindex file words, system documentation
                   13107: 
                   13108: @menu
                   13109: * file-idef::                   Implementation Defined Options
                   13110: * file-ambcond::                Ambiguous Conditions                
                   13111: @end menu
                   13112: 
                   13113: @c ---------------------------------------------------------------------
                   13114: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
                   13115: @subsection Implementation Defined Options
                   13116: @c ---------------------------------------------------------------------
                   13117: @cindex implementation-defined options, file words
                   13118: @cindex file words, implementation-defined options
                   13119: 
                   13120: @table @i
                   13121: @item file access methods used:
                   13122: @cindex file access methods used
                   13123: @code{R/O}, @code{R/W} and @code{BIN} work as you would
                   13124: expect. @code{W/O} translates into the C file opening mode @code{w} (or
                   13125: @code{wb}): The file is cleared, if it exists, and created, if it does
                   13126: not (with both @code{open-file} and @code{create-file}).  Under Unix
                   13127: @code{create-file} creates a file with 666 permissions modified by your
                   13128: umask.
                   13129: 
                   13130: @item file exceptions:
                   13131: @cindex file exceptions
                   13132: The file words do not raise exceptions (except, perhaps, memory access
                   13133: faults when you pass illegal addresses or file-ids).
                   13134: 
                   13135: @item file line terminator:
                   13136: @cindex file line terminator
                   13137: System-dependent. Gforth uses C's newline character as line
                   13138: terminator. What the actual character code(s) of this are is
                   13139: system-dependent.
                   13140: 
                   13141: @item file name format:
                   13142: @cindex file name format
                   13143: System dependent. Gforth just uses the file name format of your OS.
                   13144: 
                   13145: @item information returned by @code{FILE-STATUS}:
                   13146: @cindex @code{FILE-STATUS}, returned information
                   13147: @code{FILE-STATUS} returns the most powerful file access mode allowed
                   13148: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
                   13149: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
                   13150: along with the returned mode.
                   13151: 
                   13152: @item input file state after an exception when including source:
                   13153: @cindex exception when including source
                   13154: All files that are left via the exception are closed.
                   13155: 
1.29      crook    13156: @item @i{ior} values and meaning:
                   13157: @cindex @i{ior} values and meaning
1.68      anton    13158: @cindex @i{wior} values and meaning
1.29      crook    13159: The @i{ior}s returned by the file and memory allocation words are
1.1       anton    13160: intended as throw codes. They typically are in the range
                   13161: -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
1.29      crook    13162: @i{ior}s is -512@minus{}@i{errno}.
1.1       anton    13163: 
                   13164: @item maximum depth of file input nesting:
                   13165: @cindex maximum depth of file input nesting
                   13166: @cindex file input nesting, maximum depth
                   13167: limited by the amount of return stack, locals/TIB stack, and the number
                   13168: of open files available. This should not give you troubles.
                   13169: 
                   13170: @item maximum size of input line:
                   13171: @cindex maximum size of input line
                   13172: @cindex input line size, maximum
                   13173: @code{/line}. Currently 255.
                   13174: 
                   13175: @item methods of mapping block ranges to files:
                   13176: @cindex mapping block ranges to files
                   13177: @cindex files containing blocks
                   13178: @cindex blocks in files
                   13179: By default, blocks are accessed in the file @file{blocks.fb} in the
                   13180: current working directory. The file can be switched with @code{USE}.
                   13181: 
                   13182: @item number of string buffers provided by @code{S"}:
                   13183: @cindex @code{S"}, number of string buffers
                   13184: 1
                   13185: 
                   13186: @item size of string buffer used by @code{S"}:
                   13187: @cindex @code{S"}, size of string buffer
                   13188: @code{/line}. currently 255.
                   13189: 
                   13190: @end table
                   13191: 
                   13192: @c ---------------------------------------------------------------------
                   13193: @node file-ambcond,  , file-idef, The optional File-Access word set
                   13194: @subsection Ambiguous conditions
                   13195: @c ---------------------------------------------------------------------
                   13196: @cindex file words, ambiguous conditions
                   13197: @cindex ambiguous conditions, file words
                   13198: 
                   13199: @table @i
                   13200: @item attempting to position a file outside its boundaries:
                   13201: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
                   13202: @code{REPOSITION-FILE} is performed as usual: Afterwards,
                   13203: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
                   13204: 
                   13205: @item attempting to read from file positions not yet written:
                   13206: @cindex reading from file positions not yet written
                   13207: End-of-file, i.e., zero characters are read and no error is reported.
                   13208: 
1.29      crook    13209: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
                   13210: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid 
1.1       anton    13211: An appropriate exception may be thrown, but a memory fault or other
                   13212: problem is more probable.
                   13213: 
1.29      crook    13214: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
                   13215: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
                   13216: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
                   13217: The @i{ior} produced by the operation, that discovered the problem, is
1.1       anton    13218: thrown.
                   13219: 
                   13220: @item named file cannot be opened (@code{INCLUDED}):
                   13221: @cindex @code{INCLUDED}, named file cannot be opened
1.29      crook    13222: The @i{ior} produced by @code{open-file} is thrown.
1.1       anton    13223: 
                   13224: @item requesting an unmapped block number:
                   13225: @cindex unmapped block numbers
                   13226: There are no unmapped legal block numbers. On some operating systems,
                   13227: writing a block with a large number may overflow the file system and
                   13228: have an error message as consequence.
                   13229: 
                   13230: @item using @code{source-id} when @code{blk} is non-zero:
                   13231: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
                   13232: @code{source-id} performs its function. Typically it will give the id of
                   13233: the source which loaded the block. (Better ideas?)
                   13234: 
                   13235: @end table
                   13236: 
                   13237: 
                   13238: @c =====================================================================
                   13239: @node  The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
                   13240: @section The optional Floating-Point word set
                   13241: @c =====================================================================
                   13242: @cindex system documentation, floating-point words
                   13243: @cindex floating-point words, system documentation
                   13244: 
                   13245: @menu
                   13246: * floating-idef::               Implementation Defined Options
                   13247: * floating-ambcond::            Ambiguous Conditions            
                   13248: @end menu
                   13249: 
                   13250: 
                   13251: @c ---------------------------------------------------------------------
                   13252: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
                   13253: @subsection Implementation Defined Options
                   13254: @c ---------------------------------------------------------------------
                   13255: @cindex implementation-defined options, floating-point words
                   13256: @cindex floating-point words, implementation-defined options
                   13257: 
                   13258: @table @i
                   13259: @item format and range of floating point numbers:
                   13260: @cindex format and range of floating point numbers
                   13261: @cindex floating point numbers, format and range
                   13262: System-dependent; the @code{double} type of C.
                   13263: 
1.29      crook    13264: @item results of @code{REPRESENT} when @i{float} is out of range:
                   13265: @cindex  @code{REPRESENT}, results when @i{float} is out of range
1.1       anton    13266: System dependent; @code{REPRESENT} is implemented using the C library
                   13267: function @code{ecvt()} and inherits its behaviour in this respect.
                   13268: 
                   13269: @item rounding or truncation of floating-point numbers:
                   13270: @cindex rounding of floating-point numbers
                   13271: @cindex truncation of floating-point numbers
                   13272: @cindex floating-point numbers, rounding or truncation
                   13273: System dependent; the rounding behaviour is inherited from the hosting C
                   13274: compiler. IEEE-FP-based (i.e., most) systems by default round to
                   13275: nearest, and break ties by rounding to even (i.e., such that the last
                   13276: bit of the mantissa is 0).
                   13277: 
                   13278: @item size of floating-point stack:
                   13279: @cindex floating-point stack size
                   13280: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
                   13281: the floating-point stack (in floats). You can specify this on startup
                   13282: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
                   13283: 
                   13284: @item width of floating-point stack:
                   13285: @cindex floating-point stack width 
                   13286: @code{1 floats}.
                   13287: 
                   13288: @end table
                   13289: 
                   13290: 
                   13291: @c ---------------------------------------------------------------------
                   13292: @node floating-ambcond,  , floating-idef, The optional Floating-Point word set
                   13293: @subsection Ambiguous conditions
                   13294: @c ---------------------------------------------------------------------
                   13295: @cindex floating-point words, ambiguous conditions
                   13296: @cindex ambiguous conditions, floating-point words
                   13297: 
                   13298: @table @i
                   13299: @item @code{df@@} or @code{df!} used with an address that is not double-float  aligned:
                   13300: @cindex @code{df@@} or @code{df!} used with an address that is not double-float  aligned
                   13301: System-dependent. Typically results in a @code{-23 THROW} like other
                   13302: alignment violations.
                   13303: 
                   13304: @item @code{f@@} or @code{f!} used with an address that is not float  aligned:
                   13305: @cindex @code{f@@} used with an address that is not float aligned
                   13306: @cindex @code{f!} used with an address that is not float aligned
                   13307: System-dependent. Typically results in a @code{-23 THROW} like other
                   13308: alignment violations.
                   13309: 
                   13310: @item floating-point result out of range:
                   13311: @cindex floating-point result out of range
1.80      anton    13312: System-dependent. Can result in a @code{-43 throw} (floating point
                   13313: overflow), @code{-54 throw} (floating point underflow), @code{-41 throw}
                   13314: (floating point inexact result), @code{-55 THROW} (Floating-point
1.1       anton    13315: unidentified fault), or can produce a special value representing, e.g.,
                   13316: Infinity.
                   13317: 
                   13318: @item @code{sf@@} or @code{sf!} used with an address that is not single-float  aligned:
                   13319: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float  aligned
                   13320: System-dependent. Typically results in an alignment fault like other
                   13321: alignment violations.
                   13322: 
1.35      anton    13323: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
                   13324: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
1.1       anton    13325: The floating-point number is converted into decimal nonetheless.
                   13326: 
                   13327: @item Both arguments are equal to zero (@code{FATAN2}):
                   13328: @cindex @code{FATAN2}, both arguments are equal to zero
                   13329: System-dependent. @code{FATAN2} is implemented using the C library
                   13330: function @code{atan2()}.
                   13331: 
1.29      crook    13332: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
                   13333: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
                   13334: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
1.1       anton    13335: because of small errors and the tan will be a very large (or very small)
                   13336: but finite number.
                   13337: 
1.29      crook    13338: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
                   13339: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
1.1       anton    13340: The result is rounded to the nearest float.
                   13341: 
                   13342: @item dividing by zero:
                   13343: @cindex dividing by zero, floating-point
                   13344: @cindex floating-point dividing by zero
                   13345: @cindex floating-point unidentified fault, FP divide-by-zero
1.80      anton    13346: Platform-dependent; can produce an Infinity, NaN, @code{-42 throw}
                   13347: (floating point divide by zero) or @code{-55 throw} (Floating-point
                   13348: unidentified fault).
1.1       anton    13349: 
                   13350: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
                   13351: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
                   13352: System dependent. On IEEE-FP based systems the number is converted into
                   13353: an infinity.
                   13354: 
1.29      crook    13355: @item @i{float}<1 (@code{FACOSH}):
                   13356: @cindex @code{FACOSH}, @i{float}<1
1.1       anton    13357: @cindex floating-point unidentified fault, @code{FACOSH}
1.80      anton    13358: Platform-dependent; on IEEE-FP systems typically produces a NaN.
1.1       anton    13359: 
1.29      crook    13360: @item @i{float}=<-1 (@code{FLNP1}):
                   13361: @cindex @code{FLNP1}, @i{float}=<-1
1.1       anton    13362: @cindex floating-point unidentified fault, @code{FLNP1}
1.80      anton    13363: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
                   13364: negative infinity for @i{float}=-1).
1.1       anton    13365: 
1.29      crook    13366: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
                   13367: @cindex @code{FLN}, @i{float}=<0
                   13368: @cindex @code{FLOG}, @i{float}=<0
1.1       anton    13369: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
1.80      anton    13370: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
                   13371: negative infinity for @i{float}=0).
1.1       anton    13372: 
1.29      crook    13373: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
                   13374: @cindex @code{FASINH}, @i{float}<0
                   13375: @cindex @code{FSQRT}, @i{float}<0
1.1       anton    13376: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
1.80      anton    13377: Platform-dependent; for @code{fsqrt} this typically gives a NaN, for
                   13378: @code{fasinh} some platforms produce a NaN, others a number (bug in the
                   13379: C library?).
1.1       anton    13380: 
1.29      crook    13381: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
                   13382: @cindex @code{FACOS}, |@i{float}|>1
                   13383: @cindex @code{FASIN}, |@i{float}|>1
                   13384: @cindex @code{FATANH}, |@i{float}|>1
1.1       anton    13385: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
1.80      anton    13386: Platform-dependent; IEEE-FP systems typically produce a NaN.
1.1       anton    13387: 
1.29      crook    13388: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
                   13389: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
1.1       anton    13390: @cindex floating-point unidentified fault, @code{F>D}
1.80      anton    13391: Platform-dependent; typically, some double number is produced and no
                   13392: error is reported.
1.1       anton    13393: 
                   13394: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
                   13395: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
1.80      anton    13396: @code{Precision} characters of the numeric output area are used.  If
                   13397: @code{precision} is too high, these words will smash the data or code
                   13398: close to @code{here}.
1.1       anton    13399: @end table
                   13400: 
                   13401: @c =====================================================================
                   13402: @node  The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
                   13403: @section The optional Locals word set
                   13404: @c =====================================================================
                   13405: @cindex system documentation, locals words
                   13406: @cindex locals words, system documentation
                   13407: 
                   13408: @menu
                   13409: * locals-idef::                 Implementation Defined Options                 
                   13410: * locals-ambcond::              Ambiguous Conditions              
                   13411: @end menu
                   13412: 
                   13413: 
                   13414: @c ---------------------------------------------------------------------
                   13415: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
                   13416: @subsection Implementation Defined Options
                   13417: @c ---------------------------------------------------------------------
                   13418: @cindex implementation-defined options, locals words
                   13419: @cindex locals words, implementation-defined options
                   13420: 
                   13421: @table @i
                   13422: @item maximum number of locals in a definition:
                   13423: @cindex maximum number of locals in a definition
                   13424: @cindex locals, maximum number in a definition
                   13425: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
                   13426: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
                   13427: characters. The number of locals in a definition is bounded by the size
                   13428: of locals-buffer, which contains the names of the locals.
                   13429: 
                   13430: @end table
                   13431: 
                   13432: 
                   13433: @c ---------------------------------------------------------------------
                   13434: @node locals-ambcond,  , locals-idef, The optional Locals word set
                   13435: @subsection Ambiguous conditions
                   13436: @c ---------------------------------------------------------------------
                   13437: @cindex locals words, ambiguous conditions
                   13438: @cindex ambiguous conditions, locals words
                   13439: 
                   13440: @table @i
                   13441: @item executing a named local in interpretation state:
                   13442: @cindex local in interpretation state
                   13443: @cindex Interpreting a compile-only word, for a local
                   13444: Locals have no interpretation semantics. If you try to perform the
                   13445: interpretation semantics, you will get a @code{-14 throw} somewhere
                   13446: (Interpreting a compile-only word). If you perform the compilation
                   13447: semantics, the locals access will be compiled (irrespective of state).
                   13448: 
1.29      crook    13449: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
1.1       anton    13450: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
                   13451: @cindex @code{TO} on non-@code{VALUE}s and non-locals
                   13452: @cindex Invalid name argument, @code{TO}
                   13453: @code{-32 throw} (Invalid name argument)
                   13454: 
                   13455: @end table
                   13456: 
                   13457: 
                   13458: @c =====================================================================
                   13459: @node  The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
                   13460: @section The optional Memory-Allocation word set
                   13461: @c =====================================================================
                   13462: @cindex system documentation, memory-allocation words
                   13463: @cindex memory-allocation words, system documentation
                   13464: 
                   13465: @menu
                   13466: * memory-idef::                 Implementation Defined Options                 
                   13467: @end menu
                   13468: 
                   13469: 
                   13470: @c ---------------------------------------------------------------------
                   13471: @node memory-idef,  , The optional Memory-Allocation word set, The optional Memory-Allocation word set
                   13472: @subsection Implementation Defined Options
                   13473: @c ---------------------------------------------------------------------
                   13474: @cindex implementation-defined options, memory-allocation words
                   13475: @cindex memory-allocation words, implementation-defined options
                   13476: 
                   13477: @table @i
1.29      crook    13478: @item values and meaning of @i{ior}:
                   13479: @cindex  @i{ior} values and meaning
                   13480: The @i{ior}s returned by the file and memory allocation words are
1.1       anton    13481: intended as throw codes. They typically are in the range
                   13482: -512@minus{}-2047 of OS errors.  The mapping from OS error numbers to
1.29      crook    13483: @i{ior}s is -512@minus{}@i{errno}.
1.1       anton    13484: 
                   13485: @end table
                   13486: 
                   13487: @c =====================================================================
                   13488: @node  The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
                   13489: @section The optional Programming-Tools word set
                   13490: @c =====================================================================
                   13491: @cindex system documentation, programming-tools words
                   13492: @cindex programming-tools words, system documentation
                   13493: 
                   13494: @menu
                   13495: * programming-idef::            Implementation Defined Options            
                   13496: * programming-ambcond::         Ambiguous Conditions         
                   13497: @end menu
                   13498: 
                   13499: 
                   13500: @c ---------------------------------------------------------------------
                   13501: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
                   13502: @subsection Implementation Defined Options
                   13503: @c ---------------------------------------------------------------------
                   13504: @cindex implementation-defined options, programming-tools words
                   13505: @cindex programming-tools words, implementation-defined options
                   13506: 
                   13507: @table @i
                   13508: @item ending sequence for input following @code{;CODE} and @code{CODE}:
                   13509: @cindex @code{;CODE} ending sequence
                   13510: @cindex @code{CODE} ending sequence
                   13511: @code{END-CODE}
                   13512: 
                   13513: @item manner of processing input following @code{;CODE} and @code{CODE}:
                   13514: @cindex @code{;CODE}, processing input
                   13515: @cindex @code{CODE}, processing input
                   13516: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
                   13517: the input is processed by the text interpreter, (starting) in interpret
                   13518: state.
                   13519: 
                   13520: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
                   13521: @cindex @code{ASSEMBLER}, search order capability
                   13522: The ANS Forth search order word set.
                   13523: 
                   13524: @item source and format of display by @code{SEE}:
                   13525: @cindex @code{SEE}, source and format of output
1.80      anton    13526: The source for @code{see} is the executable code used by the inner
1.1       anton    13527: interpreter.  The current @code{see} tries to output Forth source code
1.80      anton    13528: (and on some platforms, assembly code for primitives) as well as
                   13529: possible.
1.1       anton    13530: 
                   13531: @end table
                   13532: 
                   13533: @c ---------------------------------------------------------------------
                   13534: @node programming-ambcond,  , programming-idef, The optional Programming-Tools word set
                   13535: @subsection Ambiguous conditions
                   13536: @c ---------------------------------------------------------------------
                   13537: @cindex programming-tools words, ambiguous conditions
                   13538: @cindex ambiguous conditions, programming-tools words
                   13539: 
                   13540: @table @i
                   13541: 
1.21      crook    13542: @item deleting the compilation word list (@code{FORGET}):
                   13543: @cindex @code{FORGET}, deleting the compilation word list
1.1       anton    13544: Not implemented (yet).
                   13545: 
1.29      crook    13546: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
                   13547: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
                   13548: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
1.1       anton    13549: @cindex control-flow stack underflow
                   13550: This typically results in an @code{abort"} with a descriptive error
                   13551: message (may change into a @code{-22 throw} (Control structure mismatch)
                   13552: in the future). You may also get a memory access error. If you are
                   13553: unlucky, this ambiguous condition is not caught.
                   13554: 
1.29      crook    13555: @item @i{name} can't be found (@code{FORGET}):
                   13556: @cindex @code{FORGET}, @i{name} can't be found
1.1       anton    13557: Not implemented (yet).
                   13558: 
1.29      crook    13559: @item @i{name} not defined via @code{CREATE}:
                   13560: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
1.1       anton    13561: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
                   13562: the execution semantics of the last defined word no matter how it was
                   13563: defined.
                   13564: 
                   13565: @item @code{POSTPONE} applied to @code{[IF]}:
                   13566: @cindex @code{POSTPONE} applied to @code{[IF]}
                   13567: @cindex @code{[IF]} and @code{POSTPONE}
                   13568: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
                   13569: equivalent to @code{[IF]}.
                   13570: 
                   13571: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
                   13572: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
                   13573: Continue in the same state of conditional compilation in the next outer
                   13574: input source. Currently there is no warning to the user about this.
                   13575: 
                   13576: @item removing a needed definition (@code{FORGET}):
                   13577: @cindex @code{FORGET}, removing a needed definition
                   13578: Not implemented (yet).
                   13579: 
                   13580: @end table
                   13581: 
                   13582: 
                   13583: @c =====================================================================
                   13584: @node  The optional Search-Order word set,  , The optional Programming-Tools word set, ANS conformance
                   13585: @section The optional Search-Order word set
                   13586: @c =====================================================================
                   13587: @cindex system documentation, search-order words
                   13588: @cindex search-order words, system documentation
                   13589: 
                   13590: @menu
                   13591: * search-idef::                 Implementation Defined Options                 
                   13592: * search-ambcond::              Ambiguous Conditions              
                   13593: @end menu
                   13594: 
                   13595: 
                   13596: @c ---------------------------------------------------------------------
                   13597: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
                   13598: @subsection Implementation Defined Options
                   13599: @c ---------------------------------------------------------------------
                   13600: @cindex implementation-defined options, search-order words
                   13601: @cindex search-order words, implementation-defined options
                   13602: 
                   13603: @table @i
                   13604: @item maximum number of word lists in search order:
                   13605: @cindex maximum number of word lists in search order
                   13606: @cindex search order, maximum depth
                   13607: @code{s" wordlists" environment? drop .}. Currently 16.
                   13608: 
                   13609: @item minimum search order:
                   13610: @cindex minimum search order
                   13611: @cindex search order, minimum
                   13612: @code{root root}.
                   13613: 
                   13614: @end table
                   13615: 
                   13616: @c ---------------------------------------------------------------------
                   13617: @node search-ambcond,  , search-idef, The optional Search-Order word set
                   13618: @subsection Ambiguous conditions
                   13619: @c ---------------------------------------------------------------------
                   13620: @cindex search-order words, ambiguous conditions
                   13621: @cindex ambiguous conditions, search-order words
                   13622: 
                   13623: @table @i
1.21      crook    13624: @item changing the compilation word list (during compilation):
                   13625: @cindex changing the compilation word list (during compilation)
                   13626: @cindex compilation word list, change before definition ends
                   13627: The word is entered into the word list that was the compilation word list
1.1       anton    13628: at the start of the definition. Any changes to the name field (e.g.,
                   13629: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
                   13630: are applied to the latest defined word (as reported by @code{last} or
1.21      crook    13631: @code{lastxt}), if possible, irrespective of the compilation word list.
1.1       anton    13632: 
                   13633: @item search order empty (@code{previous}):
                   13634: @cindex @code{previous}, search order empty
1.26      crook    13635: @cindex vocstack empty, @code{previous}
1.1       anton    13636: @code{abort" Vocstack empty"}.
                   13637: 
                   13638: @item too many word lists in search order (@code{also}):
                   13639: @cindex @code{also}, too many word lists in search order
1.26      crook    13640: @cindex vocstack full, @code{also}
1.1       anton    13641: @code{abort" Vocstack full"}.
                   13642: 
                   13643: @end table
                   13644: 
                   13645: @c ***************************************************************
1.65      anton    13646: @node Standard vs Extensions, Model, ANS conformance, Top
                   13647: @chapter Should I use Gforth extensions?
                   13648: @cindex Gforth extensions
                   13649: 
                   13650: As you read through the rest of this manual, you will see documentation
                   13651: for @i{Standard} words, and documentation for some appealing Gforth
                   13652: @i{extensions}. You might ask yourself the question: @i{``Should I
                   13653: restrict myself to the standard, or should I use the extensions?''}
                   13654: 
                   13655: The answer depends on the goals you have for the program you are working
                   13656: on:
                   13657: 
                   13658: @itemize @bullet
                   13659: 
                   13660: @item Is it just for yourself or do you want to share it with others?
                   13661: 
                   13662: @item
                   13663: If you want to share it, do the others all use Gforth?
                   13664: 
                   13665: @item
                   13666: If it is just for yourself, do you want to restrict yourself to Gforth?
                   13667: 
                   13668: @end itemize
                   13669: 
                   13670: If restricting the program to Gforth is ok, then there is no reason not
                   13671: to use extensions.  It is still a good idea to keep to the standard
                   13672: where it is easy, in case you want to reuse these parts in another
                   13673: program that you want to be portable.
                   13674: 
                   13675: If you want to be able to port the program to other Forth systems, there
                   13676: are the following points to consider:
                   13677: 
                   13678: @itemize @bullet
                   13679: 
                   13680: @item
                   13681: Most Forth systems that are being maintained support the ANS Forth
                   13682: standard.  So if your program complies with the standard, it will be
                   13683: portable among many systems.
                   13684: 
                   13685: @item
                   13686: A number of the Gforth extensions can be implemented in ANS Forth using
                   13687: public-domain files provided in the @file{compat/} directory. These are
                   13688: mentioned in the text in passing.  There is no reason not to use these
                   13689: extensions, your program will still be ANS Forth compliant; just include
                   13690: the appropriate compat files with your program.
                   13691: 
                   13692: @item
                   13693: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
                   13694: analyse your program and determine what non-Standard words it relies
                   13695: upon.  However, it does not check whether you use standard words in a
                   13696: non-standard way.
                   13697: 
                   13698: @item
                   13699: Some techniques are not standardized by ANS Forth, and are hard or
                   13700: impossible to implement in a standard way, but can be implemented in
                   13701: most Forth systems easily, and usually in similar ways (e.g., accessing
                   13702: word headers).  Forth has a rich historical precedent for programmers
                   13703: taking advantage of implementation-dependent features of their tools
                   13704: (for example, relying on a knowledge of the dictionary
                   13705: structure). Sometimes these techniques are necessary to extract every
                   13706: last bit of performance from the hardware, sometimes they are just a
                   13707: programming shorthand.
                   13708: 
                   13709: @item
                   13710: Does using a Gforth extension save more work than the porting this part
                   13711: to other Forth systems (if any) will cost?
                   13712: 
                   13713: @item
                   13714: Is the additional functionality worth the reduction in portability and
                   13715: the additional porting problems?
                   13716: 
                   13717: @end itemize
                   13718: 
                   13719: In order to perform these consideratios, you need to know what's
                   13720: standard and what's not.  This manual generally states if something is
1.81      anton    13721: non-standard, but the authoritative source is the
                   13722: @uref{http://www.taygeta.com/forth/dpans.html,standard document}.
1.65      anton    13723: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
                   13724: into the thought processes of the technical committee.
                   13725: 
                   13726: Note also that portability between Forth systems is not the only
                   13727: portability issue; there is also the issue of portability between
                   13728: different platforms (processor/OS combinations).
                   13729: 
                   13730: @c ***************************************************************
                   13731: @node Model, Integrating Gforth, Standard vs Extensions, Top
1.1       anton    13732: @chapter Model
                   13733: 
                   13734: This chapter has yet to be written. It will contain information, on
                   13735: which internal structures you can rely.
                   13736: 
                   13737: @c ***************************************************************
                   13738: @node Integrating Gforth, Emacs and Gforth, Model, Top
                   13739: @chapter Integrating Gforth into C programs
                   13740: 
                   13741: This is not yet implemented.
                   13742: 
                   13743: Several people like to use Forth as scripting language for applications
                   13744: that are otherwise written in C, C++, or some other language.
                   13745: 
                   13746: The Forth system ATLAST provides facilities for embedding it into
                   13747: applications; unfortunately it has several disadvantages: most
                   13748: importantly, it is not based on ANS Forth, and it is apparently dead
                   13749: (i.e., not developed further and not supported). The facilities
1.21      crook    13750: provided by Gforth in this area are inspired by ATLAST's facilities, so
1.1       anton    13751: making the switch should not be hard.
                   13752: 
                   13753: We also tried to design the interface such that it can easily be
                   13754: implemented by other Forth systems, so that we may one day arrive at a
                   13755: standardized interface. Such a standard interface would allow you to
                   13756: replace the Forth system without having to rewrite C code.
                   13757: 
                   13758: You embed the Gforth interpreter by linking with the library
                   13759: @code{libgforth.a} (give the compiler the option @code{-lgforth}).  All
                   13760: global symbols in this library that belong to the interface, have the
                   13761: prefix @code{forth_}. (Global symbols that are used internally have the
                   13762: prefix @code{gforth_}).
                   13763: 
                   13764: You can include the declarations of Forth types and the functions and
                   13765: variables of the interface with @code{#include <forth.h>}.
                   13766: 
                   13767: Types.
                   13768: 
                   13769: Variables.
                   13770: 
                   13771: Data and FP Stack pointer. Area sizes.
                   13772: 
                   13773: functions.
                   13774: 
                   13775: forth_init(imagefile)
                   13776: forth_evaluate(string) exceptions?
                   13777: forth_goto(address) (or forth_execute(xt)?)
                   13778: forth_continue() (a corountining mechanism)
                   13779: 
                   13780: Adding primitives.
                   13781: 
                   13782: No checking.
                   13783: 
                   13784: Signals?
                   13785: 
                   13786: Accessing the Stacks
                   13787: 
1.26      crook    13788: @c ******************************************************************
1.1       anton    13789: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
                   13790: @chapter Emacs and Gforth
                   13791: @cindex Emacs and Gforth
                   13792: 
                   13793: @cindex @file{gforth.el}
                   13794: @cindex @file{forth.el}
                   13795: @cindex Rydqvist, Goran
                   13796: @cindex comment editing commands
                   13797: @cindex @code{\}, editing with Emacs
                   13798: @cindex debug tracer editing commands
                   13799: @cindex @code{~~}, removal with Emacs
                   13800: @cindex Forth mode in Emacs
                   13801: Gforth comes with @file{gforth.el}, an improved version of
                   13802: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
1.26      crook    13803: improvements are:
                   13804: 
                   13805: @itemize @bullet
                   13806: @item
                   13807: A better (but still not perfect) handling of indentation.
                   13808: @item
                   13809: Comment paragraph filling (@kbd{M-q})
                   13810: @item
                   13811: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
                   13812: @item
                   13813: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
1.41      anton    13814: @item
                   13815: Support of the @code{info-lookup} feature for looking up the
                   13816: documentation of a word.
1.26      crook    13817: @end itemize
                   13818: 
                   13819: I left the stuff I do not use alone, even though some of it only makes
                   13820: sense for TILE. To get a description of these features, enter Forth mode
                   13821: and type @kbd{C-h m}.
1.1       anton    13822: 
                   13823: @cindex source location of error or debugging output in Emacs
                   13824: @cindex error output, finding the source location in Emacs
                   13825: @cindex debugging output, finding the source location in Emacs
                   13826: In addition, Gforth supports Emacs quite well: The source code locations
                   13827: given in error messages, debugging output (from @code{~~}) and failed
                   13828: assertion messages are in the right format for Emacs' compilation mode
                   13829: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
                   13830: Manual}) so the source location corresponding to an error or other
                   13831: message is only a few keystrokes away (@kbd{C-x `} for the next error,
                   13832: @kbd{C-c C-c} for the error under the cursor).
                   13833: 
                   13834: @cindex @file{TAGS} file
                   13835: @cindex @file{etags.fs}
                   13836: @cindex viewing the source of a word in Emacs
1.43      anton    13837: @cindex @code{require}, placement in files
                   13838: @cindex @code{include}, placement in files
                   13839: Also, if you @code{require} @file{etags.fs}, a new @file{TAGS} file will
1.26      crook    13840: be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
1.1       anton    13841: contains the definitions of all words defined afterwards. You can then
                   13842: find the source for a word using @kbd{M-.}. Note that emacs can use
                   13843: several tags files at the same time (e.g., one for the Gforth sources
                   13844: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
                   13845: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
                   13846: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
1.43      anton    13847: @file{/usr/local/share/gforth/0.2.0/TAGS}).  To get the best behaviour
                   13848: with @file{etags.fs}, you should avoid putting definitions both before
                   13849: and after @code{require} etc., otherwise you will see the same file
                   13850: visited several times by commands like @code{tags-search}.
1.1       anton    13851: 
1.41      anton    13852: @cindex viewing the documentation of a word in Emacs
                   13853: @cindex context-sensitive help
                   13854: Moreover, for words documented in this manual, you can look up the
                   13855: glossary entry quickly by using @kbd{C-h TAB}
1.80      anton    13856: (@code{info-lookup-symbol}, @pxref{Documentation, ,Documentation
1.41      anton    13857: Commands, emacs, Emacs Manual}).  This feature requires Emacs 20.3 or
1.42      anton    13858: later and does not work for words containing @code{:}.
1.41      anton    13859: 
                   13860: 
1.1       anton    13861: @cindex @file{.emacs}
                   13862: To get all these benefits, add the following lines to your @file{.emacs}
                   13863: file:
                   13864: 
                   13865: @example
                   13866: (autoload 'forth-mode "gforth.el")
                   13867: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
                   13868: @end example
                   13869: 
1.26      crook    13870: @c ******************************************************************
1.1       anton    13871: @node Image Files, Engine, Emacs and Gforth, Top
                   13872: @chapter Image Files
1.26      crook    13873: @cindex image file
                   13874: @cindex @file{.fi} files
1.1       anton    13875: @cindex precompiled Forth code
                   13876: @cindex dictionary in persistent form
                   13877: @cindex persistent form of dictionary
                   13878: 
                   13879: An image file is a file containing an image of the Forth dictionary,
                   13880: i.e., compiled Forth code and data residing in the dictionary.  By
                   13881: convention, we use the extension @code{.fi} for image files.
                   13882: 
                   13883: @menu
1.18      anton    13884: * Image Licensing Issues::      Distribution terms for images.
                   13885: * Image File Background::       Why have image files?
1.67      anton    13886: * Non-Relocatable Image Files::  don't always work.
1.18      anton    13887: * Data-Relocatable Image Files::  are better.
1.67      anton    13888: * Fully Relocatable Image Files::  better yet.
1.18      anton    13889: * Stack and Dictionary Sizes::  Setting the default sizes for an image.
1.29      crook    13890: * Running Image Files::         @code{gforth -i @i{file}} or @i{file}.
1.18      anton    13891: * Modifying the Startup Sequence::  and turnkey applications.
1.1       anton    13892: @end menu
                   13893: 
1.18      anton    13894: @node Image Licensing Issues, Image File Background, Image Files, Image Files
                   13895: @section Image Licensing Issues
                   13896: @cindex license for images
                   13897: @cindex image license
                   13898: 
                   13899: An image created with @code{gforthmi} (@pxref{gforthmi}) or
                   13900: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
                   13901: original image; i.e., according to copyright law it is a derived work of
                   13902: the original image.
                   13903: 
                   13904: Since Gforth is distributed under the GNU GPL, the newly created image
                   13905: falls under the GNU GPL, too. In particular, this means that if you
                   13906: distribute the image, you have to make all of the sources for the image
                   13907: available, including those you wrote.  For details see @ref{License, ,
                   13908: GNU General Public License (Section 3)}.
                   13909: 
                   13910: If you create an image with @code{cross} (@pxref{cross.fs}), the image
                   13911: contains only code compiled from the sources you gave it; if none of
                   13912: these sources is under the GPL, the terms discussed above do not apply
                   13913: to the image. However, if your image needs an engine (a gforth binary)
                   13914: that is under the GPL, you should make sure that you distribute both in
                   13915: a way that is at most a @emph{mere aggregation}, if you don't want the
                   13916: terms of the GPL to apply to the image.
                   13917: 
                   13918: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
1.1       anton    13919: @section Image File Background
                   13920: @cindex image file background
                   13921: 
1.80      anton    13922: Gforth consists not only of primitives (in the engine), but also of
1.1       anton    13923: definitions written in Forth. Since the Forth compiler itself belongs to
                   13924: those definitions, it is not possible to start the system with the
1.80      anton    13925: engine and the Forth source alone. Therefore we provide the Forth
1.26      crook    13926: code as an image file in nearly executable form. When Gforth starts up,
                   13927: a C routine loads the image file into memory, optionally relocates the
                   13928: addresses, then sets up the memory (stacks etc.) according to
                   13929: information in the image file, and (finally) starts executing Forth
                   13930: code.
1.1       anton    13931: 
                   13932: The image file variants represent different compromises between the
                   13933: goals of making it easy to generate image files and making them
                   13934: portable.
                   13935: 
                   13936: @cindex relocation at run-time
1.26      crook    13937: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
1.1       anton    13938: run-time. This avoids many of the complications discussed below (image
                   13939: files are data relocatable without further ado), but costs performance
                   13940: (one addition per memory access).
                   13941: 
                   13942: @cindex relocation at load-time
1.26      crook    13943: By contrast, the Gforth loader performs relocation at image load time. The
                   13944: loader also has to replace tokens that represent primitive calls with the
1.1       anton    13945: appropriate code-field addresses (or code addresses in the case of
                   13946: direct threading).
                   13947: 
                   13948: There are three kinds of image files, with different degrees of
                   13949: relocatability: non-relocatable, data-relocatable, and fully relocatable
                   13950: image files.
                   13951: 
                   13952: @cindex image file loader
                   13953: @cindex relocating loader
                   13954: @cindex loader for image files
                   13955: These image file variants have several restrictions in common; they are
                   13956: caused by the design of the image file loader:
                   13957: 
                   13958: @itemize @bullet
                   13959: @item
                   13960: There is only one segment; in particular, this means, that an image file
                   13961: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
1.26      crook    13962: them). The contents of the stacks are not represented, either.
1.1       anton    13963: 
                   13964: @item
                   13965: The only kinds of relocation supported are: adding the same offset to
                   13966: all cells that represent data addresses; and replacing special tokens
                   13967: with code addresses or with pieces of machine code.
                   13968: 
                   13969: If any complex computations involving addresses are performed, the
                   13970: results cannot be represented in the image file. Several applications that
                   13971: use such computations come to mind:
                   13972: @itemize @minus
                   13973: @item
                   13974: Hashing addresses (or data structures which contain addresses) for table
                   13975: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
                   13976: purpose, you will have no problem, because the hash tables are
                   13977: recomputed automatically when the system is started. If you use your own
                   13978: hash tables, you will have to do something similar.
                   13979: 
                   13980: @item
                   13981: There's a cute implementation of doubly-linked lists that uses
                   13982: @code{XOR}ed addresses. You could represent such lists as singly-linked
                   13983: in the image file, and restore the doubly-linked representation on
                   13984: startup.@footnote{In my opinion, though, you should think thrice before
                   13985: using a doubly-linked list (whatever implementation).}
                   13986: 
                   13987: @item
                   13988: The code addresses of run-time routines like @code{docol:} cannot be
                   13989: represented in the image file (because their tokens would be replaced by
                   13990: machine code in direct threaded implementations). As a workaround,
                   13991: compute these addresses at run-time with @code{>code-address} from the
                   13992: executions tokens of appropriate words (see the definitions of
1.80      anton    13993: @code{docol:} and friends in @file{kernel/getdoers.fs}).
1.1       anton    13994: 
                   13995: @item
                   13996: On many architectures addresses are represented in machine code in some
                   13997: shifted or mangled form. You cannot put @code{CODE} words that contain
                   13998: absolute addresses in this form in a relocatable image file. Workarounds
                   13999: are representing the address in some relative form (e.g., relative to
                   14000: the CFA, which is present in some register), or loading the address from
                   14001: a place where it is stored in a non-mangled form.
                   14002: @end itemize
                   14003: @end itemize
                   14004: 
                   14005: @node  Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
                   14006: @section Non-Relocatable Image Files
                   14007: @cindex non-relocatable image files
1.26      crook    14008: @cindex image file, non-relocatable
1.1       anton    14009: 
                   14010: These files are simple memory dumps of the dictionary. They are specific
                   14011: to the executable (i.e., @file{gforth} file) they were created
                   14012: with. What's worse, they are specific to the place on which the
                   14013: dictionary resided when the image was created. Now, there is no
                   14014: guarantee that the dictionary will reside at the same place the next
                   14015: time you start Gforth, so there's no guarantee that a non-relocatable
                   14016: image will work the next time (Gforth will complain instead of crashing,
                   14017: though).
                   14018: 
                   14019: You can create a non-relocatable image file with
                   14020: 
1.44      crook    14021: 
1.1       anton    14022: doc-savesystem
                   14023: 
1.44      crook    14024: 
1.1       anton    14025: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
                   14026: @section Data-Relocatable Image Files
                   14027: @cindex data-relocatable image files
1.26      crook    14028: @cindex image file, data-relocatable
1.1       anton    14029: 
                   14030: These files contain relocatable data addresses, but fixed code addresses
                   14031: (instead of tokens). They are specific to the executable (i.e.,
                   14032: @file{gforth} file) they were created with. For direct threading on some
                   14033: architectures (e.g., the i386), data-relocatable images do not work. You
                   14034: get a data-relocatable image, if you use @file{gforthmi} with a
                   14035: Gforth binary that is not doubly indirect threaded (@pxref{Fully
                   14036: Relocatable Image Files}).
                   14037: 
                   14038: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
                   14039: @section Fully Relocatable Image Files
                   14040: @cindex fully relocatable image files
1.26      crook    14041: @cindex image file, fully relocatable
1.1       anton    14042: 
                   14043: @cindex @file{kern*.fi}, relocatability
                   14044: @cindex @file{gforth.fi}, relocatability
                   14045: These image files have relocatable data addresses, and tokens for code
                   14046: addresses. They can be used with different binaries (e.g., with and
                   14047: without debugging) on the same machine, and even across machines with
                   14048: the same data formats (byte order, cell size, floating point
                   14049: format). However, they are usually specific to the version of Gforth
                   14050: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
                   14051: are fully relocatable.
                   14052: 
                   14053: There are two ways to create a fully relocatable image file:
                   14054: 
                   14055: @menu
1.29      crook    14056: * gforthmi::                    The normal way
1.1       anton    14057: * cross.fs::                    The hard way
                   14058: @end menu
                   14059: 
                   14060: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
                   14061: @subsection @file{gforthmi}
                   14062: @cindex @file{comp-i.fs}
                   14063: @cindex @file{gforthmi}
                   14064: 
                   14065: You will usually use @file{gforthmi}. If you want to create an
1.29      crook    14066: image @i{file} that contains everything you would load by invoking
                   14067: Gforth with @code{gforth @i{options}}, you simply say:
1.1       anton    14068: @example
1.29      crook    14069: gforthmi @i{file} @i{options}
1.1       anton    14070: @end example
                   14071: 
                   14072: E.g., if you want to create an image @file{asm.fi} that has the file
                   14073: @file{asm.fs} loaded in addition to the usual stuff, you could do it
                   14074: like this:
                   14075: 
                   14076: @example
                   14077: gforthmi asm.fi asm.fs
                   14078: @end example
                   14079: 
1.27      crook    14080: @file{gforthmi} is implemented as a sh script and works like this: It
                   14081: produces two non-relocatable images for different addresses and then
                   14082: compares them. Its output reflects this: first you see the output (if
1.62      crook    14083: any) of the two Gforth invocations that produce the non-relocatable image
1.27      crook    14084: files, then you see the output of the comparing program: It displays the
                   14085: offset used for data addresses and the offset used for code addresses;
1.1       anton    14086: moreover, for each cell that cannot be represented correctly in the
1.44      crook    14087: image files, it displays a line like this:
1.1       anton    14088: 
                   14089: @example
                   14090:      78DC         BFFFFA50         BFFFFA40
                   14091: @end example
                   14092: 
                   14093: This means that at offset $78dc from @code{forthstart}, one input image
                   14094: contains $bffffa50, and the other contains $bffffa40. Since these cells
                   14095: cannot be represented correctly in the output image, you should examine
                   14096: these places in the dictionary and verify that these cells are dead
                   14097: (i.e., not read before they are written).
1.39      anton    14098: 
                   14099: @cindex --application, @code{gforthmi} option
                   14100: If you insert the option @code{--application} in front of the image file
                   14101: name, you will get an image that uses the @code{--appl-image} option
                   14102: instead of the @code{--image-file} option (@pxref{Invoking
                   14103: Gforth}). When you execute such an image on Unix (by typing the image
                   14104: name as command), the Gforth engine will pass all options to the image
                   14105: instead of trying to interpret them as engine options.
1.1       anton    14106: 
1.27      crook    14107: If you type @file{gforthmi} with no arguments, it prints some usage
                   14108: instructions.
                   14109: 
1.1       anton    14110: @cindex @code{savesystem} during @file{gforthmi}
                   14111: @cindex @code{bye} during @file{gforthmi}
                   14112: @cindex doubly indirect threaded code
1.44      crook    14113: @cindex environment variables
                   14114: @cindex @code{GFORTHD} -- environment variable
                   14115: @cindex @code{GFORTH} -- environment variable
1.1       anton    14116: @cindex @code{gforth-ditc}
1.29      crook    14117: There are a few wrinkles: After processing the passed @i{options}, the
1.1       anton    14118: words @code{savesystem} and @code{bye} must be visible. A special doubly
                   14119: indirect threaded version of the @file{gforth} executable is used for
1.62      crook    14120: creating the non-relocatable images; you can pass the exact filename of
1.1       anton    14121: this executable through the environment variable @code{GFORTHD}
                   14122: (default: @file{gforth-ditc}); if you pass a version that is not doubly
                   14123: indirect threaded, you will not get a fully relocatable image, but a
1.27      crook    14124: data-relocatable image (because there is no code address offset). The
                   14125: normal @file{gforth} executable is used for creating the relocatable
                   14126: image; you can pass the exact filename of this executable through the
                   14127: environment variable @code{GFORTH}.
1.1       anton    14128: 
                   14129: @node cross.fs,  , gforthmi, Fully Relocatable Image Files
                   14130: @subsection @file{cross.fs}
                   14131: @cindex @file{cross.fs}
                   14132: @cindex cross-compiler
                   14133: @cindex metacompiler
1.47      crook    14134: @cindex target compiler
1.1       anton    14135: 
                   14136: You can also use @code{cross}, a batch compiler that accepts a Forth-like
1.47      crook    14137: programming language (@pxref{Cross Compiler}).
1.1       anton    14138: 
1.47      crook    14139: @code{cross} allows you to create image files for machines with
1.1       anton    14140: different data sizes and data formats than the one used for generating
                   14141: the image file. You can also use it to create an application image that
                   14142: does not contain a Forth compiler. These features are bought with
                   14143: restrictions and inconveniences in programming. E.g., addresses have to
                   14144: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
                   14145: order to make the code relocatable.
                   14146: 
                   14147: 
                   14148: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
                   14149: @section Stack and Dictionary Sizes
                   14150: @cindex image file, stack and dictionary sizes
                   14151: @cindex dictionary size default
                   14152: @cindex stack size default
                   14153: 
                   14154: If you invoke Gforth with a command line flag for the size
                   14155: (@pxref{Invoking Gforth}), the size you specify is stored in the
                   14156: dictionary. If you save the dictionary with @code{savesystem} or create
                   14157: an image with @file{gforthmi}, this size will become the default
                   14158: for the resulting image file. E.g., the following will create a
1.21      crook    14159: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
1.1       anton    14160: 
                   14161: @example
                   14162: gforthmi gforth.fi -m 1M
                   14163: @end example
                   14164: 
                   14165: In other words, if you want to set the default size for the dictionary
                   14166: and the stacks of an image, just invoke @file{gforthmi} with the
                   14167: appropriate options when creating the image.
                   14168: 
                   14169: @cindex stack size, cache-friendly
                   14170: Note: For cache-friendly behaviour (i.e., good performance), you should
                   14171: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
                   14172: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
                   14173: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
                   14174: 
                   14175: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
                   14176: @section Running Image Files
                   14177: @cindex running image files
                   14178: @cindex invoking image files
                   14179: @cindex image file invocation
                   14180: 
                   14181: @cindex -i, invoke image file
                   14182: @cindex --image file, invoke image file
1.29      crook    14183: You can invoke Gforth with an image file @i{image} instead of the
1.1       anton    14184: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
                   14185: @example
1.29      crook    14186: gforth -i @i{image}
1.1       anton    14187: @end example
                   14188: 
                   14189: @cindex executable image file
1.26      crook    14190: @cindex image file, executable
1.1       anton    14191: If your operating system supports starting scripts with a line of the
                   14192: form @code{#! ...}, you just have to type the image file name to start
                   14193: Gforth with this image file (note that the file extension @code{.fi} is
1.29      crook    14194: just a convention). I.e., to run Gforth with the image file @i{image},
                   14195: you can just type @i{image} instead of @code{gforth -i @i{image}}.
1.27      crook    14196: This works because every @code{.fi} file starts with a line of this
                   14197: format:
                   14198: 
                   14199: @example
                   14200: #! /usr/local/bin/gforth-0.4.0 -i
                   14201: @end example
                   14202: 
                   14203: The file and pathname for the Gforth engine specified on this line is
                   14204: the specific Gforth executable that it was built against; i.e. the value
                   14205: of the environment variable @code{GFORTH} at the time that
                   14206: @file{gforthmi} was executed.
1.1       anton    14207: 
1.27      crook    14208: You can make use of the same shell capability to make a Forth source
                   14209: file into an executable. For example, if you place this text in a file:
1.26      crook    14210: 
                   14211: @example
                   14212: #! /usr/local/bin/gforth
                   14213: 
                   14214: ." Hello, world" CR
                   14215: bye
                   14216: @end example
                   14217: 
                   14218: @noindent
1.27      crook    14219: and then make the file executable (chmod +x in Unix), you can run it
1.26      crook    14220: directly from the command line. The sequence @code{#!} is used in two
                   14221: ways; firstly, it is recognised as a ``magic sequence'' by the operating
1.29      crook    14222: system@footnote{The Unix kernel actually recognises two types of files:
                   14223: executable files and files of data, where the data is processed by an
                   14224: interpreter that is specified on the ``interpreter line'' -- the first
                   14225: line of the file, starting with the sequence #!. There may be a small
                   14226: limit (e.g., 32) on the number of characters that may be specified on
                   14227: the interpreter line.} secondly it is treated as a comment character by
                   14228: Gforth. Because of the second usage, a space is required between
1.80      anton    14229: @code{#!} and the path to the executable (moreover, some Unixes
                   14230: require the sequence @code{#! /}).
1.27      crook    14231: 
                   14232: The disadvantage of this latter technique, compared with using
1.80      anton    14233: @file{gforthmi}, is that it is slightly slower; the Forth source code is
                   14234: compiled on-the-fly, each time the program is invoked.
1.26      crook    14235: 
1.1       anton    14236: doc-#!
                   14237: 
1.44      crook    14238: 
1.1       anton    14239: @node Modifying the Startup Sequence,  , Running Image Files, Image Files
                   14240: @section Modifying the Startup Sequence
                   14241: @cindex startup sequence for image file
                   14242: @cindex image file initialization sequence
                   14243: @cindex initialization sequence of image file
                   14244: 
                   14245: You can add your own initialization to the startup sequence through the
1.26      crook    14246: deferred word @code{'cold}. @code{'cold} is invoked just before the
1.80      anton    14247: image-specific command line processing (i.e., loading files and
1.26      crook    14248: evaluating (@code{-e}) strings) starts.
1.1       anton    14249: 
                   14250: A sequence for adding your initialization usually looks like this:
                   14251: 
                   14252: @example
                   14253: :noname
                   14254:     Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
                   14255:     ... \ your stuff
                   14256: ; IS 'cold
                   14257: @end example
                   14258: 
                   14259: @cindex turnkey image files
1.26      crook    14260: @cindex image file, turnkey applications
1.1       anton    14261: You can make a turnkey image by letting @code{'cold} execute a word
                   14262: (your turnkey application) that never returns; instead, it exits Gforth
                   14263: via @code{bye} or @code{throw}.
                   14264: 
                   14265: @cindex command-line arguments, access
                   14266: @cindex arguments on the command line, access
                   14267: You can access the (image-specific) command-line arguments through the
1.26      crook    14268: variables @code{argc} and @code{argv}. @code{arg} provides convenient
1.1       anton    14269: access to @code{argv}.
                   14270: 
1.26      crook    14271: If @code{'cold} exits normally, Gforth processes the command-line
                   14272: arguments as files to be loaded and strings to be evaluated.  Therefore,
                   14273: @code{'cold} should remove the arguments it has used in this case.
                   14274: 
1.44      crook    14275: 
                   14276: 
1.26      crook    14277: doc-'cold
1.1       anton    14278: doc-argc
                   14279: doc-argv
                   14280: doc-arg
                   14281: 
                   14282: 
1.44      crook    14283: 
1.1       anton    14284: @c ******************************************************************
1.13      pazsan   14285: @node Engine, Binding to System Library, Image Files, Top
1.1       anton    14286: @chapter Engine
                   14287: @cindex engine
                   14288: @cindex virtual machine
                   14289: 
1.26      crook    14290: Reading this chapter is not necessary for programming with Gforth. It
1.1       anton    14291: may be helpful for finding your way in the Gforth sources.
                   14292: 
1.66      anton    14293: The ideas in this section have also been published in Bernd Paysan,
                   14294: @cite{ANS fig/GNU/??? Forth} (in German), Forth-Tagung '93 and M. Anton
                   14295: Ertl, @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
                   14296: Portable Forth Engine}}, EuroForth '93.
1.1       anton    14297: 
                   14298: @menu
                   14299: * Portability::                 
                   14300: * Threading::                   
                   14301: * Primitives::                  
                   14302: * Performance::                 
                   14303: @end menu
                   14304: 
                   14305: @node Portability, Threading, Engine, Engine
                   14306: @section Portability
                   14307: @cindex engine portability
                   14308: 
1.26      crook    14309: An important goal of the Gforth Project is availability across a wide
                   14310: range of personal machines. fig-Forth, and, to a lesser extent, F83,
                   14311: achieved this goal by manually coding the engine in assembly language
                   14312: for several then-popular processors. This approach is very
                   14313: labor-intensive and the results are short-lived due to progress in
                   14314: computer architecture.
1.1       anton    14315: 
                   14316: @cindex C, using C for the engine
                   14317: Others have avoided this problem by coding in C, e.g., Mitch Bradley
                   14318: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
                   14319: particularly popular for UNIX-based Forths due to the large variety of
                   14320: architectures of UNIX machines. Unfortunately an implementation in C
                   14321: does not mix well with the goals of efficiency and with using
                   14322: traditional techniques: Indirect or direct threading cannot be expressed
                   14323: in C, and switch threading, the fastest technique available in C, is
                   14324: significantly slower. Another problem with C is that it is very
                   14325: cumbersome to express double integer arithmetic.
                   14326: 
                   14327: @cindex GNU C for the engine
                   14328: @cindex long long
                   14329: Fortunately, there is a portable language that does not have these
                   14330: limitations: GNU C, the version of C processed by the GNU C compiler
                   14331: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
                   14332: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
                   14333: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
                   14334: threading possible, its @code{long long} type (@pxref{Long Long, ,
                   14335: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
                   14336: double numbers@footnote{Unfortunately, long longs are not implemented
                   14337: properly on all machines (e.g., on alpha-osf1, long longs are only 64
                   14338: bits, the same size as longs (and pointers), but they should be twice as
1.4       anton    14339: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
1.1       anton    14340: C Manual}). So, we had to implement doubles in C after all. Still, on
                   14341: most machines we can use long longs and achieve better performance than
                   14342: with the emulation package.}. GNU C is available for free on all
                   14343: important (and many unimportant) UNIX machines, VMS, 80386s running
                   14344: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
                   14345: on all these machines.
                   14346: 
                   14347: Writing in a portable language has the reputation of producing code that
                   14348: is slower than assembly. For our Forth engine we repeatedly looked at
                   14349: the code produced by the compiler and eliminated most compiler-induced
                   14350: inefficiencies by appropriate changes in the source code.
                   14351: 
                   14352: @cindex explicit register declarations
                   14353: @cindex --enable-force-reg, configuration flag
                   14354: @cindex -DFORCE_REG
                   14355: However, register allocation cannot be portably influenced by the
                   14356: programmer, leading to some inefficiencies on register-starved
                   14357: machines. We use explicit register declarations (@pxref{Explicit Reg
                   14358: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
                   14359: improve the speed on some machines. They are turned on by using the
                   14360: configuration flag @code{--enable-force-reg} (@code{gcc} switch
                   14361: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
                   14362: machine, but also on the compiler version: On some machines some
                   14363: compiler versions produce incorrect code when certain explicit register
                   14364: declarations are used. So by default @code{-DFORCE_REG} is not used.
                   14365: 
                   14366: @node Threading, Primitives, Portability, Engine
                   14367: @section Threading
                   14368: @cindex inner interpreter implementation
                   14369: @cindex threaded code implementation
                   14370: 
                   14371: @cindex labels as values
                   14372: GNU C's labels as values extension (available since @code{gcc-2.0},
                   14373: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
1.29      crook    14374: makes it possible to take the address of @i{label} by writing
                   14375: @code{&&@i{label}}.  This address can then be used in a statement like
                   14376: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
1.1       anton    14377: @code{goto x}.
                   14378: 
1.26      crook    14379: @cindex @code{NEXT}, indirect threaded
1.1       anton    14380: @cindex indirect threaded inner interpreter
                   14381: @cindex inner interpreter, indirect threaded
1.26      crook    14382: With this feature an indirect threaded @code{NEXT} looks like:
1.1       anton    14383: @example
                   14384: cfa = *ip++;
                   14385: ca = *cfa;
                   14386: goto *ca;
                   14387: @end example
                   14388: @cindex instruction pointer
                   14389: For those unfamiliar with the names: @code{ip} is the Forth instruction
                   14390: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
                   14391: execution token and points to the code field of the next word to be
                   14392: executed; The @code{ca} (code address) fetched from there points to some
                   14393: executable code, e.g., a primitive or the colon definition handler
                   14394: @code{docol}.
                   14395: 
1.26      crook    14396: @cindex @code{NEXT}, direct threaded
1.1       anton    14397: @cindex direct threaded inner interpreter
                   14398: @cindex inner interpreter, direct threaded
                   14399: Direct threading is even simpler:
                   14400: @example
                   14401: ca = *ip++;
                   14402: goto *ca;
                   14403: @end example
                   14404: 
                   14405: Of course we have packaged the whole thing neatly in macros called
1.26      crook    14406: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
1.1       anton    14407: 
                   14408: @menu
                   14409: * Scheduling::                  
                   14410: * Direct or Indirect Threaded?::  
                   14411: * DOES>::                       
                   14412: @end menu
                   14413: 
                   14414: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
                   14415: @subsection Scheduling
                   14416: @cindex inner interpreter optimization
                   14417: 
                   14418: There is a little complication: Pipelined and superscalar processors,
                   14419: i.e., RISC and some modern CISC machines can process independent
                   14420: instructions while waiting for the results of an instruction. The
                   14421: compiler usually reorders (schedules) the instructions in a way that
                   14422: achieves good usage of these delay slots. However, on our first tries
                   14423: the compiler did not do well on scheduling primitives. E.g., for
                   14424: @code{+} implemented as
                   14425: @example
                   14426: n=sp[0]+sp[1];
                   14427: sp++;
                   14428: sp[0]=n;
                   14429: NEXT;
                   14430: @end example
1.81      anton    14431: the @code{NEXT} comes strictly after the other code, i.e., there is
                   14432: nearly no scheduling. After a little thought the problem becomes clear:
                   14433: The compiler cannot know that @code{sp} and @code{ip} point to different
1.21      crook    14434: addresses (and the version of @code{gcc} we used would not know it even
                   14435: if it was possible), so it could not move the load of the cfa above the
                   14436: store to the TOS. Indeed the pointers could be the same, if code on or
                   14437: very near the top of stack were executed. In the interest of speed we
                   14438: chose to forbid this probably unused ``feature'' and helped the compiler
1.81      anton    14439: in scheduling: @code{NEXT} is divided into several parts:
                   14440: @code{NEXT_P0}, @code{NEXT_P1} and @code{NEXT_P2}). @code{+} now looks
                   14441: like:
1.1       anton    14442: @example
1.81      anton    14443: NEXT_P0;
1.1       anton    14444: n=sp[0]+sp[1];
                   14445: sp++;
                   14446: NEXT_P1;
                   14447: sp[0]=n;
                   14448: NEXT_P2;
                   14449: @end example
                   14450: 
1.81      anton    14451: There are various schemes that distribute the different operations of
                   14452: NEXT between these parts in several ways; in general, different schemes
                   14453: perform best on different processors.  We use a scheme for most
                   14454: architectures that performs well for most processors of this
                   14455: architecture; in the furture we may switch to benchmarking and chosing
                   14456: the scheme on installation time.
                   14457: 
1.1       anton    14458: 
                   14459: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
                   14460: @subsection Direct or Indirect Threaded?
                   14461: @cindex threading, direct or indirect?
                   14462: 
                   14463: @cindex -DDIRECT_THREADED
                   14464: Both! After packaging the nasty details in macro definitions we
                   14465: realized that we could switch between direct and indirect threading by
                   14466: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
                   14467: defining a few machine-specific macros for the direct-threading case.
                   14468: On the Forth level we also offer access words that hide the
                   14469: differences between the threading methods (@pxref{Threading Words}).
                   14470: 
                   14471: Indirect threading is implemented completely machine-independently.
                   14472: Direct threading needs routines for creating jumps to the executable
1.21      crook    14473: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
                   14474: machine-dependent, but they do not amount to many source lines. Therefore,
                   14475: even porting direct threading to a new machine requires little effort.
1.1       anton    14476: 
                   14477: @cindex --enable-indirect-threaded, configuration flag
                   14478: @cindex --enable-direct-threaded, configuration flag
                   14479: The default threading method is machine-dependent. You can enforce a
                   14480: specific threading method when building Gforth with the configuration
                   14481: flag @code{--enable-direct-threaded} or
                   14482: @code{--enable-indirect-threaded}. Note that direct threading is not
                   14483: supported on all machines.
                   14484: 
                   14485: @node DOES>,  , Direct or Indirect Threaded?, Threading
                   14486: @subsection DOES>
                   14487: @cindex @code{DOES>} implementation
                   14488: 
1.26      crook    14489: @cindex @code{dodoes} routine
                   14490: @cindex @code{DOES>}-code
1.1       anton    14491: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
                   14492: the chunk of code executed by every word defined by a
                   14493: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
                   14494: the Forth code to be executed, i.e. the code after the
1.26      crook    14495: @code{DOES>} (the @code{DOES>}-code)? There are two solutions:
1.1       anton    14496: 
1.21      crook    14497: In fig-Forth the code field points directly to the @code{dodoes} and the
1.45      crook    14498: @code{DOES>}-code address is stored in the cell after the code address (i.e. at
1.29      crook    14499: @code{@i{CFA} cell+}). It may seem that this solution is illegal in
1.1       anton    14500: the Forth-79 and all later standards, because in fig-Forth this address
                   14501: lies in the body (which is illegal in these standards). However, by
                   14502: making the code field larger for all words this solution becomes legal
                   14503: again. We use this approach for the indirect threaded version and for
                   14504: direct threading on some machines. Leaving a cell unused in most words
                   14505: is a bit wasteful, but on the machines we are targeting this is hardly a
                   14506: problem. The other reason for having a code field size of two cells is
                   14507: to avoid having different image files for direct and indirect threaded
                   14508: systems (direct threaded systems require two-cell code fields on many
                   14509: machines).
                   14510: 
1.26      crook    14511: @cindex @code{DOES>}-handler
1.1       anton    14512: The other approach is that the code field points or jumps to the cell
1.26      crook    14513: after @code{DOES>}. In this variant there is a jump to @code{dodoes} at
                   14514: this address (the @code{DOES>}-handler). @code{dodoes} can then get the
                   14515: @code{DOES>}-code address by computing the code address, i.e., the address of
1.45      crook    14516: the jump to @code{dodoes}, and add the length of that jump field. A variant of
1.1       anton    14517: this is to have a call to @code{dodoes} after the @code{DOES>}; then the
                   14518: return address (which can be found in the return register on RISCs) is
1.26      crook    14519: the @code{DOES>}-code address. Since the two cells available in the code field
1.1       anton    14520: are used up by the jump to the code address in direct threading on many
                   14521: architectures, we use this approach for direct threading on these
                   14522: architectures. We did not want to add another cell to the code field.
                   14523: 
                   14524: @node Primitives, Performance, Threading, Engine
                   14525: @section Primitives
                   14526: @cindex primitives, implementation
                   14527: @cindex virtual machine instructions, implementation
                   14528: 
                   14529: @menu
                   14530: * Automatic Generation::        
                   14531: * TOS Optimization::            
                   14532: * Produced code::               
                   14533: @end menu
                   14534: 
                   14535: @node Automatic Generation, TOS Optimization, Primitives, Primitives
                   14536: @subsection Automatic Generation
                   14537: @cindex primitives, automatic generation
                   14538: 
                   14539: @cindex @file{prims2x.fs}
                   14540: Since the primitives are implemented in a portable language, there is no
                   14541: longer any need to minimize the number of primitives. On the contrary,
                   14542: having many primitives has an advantage: speed. In order to reduce the
                   14543: number of errors in primitives and to make programming them easier, we
                   14544: provide a tool, the primitive generator (@file{prims2x.fs}), that
                   14545: automatically generates most (and sometimes all) of the C code for a
                   14546: primitive from the stack effect notation.  The source for a primitive
                   14547: has the following form:
                   14548: 
                   14549: @cindex primitive source format
                   14550: @format
1.58      anton    14551: @i{Forth-name}  ( @i{stack-effect} )        @i{category}    [@i{pronounc.}]
1.29      crook    14552: [@code{""}@i{glossary entry}@code{""}]
                   14553: @i{C code}
1.1       anton    14554: [@code{:}
1.29      crook    14555: @i{Forth code}]
1.1       anton    14556: @end format
                   14557: 
                   14558: The items in brackets are optional. The category and glossary fields
                   14559: are there for generating the documentation, the Forth code is there
                   14560: for manual implementations on machines without GNU C. E.g., the source
                   14561: for the primitive @code{+} is:
                   14562: @example
1.58      anton    14563: +    ( n1 n2 -- n )   core    plus
1.1       anton    14564: n = n1+n2;
                   14565: @end example
                   14566: 
                   14567: This looks like a specification, but in fact @code{n = n1+n2} is C
                   14568: code. Our primitive generation tool extracts a lot of information from
                   14569: the stack effect notations@footnote{We use a one-stack notation, even
                   14570: though we have separate data and floating-point stacks; The separate
                   14571: notation can be generated easily from the unified notation.}: The number
                   14572: of items popped from and pushed on the stack, their type, and by what
                   14573: name they are referred to in the C code. It then generates a C code
                   14574: prelude and postlude for each primitive. The final C code for @code{+}
                   14575: looks like this:
                   14576: 
                   14577: @example
1.46      pazsan   14578: I_plus: /* + ( n1 n2 -- n ) */  /* label, stack effect */
1.1       anton    14579: /*  */                          /* documentation */
1.81      anton    14580: NAME("+")                       /* debugging output (with -DDEBUG) */
1.1       anton    14581: @{
                   14582: DEF_CA                          /* definition of variable ca (indirect threading) */
                   14583: Cell n1;                        /* definitions of variables */
                   14584: Cell n2;
                   14585: Cell n;
1.81      anton    14586: NEXT_P0;                        /* NEXT part 0 */
1.1       anton    14587: n1 = (Cell) sp[1];              /* input */
                   14588: n2 = (Cell) TOS;
                   14589: sp += 1;                        /* stack adjustment */
                   14590: @{
                   14591: n = n1+n2;                      /* C code taken from the source */
                   14592: @}
                   14593: NEXT_P1;                        /* NEXT part 1 */
                   14594: TOS = (Cell)n;                  /* output */
                   14595: NEXT_P2;                        /* NEXT part 2 */
                   14596: @}
                   14597: @end example
                   14598: 
                   14599: This looks long and inefficient, but the GNU C compiler optimizes quite
                   14600: well and produces optimal code for @code{+} on, e.g., the R3000 and the
                   14601: HP RISC machines: Defining the @code{n}s does not produce any code, and
                   14602: using them as intermediate storage also adds no cost.
                   14603: 
1.26      crook    14604: There are also other optimizations that are not illustrated by this
                   14605: example: assignments between simple variables are usually for free (copy
1.1       anton    14606: propagation). If one of the stack items is not used by the primitive
                   14607: (e.g.  in @code{drop}), the compiler eliminates the load from the stack
                   14608: (dead code elimination). On the other hand, there are some things that
                   14609: the compiler does not do, therefore they are performed by
                   14610: @file{prims2x.fs}: The compiler does not optimize code away that stores
                   14611: a stack item to the place where it just came from (e.g., @code{over}).
                   14612: 
                   14613: While programming a primitive is usually easy, there are a few cases
                   14614: where the programmer has to take the actions of the generator into
                   14615: account, most notably @code{?dup}, but also words that do not (always)
1.26      crook    14616: fall through to @code{NEXT}.
1.1       anton    14617: 
                   14618: @node TOS Optimization, Produced code, Automatic Generation, Primitives
                   14619: @subsection TOS Optimization
                   14620: @cindex TOS optimization for primitives
                   14621: @cindex primitives, keeping the TOS in a register
                   14622: 
                   14623: An important optimization for stack machine emulators, e.g., Forth
                   14624: engines, is keeping  one or more of the top stack items in
1.29      crook    14625: registers.  If a word has the stack effect @i{in1}...@i{inx} @code{--}
                   14626: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
1.1       anton    14627: @itemize @bullet
                   14628: @item
1.29      crook    14629: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
1.1       anton    14630: due to fewer loads from and stores to the stack.
1.29      crook    14631: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
                   14632: @i{y<n}, due to additional moves between registers.
1.1       anton    14633: @end itemize
                   14634: 
                   14635: @cindex -DUSE_TOS
                   14636: @cindex -DUSE_NO_TOS
                   14637: In particular, keeping one item in a register is never a disadvantage,
                   14638: if there are enough registers. Keeping two items in registers is a
                   14639: disadvantage for frequent words like @code{?branch}, constants,
                   14640: variables, literals and @code{i}. Therefore our generator only produces
                   14641: code that keeps zero or one items in registers. The generated C code
                   14642: covers both cases; the selection between these alternatives is made at
                   14643: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
                   14644: code for @code{+} is just a simple variable name in the one-item case,
                   14645: otherwise it is a macro that expands into @code{sp[0]}. Note that the
                   14646: GNU C compiler tries to keep simple variables like @code{TOS} in
                   14647: registers, and it usually succeeds, if there are enough registers.
                   14648: 
                   14649: @cindex -DUSE_FTOS
                   14650: @cindex -DUSE_NO_FTOS
                   14651: The primitive generator performs the TOS optimization for the
                   14652: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
                   14653: operations the benefit of this optimization is even larger:
                   14654: floating-point operations take quite long on most processors, but can be
                   14655: performed in parallel with other operations as long as their results are
                   14656: not used. If the FP-TOS is kept in a register, this works. If
                   14657: it is kept on the stack, i.e., in memory, the store into memory has to
                   14658: wait for the result of the floating-point operation, lengthening the
                   14659: execution time of the primitive considerably.
                   14660: 
                   14661: The TOS optimization makes the automatic generation of primitives a
                   14662: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
                   14663: @code{TOS} is not sufficient. There are some special cases to
                   14664: consider:
                   14665: @itemize @bullet
                   14666: @item In the case of @code{dup ( w -- w w )} the generator must not
                   14667: eliminate the store to the original location of the item on the stack,
                   14668: if the TOS optimization is turned on.
                   14669: @item Primitives with stack effects of the form @code{--}
1.29      crook    14670: @i{out1}...@i{outy} must store the TOS to the stack at the start.
                   14671: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
1.1       anton    14672: must load the TOS from the stack at the end. But for the null stack
                   14673: effect @code{--} no stores or loads should be generated.
                   14674: @end itemize
                   14675: 
                   14676: @node Produced code,  , TOS Optimization, Primitives
                   14677: @subsection Produced code
                   14678: @cindex primitives, assembly code listing
                   14679: 
                   14680: @cindex @file{engine.s}
                   14681: To see what assembly code is produced for the primitives on your machine
                   14682: with your compiler and your flag settings, type @code{make engine.s} and
1.81      anton    14683: look at the resulting file @file{engine.s}.  Alternatively, you can also
                   14684: disassemble the code of primitives with @code{see} on some architectures.
1.1       anton    14685: 
                   14686: @node  Performance,  , Primitives, Engine
                   14687: @section Performance
                   14688: @cindex performance of some Forth interpreters
                   14689: @cindex engine performance
                   14690: @cindex benchmarking Forth systems
                   14691: @cindex Gforth performance
                   14692: 
                   14693: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
                   14694: impossible to write a significantly faster engine.
                   14695: 
                   14696: On register-starved machines like the 386 architecture processors
                   14697: improvements are possible, because @code{gcc} does not utilize the
                   14698: registers as well as a human, even with explicit register declarations;
                   14699: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
                   14700: and hand-tuned it for the 486; this system is 1.19 times faster on the
                   14701: Sieve benchmark on a 486DX2/66 than Gforth compiled with
1.40      anton    14702: @code{gcc-2.6.3} with @code{-DFORCE_REG}.  The situation has improved
                   14703: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
                   14704: registers fit in real registers (and we can even afford to use the TOS
                   14705: optimization), resulting in a speedup of 1.14 on the sieve over the
                   14706: earlier results.
1.1       anton    14707: 
                   14708: @cindex Win32Forth performance
                   14709: @cindex NT Forth performance
                   14710: @cindex eforth performance
                   14711: @cindex ThisForth performance
                   14712: @cindex PFE performance
                   14713: @cindex TILE performance
1.81      anton    14714: The potential advantage of assembly language implementations is not
                   14715: necessarily realized in complete Forth systems: We compared Gforth-0.4.9
                   14716: (direct threaded, compiled with @code{gcc-2.95.1} and
                   14717: @code{-DFORCE_REG}) with Win32Forth 1.2093 (newer versions are
                   14718: reportedly much faster), LMI's NT Forth (Beta, May 1994) and Eforth
                   14719: (with and without peephole (aka pinhole) optimization of the threaded
                   14720: code); all these systems were written in assembly language. We also
                   14721: compared Gforth with three systems written in C: PFE-0.9.14 (compiled
                   14722: with @code{gcc-2.6.3} with the default configuration for Linux:
                   14723: @code{-O2 -fomit-frame-pointer -DUSE_REGS -DUNROLL_NEXT}), ThisForth
                   14724: Beta (compiled with @code{gcc-2.6.3 -O3 -fomit-frame-pointer}; ThisForth
                   14725: employs peephole optimization of the threaded code) and TILE (compiled
                   14726: with @code{make opt}). We benchmarked Gforth, PFE, ThisForth and TILE on
                   14727: a 486DX2/66 under Linux. Kenneth O'Heskin kindly provided the results
                   14728: for Win32Forth and NT Forth on a 486DX2/66 with similar memory
                   14729: performance under Windows NT. Marcel Hendrix ported Eforth to Linux,
                   14730: then extended it to run the benchmarks, added the peephole optimizer,
                   14731: ran the benchmarks and reported the results.
1.40      anton    14732: 
1.1       anton    14733: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
                   14734: matrix multiplication come from the Stanford integer benchmarks and have
                   14735: been translated into Forth by Martin Fraeman; we used the versions
                   14736: included in the TILE Forth package, but with bigger data set sizes; and
                   14737: a recursive Fibonacci number computation for benchmarking calling
                   14738: performance. The following table shows the time taken for the benchmarks
                   14739: scaled by the time taken by Gforth (in other words, it shows the speedup
                   14740: factor that Gforth achieved over the other systems).
                   14741: 
                   14742: @example
1.40      anton    14743: relative      Win32-    NT       eforth       This-      
1.1       anton    14744:   time  Gforth Forth Forth eforth  +opt   PFE Forth  TILE
1.81      anton    14745: sieve     1.00  1.60  1.32   1.60  0.98  1.82  3.67  9.91
                   14746: bubble    1.00  1.55  1.66   1.75  1.04  1.78        4.58
                   14747: matmul    1.00  1.71  1.57   1.69  0.86  1.83        4.74
                   14748: fib       1.00  1.76  1.54   1.41  1.00  2.01  3.45  4.96
1.1       anton    14749: @end example
                   14750: 
1.26      crook    14751: You may be quite surprised by the good performance of Gforth when
                   14752: compared with systems written in assembly language. One important reason
                   14753: for the disappointing performance of these other systems is probably
                   14754: that they are not written optimally for the 486 (e.g., they use the
                   14755: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
                   14756: but costly method for relocating the Forth image: like @code{cforth}, it
                   14757: computes the actual addresses at run time, resulting in two address
                   14758: computations per @code{NEXT} (@pxref{Image File Background}).
                   14759: 
1.40      anton    14760: Only Eforth with the peephole optimizer performs comparable to
                   14761: Gforth. The speedups achieved with peephole optimization of threaded
                   14762: code are quite remarkable. Adding a peephole optimizer to Gforth should
                   14763: cause similar speedups.
1.1       anton    14764: 
                   14765: The speedup of Gforth over PFE, ThisForth and TILE can be easily
                   14766: explained with the self-imposed restriction of the latter systems to
                   14767: standard C, which makes efficient threading impossible (however, the
1.4       anton    14768: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
1.1       anton    14769: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
                   14770: Moreover, current C compilers have a hard time optimizing other aspects
                   14771: of the ThisForth and the TILE source.
                   14772: 
1.26      crook    14773: The performance of Gforth on 386 architecture processors varies widely
                   14774: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
                   14775: allocate any of the virtual machine registers into real machine
                   14776: registers by itself and would not work correctly with explicit register
1.40      anton    14777: declarations, giving a 1.5 times slower engine (on a 486DX2/66 running
1.26      crook    14778: the Sieve) than the one measured above.
1.1       anton    14779: 
1.26      crook    14780: Note that there have been several releases of Win32Forth since the
                   14781: release presented here, so the results presented above may have little
1.40      anton    14782: predictive value for the performance of Win32Forth today (results for
                   14783: the current release on an i486DX2/66 are welcome).
1.1       anton    14784: 
                   14785: @cindex @file{Benchres}
1.66      anton    14786: In
                   14787: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
                   14788: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
1.1       anton    14789: Maierhofer (presented at EuroForth '95), an indirect threaded version of
1.66      anton    14790: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
                   14791: several native code systems; that version of Gforth is slower on a 486
                   14792: than the direct threaded version used here. You can find a newer version
                   14793: of these measurements at
1.47      crook    14794: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
1.1       anton    14795: find numbers for Gforth on various machines in @file{Benchres}.
                   14796: 
1.26      crook    14797: @c ******************************************************************
1.13      pazsan   14798: @node Binding to System Library, Cross Compiler, Engine, Top
1.14      pazsan   14799: @chapter Binding to System Library
1.13      pazsan   14800: 
                   14801: @node Cross Compiler, Bugs, Binding to System Library, Top
1.14      pazsan   14802: @chapter Cross Compiler
1.47      crook    14803: @cindex @file{cross.fs}
                   14804: @cindex cross-compiler
                   14805: @cindex metacompiler
                   14806: @cindex target compiler
1.13      pazsan   14807: 
1.46      pazsan   14808: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
                   14809: mostly written in Forth, including crucial parts like the outer
                   14810: interpreter and compiler, it needs compiled Forth code to get
                   14811: started. The cross compiler allows to create new images for other
                   14812: architectures, even running under another Forth system.
1.13      pazsan   14813: 
                   14814: @menu
1.67      anton    14815: * Using the Cross Compiler::    
                   14816: * How the Cross Compiler Works::  
1.13      pazsan   14817: @end menu
                   14818: 
1.21      crook    14819: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
1.14      pazsan   14820: @section Using the Cross Compiler
1.46      pazsan   14821: 
                   14822: The cross compiler uses a language that resembles Forth, but isn't. The
                   14823: main difference is that you can execute Forth code after definition,
                   14824: while you usually can't execute the code compiled by cross, because the
                   14825: code you are compiling is typically for a different computer than the
                   14826: one you are compiling on.
                   14827: 
1.81      anton    14828: @c anton: This chapter is somewhat different from waht I would expect: I
                   14829: @c would expect an explanation of the cross language and how to create an
                   14830: @c application image with it.  The section explains some aspects of
                   14831: @c creating a Gforth kernel.
                   14832: 
1.46      pazsan   14833: The Makefile is already set up to allow you to create kernels for new
                   14834: architectures with a simple make command. The generic kernels using the
                   14835: GCC compiled virtual machine are created in the normal build process
                   14836: with @code{make}. To create a embedded Gforth executable for e.g. the
                   14837: 8086 processor (running on a DOS machine), type
                   14838: 
                   14839: @example
                   14840: make kernl-8086.fi
                   14841: @end example
                   14842: 
                   14843: This will use the machine description from the @file{arch/8086}
                   14844: directory to create a new kernel. A machine file may look like that:
                   14845: 
                   14846: @example
                   14847: \ Parameter for target systems                         06oct92py
                   14848: 
                   14849:     4 Constant cell             \ cell size in bytes
                   14850:     2 Constant cell<<           \ cell shift to bytes
                   14851:     5 Constant cell>bit         \ cell shift to bits
                   14852:     8 Constant bits/char        \ bits per character
                   14853:     8 Constant bits/byte        \ bits per byte [default: 8]
                   14854:     8 Constant float            \ bytes per float
                   14855:     8 Constant /maxalign        \ maximum alignment in bytes
                   14856: false Constant bigendian        \ byte order
                   14857: ( true=big, false=little )
                   14858: 
                   14859: include machpc.fs               \ feature list
                   14860: @end example
                   14861: 
                   14862: This part is obligatory for the cross compiler itself, the feature list
                   14863: is used by the kernel to conditionally compile some features in and out,
                   14864: depending on whether the target supports these features.
                   14865: 
                   14866: There are some optional features, if you define your own primitives,
                   14867: have an assembler, or need special, nonstandard preparation to make the
1.81      anton    14868: boot process work. @code{asm-include} includes an assembler,
1.46      pazsan   14869: @code{prims-include} includes primitives, and @code{>boot} prepares for
                   14870: booting.
                   14871: 
                   14872: @example
                   14873: : asm-include    ." Include assembler" cr
                   14874:   s" arch/8086/asm.fs" included ;
                   14875: 
                   14876: : prims-include  ." Include primitives" cr
                   14877:   s" arch/8086/prim.fs" included ;
                   14878: 
                   14879: : >boot          ." Prepare booting" cr
                   14880:   s" ' boot >body into-forth 1+ !" evaluate ;
                   14881: @end example
                   14882: 
                   14883: These words are used as sort of macro during the cross compilation in
1.81      anton    14884: the file @file{kernel/main.fs}. Instead of using these macros, it would
1.46      pazsan   14885: be possible --- but more complicated --- to write a new kernel project
                   14886: file, too.
                   14887: 
                   14888: @file{kernel/main.fs} expects the machine description file name on the
                   14889: stack; the cross compiler itself (@file{cross.fs}) assumes that either
                   14890: @code{mach-file} leaves a counted string on the stack, or
                   14891: @code{machine-file} leaves an address, count pair of the filename on the
                   14892: stack.
                   14893: 
                   14894: The feature list is typically controlled using @code{SetValue}, generic
                   14895: files that are used by several projects can use @code{DefaultValue}
                   14896: instead. Both functions work like @code{Value}, when the value isn't
                   14897: defined, but @code{SetValue} works like @code{to} if the value is
                   14898: defined, and @code{DefaultValue} doesn't set anything, if the value is
                   14899: defined.
                   14900: 
                   14901: @example
                   14902: \ generic mach file for pc gforth                       03sep97jaw
                   14903: 
                   14904: true DefaultValue NIL  \ relocating
                   14905: 
                   14906: >ENVIRON
                   14907: 
                   14908: true DefaultValue file          \ controls the presence of the
                   14909:                                 \ file access wordset
                   14910: true DefaultValue OS            \ flag to indicate a operating system
                   14911: 
                   14912: true DefaultValue prims         \ true: primitives are c-code
                   14913: 
                   14914: true DefaultValue floating      \ floating point wordset is present
                   14915: 
                   14916: true DefaultValue glocals       \ gforth locals are present
                   14917:                                 \ will be loaded
                   14918: true DefaultValue dcomps        \ double number comparisons
                   14919: 
                   14920: true DefaultValue hash          \ hashing primitives are loaded/present
                   14921: 
                   14922: true DefaultValue xconds        \ used together with glocals,
                   14923:                                 \ special conditionals supporting gforths'
                   14924:                                 \ local variables
                   14925: true DefaultValue header        \ save a header information
                   14926: 
                   14927: true DefaultValue backtrace     \ enables backtrace code
                   14928: 
                   14929: false DefaultValue ec
                   14930: false DefaultValue crlf
                   14931: 
                   14932: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
                   14933: 
                   14934: &16 KB          DefaultValue stack-size
                   14935: &15 KB &512 +   DefaultValue fstack-size
                   14936: &15 KB          DefaultValue rstack-size
                   14937: &14 KB &512 +   DefaultValue lstack-size
                   14938: @end example
1.13      pazsan   14939: 
1.48      anton    14940: @node How the Cross Compiler Works,  , Using the Cross Compiler, Cross Compiler
1.14      pazsan   14941: @section How the Cross Compiler Works
1.13      pazsan   14942: 
                   14943: @node Bugs, Origin, Cross Compiler, Top
1.21      crook    14944: @appendix Bugs
1.1       anton    14945: @cindex bug reporting
                   14946: 
1.21      crook    14947: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
1.1       anton    14948: 
                   14949: If you find a bug, please send a bug report to
1.33      anton    14950: @email{bug-gforth@@gnu.org}. A bug report should include this
1.21      crook    14951: information:
                   14952: 
                   14953: @itemize @bullet
                   14954: @item
1.81      anton    14955: A program (or a sequence of keyboard commands) that reproduces the bug.
                   14956: @item
                   14957: A description of what you think constitutes the buggy behaviour.
                   14958: @item
1.21      crook    14959: The Gforth version used (it is announced at the start of an
                   14960: interactive Gforth session).
                   14961: @item
                   14962: The machine and operating system (on Unix
                   14963: systems @code{uname -a} will report this information).
                   14964: @item
1.81      anton    14965: The installation options (you can find the configure options at the
                   14966: start of @file{config.status}) and configuration (@code{configure}
                   14967: output or @file{config.cache}).
1.21      crook    14968: @item
                   14969: A complete list of changes (if any) you (or your installer) have made to the
                   14970: Gforth sources.
                   14971: @end itemize
1.1       anton    14972: 
                   14973: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
                   14974: to Report Bugs, gcc.info, GNU C Manual}.
                   14975: 
                   14976: 
1.21      crook    14977: @node Origin, Forth-related information, Bugs, Top
                   14978: @appendix Authors and Ancestors of Gforth
1.1       anton    14979: 
                   14980: @section Authors and Contributors
                   14981: @cindex authors of Gforth
                   14982: @cindex contributors to Gforth
                   14983: 
                   14984: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
1.81      anton    14985: Ertl. The third major author was Jens Wilke.  Neal Crook contributed a
                   14986: lot to the manual.  Assemblers and disassemblers were contributed by
                   14987: Andrew McKewan, Christian Pirker, and Bernd Thallner.  Lennart Benschop
                   14988: (who was one of Gforth's first users, in mid-1993) and Stuart Ramsden
                   14989: inspired us with their continuous feedback. Lennart Benshop contributed
1.1       anton    14990: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
                   14991: support for calling C libraries. Helpful comments also came from Paul
                   14992: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
1.58      anton    14993: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, and
                   14994: Robert Epprecht. Since the release of Gforth-0.2.1 there were also
                   14995: helpful comments from many others; thank you all, sorry for not listing
                   14996: you here (but digging through my mailbox to extract your names is on my
1.81      anton    14997: to-do list).
1.1       anton    14998: 
                   14999: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
                   15000: and autoconf, among others), and to the creators of the Internet: Gforth
1.21      crook    15001: was developed across the Internet, and its authors did not meet
1.20      pazsan   15002: physically for the first 4 years of development.
1.1       anton    15003: 
                   15004: @section Pedigree
1.26      crook    15005: @cindex pedigree of Gforth
1.1       anton    15006: 
1.81      anton    15007: Gforth descends from bigFORTH (1993) and fig-Forth.  Of course, a
                   15008: significant part of the design of Gforth was prescribed by ANS Forth.
1.1       anton    15009: 
1.20      pazsan   15010: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
1.1       anton    15011: 32 bit native code version of VolksForth for the Atari ST, written
                   15012: mostly by Dietrich Weineck.
                   15013: 
1.81      anton    15014: VolksForth was written by Klaus Schleisiek, Bernd Pennemann, Georg
                   15015: Rehfeld and Dietrich Weineck for the C64 (called UltraForth there) in
                   15016: the mid-80s and ported to the Atari ST in 1986.  It descends from F83.
1.1       anton    15017: 
                   15018: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
                   15019: Forth-83 standard. !! Pedigree? When?
                   15020: 
                   15021: A team led by Bill Ragsdale implemented fig-Forth on many processors in
                   15022: 1979. Robert Selzer and Bill Ragsdale developed the original
                   15023: implementation of fig-Forth for the 6502 based on microForth.
                   15024: 
                   15025: The principal architect of microForth was Dean Sanderson. microForth was
                   15026: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
                   15027: the 1802, and subsequently implemented on the 8080, the 6800 and the
                   15028: Z80.
                   15029: 
                   15030: All earlier Forth systems were custom-made, usually by Charles Moore,
                   15031: who discovered (as he puts it) Forth during the late 60s. The first full
                   15032: Forth existed in 1971.
                   15033: 
1.81      anton    15034: A part of the information in this section comes from
                   15035: @cite{@uref{http://www.forth.com/Content/History/History1.htm,The
                   15036: Evolution of Forth}} by Elizabeth D. Rather, Donald R. Colburn and
                   15037: Charles H. Moore, presented at the HOPL-II conference and preprinted in
                   15038: SIGPLAN Notices 28(3), 1993.  You can find more historical and
                   15039: genealogical information about Forth there.
1.1       anton    15040: 
1.81      anton    15041: @c ------------------------------------------------------------------
1.21      crook    15042: @node Forth-related information, Word Index, Origin, Top
                   15043: @appendix Other Forth-related information
                   15044: @cindex Forth-related information
                   15045: 
1.81      anton    15046: @c anton: I threw most of this stuff out, because it can be found through
                   15047: @c the FAQ and the FAQ is more likely to be up-to-date.
1.21      crook    15048: 
                   15049: @cindex comp.lang.forth
                   15050: @cindex frequently asked questions
1.81      anton    15051: There is an active news group (comp.lang.forth) discussing Forth
                   15052: (including Gforth) and Forth-related issues. Its
                   15053: @uref{http://www.complang.tuwien.ac.at/forth/faq/faq-general-2.html,FAQs}
                   15054: (frequently asked questions and their answers) contains a lot of
                   15055: information on Forth.  You should read it before posting to
                   15056: comp.lang.forth.
1.21      crook    15057: 
1.81      anton    15058: The ANS Forth standard is most usable in its
                   15059: @uref{http://www.taygeta.com/forth/dpans.html, HTML form}.
1.21      crook    15060: 
1.81      anton    15061: @c ------------------------------------------------------------------
                   15062: @node Word Index, Concept Index, Forth-related information, Top
1.1       anton    15063: @unnumbered Word Index
                   15064: 
1.26      crook    15065: This index is a list of Forth words that have ``glossary'' entries
                   15066: within this manual. Each word is listed with its stack effect and
                   15067: wordset.
1.1       anton    15068: 
                   15069: @printindex fn
                   15070: 
1.81      anton    15071: @c anton: the name index seems superfluous given the word and concept indices.
                   15072: 
                   15073: @c @node Name Index, Concept Index, Word Index, Top
                   15074: @c @unnumbered Name Index
1.41      anton    15075: 
1.81      anton    15076: @c This index is a list of Forth words that have ``glossary'' entries
                   15077: @c within this manual.
1.41      anton    15078: 
1.81      anton    15079: @c @printindex ky
1.41      anton    15080: 
1.81      anton    15081: @node Concept Index,  , Word Index, Top
1.1       anton    15082: @unnumbered Concept and Word Index
                   15083: 
1.26      crook    15084: Not all entries listed in this index are present verbatim in the
                   15085: text. This index also duplicates, in abbreviated form, all of the words
                   15086: listed in the Word Index (only the names are listed for the words here).
1.1       anton    15087: 
                   15088: @printindex cp
                   15089: 
                   15090: @contents
                   15091: @bye
1.81      anton    15092: 
                   15093: 
1.1       anton    15094: 

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