Annotation of gforth/doc/gforth.ds, revision 1.181
1.1 anton 1: \input texinfo @c -*-texinfo-*-
2: @comment The source is gforth.ds, from which gforth.texi is generated
1.28 crook 3:
1.21 crook 4: @comment TODO: nac29jan99 - a list of things to add in the next edit:
1.28 crook 5: @comment 1. x-ref all ambiguous or implementation-defined features?
6: @comment 2. Describe the use of Auser Avariable AConstant A, etc.
7: @comment 3. words in miscellaneous section need a home.
8: @comment 4. search for TODO for other minor and major works required.
9: @comment 5. [rats] change all @var to @i in Forth source so that info
10: @comment file looks decent.
1.36 anton 11: @c Not an improvement IMO - anton
12: @c and anyway, this should be taken up
13: @c with Karl Berry (the texinfo guy) - anton
1.113 anton 14: @c
15: @c Karl Berry writes:
16: @c If they don't like the all-caps for @var Info output, all I can say is
17: @c that it's always been that way, and the usage of all-caps for
18: @c metavariables has a long tradition. I think it's best to just let it be
19: @c what it is, for the sake of consistency among manuals.
20: @c
1.29 crook 21: @comment .. would be useful to have a word that identified all deferred words
22: @comment should semantics stuff in intro be moved to another section
23:
1.66 anton 24: @c POSTPONE, COMPILE, [COMPILE], LITERAL should have their own section
1.28 crook 25:
1.1 anton 26: @comment %**start of header (This is for running Texinfo on a region.)
27: @setfilename gforth.info
1.113 anton 28: @include version.texi
1.1 anton 29: @settitle Gforth Manual
1.113 anton 30: @c @syncodeindex pg cp
1.49 anton 31:
1.12 anton 32: @macro progstyle {}
33: Programming style note:
1.3 anton 34: @end macro
1.48 anton 35:
36: @macro assignment {}
37: @table @i
38: @item Assignment:
39: @end macro
40: @macro endassignment {}
41: @end table
42: @end macro
43:
1.29 crook 44: @comment macros for beautifying glossary entries
45: @macro GLOSS-START {}
46: @iftex
47: @ninerm
48: @end iftex
49: @end macro
50:
51: @macro GLOSS-END {}
52: @iftex
53: @rm
54: @end iftex
55: @end macro
56:
1.113 anton 57: @comment %**end of header (This is for running Texinfo on a region.)
58: @copying
1.125 anton 59: This manual is for Gforth (version @value{VERSION}, @value{UPDATED}),
60: a fast and portable implementation of the ANS Forth language. It
61: serves as reference manual, but it also contains an introduction to
62: Forth and a Forth tutorial.
1.29 crook 63:
1.169 anton 64: Copyright @copyright{} 1995, 1996, 1997, 1998, 2000, 2003, 2004,2005,2006 Free Software Foundation, Inc.
1.29 crook 65:
1.113 anton 66: @quotation
67: Permission is granted to copy, distribute and/or modify this document
68: under the terms of the GNU Free Documentation License, Version 1.1 or
69: any later version published by the Free Software Foundation; with no
70: Invariant Sections, with the Front-Cover texts being ``A GNU Manual,''
71: and with the Back-Cover Texts as in (a) below. A copy of the
72: license is included in the section entitled ``GNU Free Documentation
73: License.''
74:
75: (a) The FSF's Back-Cover Text is: ``You have freedom to copy and modify
76: this GNU Manual, like GNU software. Copies published by the Free
77: Software Foundation raise funds for GNU development.''
78: @end quotation
79: @end copying
1.10 anton 80:
1.113 anton 81: @dircategory Software development
82: @direntry
83: * Gforth: (gforth). A fast interpreter for the Forth language.
84: @end direntry
85: @c The Texinfo manual also recommends doing this, but for Gforth it may
86: @c not make much sense
87: @c @dircategory Individual utilities
88: @c @direntry
89: @c * Gforth: (gforth)Invoking Gforth. gforth, gforth-fast, gforthmi
90: @c @end direntry
1.1 anton 91:
92: @titlepage
1.113 anton 93: @title Gforth
94: @subtitle for version @value{VERSION}, @value{UPDATED}
95: @author Neal Crook
96: @author Anton Ertl
1.114 anton 97: @author David Kuehling
1.113 anton 98: @author Bernd Paysan
99: @author Jens Wilke
1.1 anton 100: @page
101: @vskip 0pt plus 1filll
1.113 anton 102: @insertcopying
103: @end titlepage
1.1 anton 104:
1.113 anton 105: @contents
1.1 anton 106:
1.113 anton 107: @ifnottex
108: @node Top, Goals, (dir), (dir)
109: @top Gforth
1.1 anton 110:
1.113 anton 111: @insertcopying
1.49 anton 112: @end ifnottex
1.1 anton 113:
114: @menu
1.26 crook 115: * Goals:: About the Gforth Project
1.29 crook 116: * Gforth Environment:: Starting (and exiting) Gforth
1.48 anton 117: * Tutorial:: Hands-on Forth Tutorial
1.21 crook 118: * Introduction:: An introduction to ANS Forth
1.1 anton 119: * Words:: Forth words available in Gforth
1.24 anton 120: * Error messages:: How to interpret them
1.1 anton 121: * Tools:: Programming tools
122: * ANS conformance:: Implementation-defined options etc.
1.65 anton 123: * Standard vs Extensions:: Should I use extensions?
1.1 anton 124: * Model:: The abstract machine of Gforth
125: * Integrating Gforth:: Forth as scripting language for applications
126: * Emacs and Gforth:: The Gforth Mode
127: * Image Files:: @code{.fi} files contain compiled code
128: * Engine:: The inner interpreter and the primitives
1.13 pazsan 129: * Cross Compiler:: The Cross Compiler
1.1 anton 130: * Bugs:: How to report them
131: * Origin:: Authors and ancestors of Gforth
1.21 crook 132: * Forth-related information:: Books and places to look on the WWW
1.113 anton 133: * Licenses::
1.1 anton 134: * Word Index:: An item for each Forth word
135: * Concept Index:: A menu covering many topics
1.12 anton 136:
1.91 anton 137: @detailmenu
138: --- The Detailed Node Listing ---
1.12 anton 139:
1.29 crook 140: Gforth Environment
141:
1.32 anton 142: * Invoking Gforth:: Getting in
143: * Leaving Gforth:: Getting out
144: * Command-line editing::
1.48 anton 145: * Environment variables:: that affect how Gforth starts up
1.32 anton 146: * Gforth Files:: What gets installed and where
1.112 anton 147: * Gforth in pipes::
1.48 anton 148: * Startup speed:: When 35ms is not fast enough ...
149:
150: Forth Tutorial
151:
152: * Starting Gforth Tutorial::
153: * Syntax Tutorial::
154: * Crash Course Tutorial::
155: * Stack Tutorial::
156: * Arithmetics Tutorial::
157: * Stack Manipulation Tutorial::
158: * Using files for Forth code Tutorial::
159: * Comments Tutorial::
160: * Colon Definitions Tutorial::
161: * Decompilation Tutorial::
162: * Stack-Effect Comments Tutorial::
163: * Types Tutorial::
164: * Factoring Tutorial::
165: * Designing the stack effect Tutorial::
166: * Local Variables Tutorial::
167: * Conditional execution Tutorial::
168: * Flags and Comparisons Tutorial::
169: * General Loops Tutorial::
170: * Counted loops Tutorial::
171: * Recursion Tutorial::
172: * Leaving definitions or loops Tutorial::
173: * Return Stack Tutorial::
174: * Memory Tutorial::
175: * Characters and Strings Tutorial::
176: * Alignment Tutorial::
1.87 anton 177: * Files Tutorial::
1.48 anton 178: * Interpretation and Compilation Semantics and Immediacy Tutorial::
179: * Execution Tokens Tutorial::
180: * Exceptions Tutorial::
181: * Defining Words Tutorial::
182: * Arrays and Records Tutorial::
183: * POSTPONE Tutorial::
184: * Literal Tutorial::
185: * Advanced macros Tutorial::
186: * Compilation Tokens Tutorial::
187: * Wordlists and Search Order Tutorial::
1.29 crook 188:
1.24 anton 189: An Introduction to ANS Forth
190:
1.67 anton 191: * Introducing the Text Interpreter::
192: * Stacks and Postfix notation::
193: * Your first definition::
194: * How does that work?::
195: * Forth is written in Forth::
196: * Review - elements of a Forth system::
197: * Where to go next::
198: * Exercises::
1.24 anton 199:
1.12 anton 200: Forth Words
201:
202: * Notation::
1.65 anton 203: * Case insensitivity::
204: * Comments::
205: * Boolean Flags::
1.12 anton 206: * Arithmetic::
207: * Stack Manipulation::
208: * Memory::
209: * Control Structures::
210: * Defining Words::
1.65 anton 211: * Interpretation and Compilation Semantics::
1.47 crook 212: * Tokens for Words::
1.81 anton 213: * Compiling words::
1.65 anton 214: * The Text Interpreter::
1.111 anton 215: * The Input Stream::
1.65 anton 216: * Word Lists::
217: * Environmental Queries::
1.12 anton 218: * Files::
219: * Blocks::
220: * Other I/O::
1.121 anton 221: * OS command line arguments::
1.78 anton 222: * Locals::
223: * Structures::
224: * Object-oriented Forth::
1.12 anton 225: * Programming Tools::
1.150 anton 226: * C Interface::
1.12 anton 227: * Assembler and Code Words::
228: * Threading Words::
1.65 anton 229: * Passing Commands to the OS::
230: * Keeping track of Time::
231: * Miscellaneous Words::
1.12 anton 232:
233: Arithmetic
234:
235: * Single precision::
1.67 anton 236: * Double precision:: Double-cell integer arithmetic
1.12 anton 237: * Bitwise operations::
1.67 anton 238: * Numeric comparison::
1.32 anton 239: * Mixed precision:: Operations with single and double-cell integers
1.12 anton 240: * Floating Point::
241:
242: Stack Manipulation
243:
244: * Data stack::
245: * Floating point stack::
246: * Return stack::
247: * Locals stack::
248: * Stack pointer manipulation::
249:
250: Memory
251:
1.32 anton 252: * Memory model::
253: * Dictionary allocation::
254: * Heap Allocation::
255: * Memory Access::
256: * Address arithmetic::
257: * Memory Blocks::
1.12 anton 258:
259: Control Structures
260:
1.41 anton 261: * Selection:: IF ... ELSE ... ENDIF
262: * Simple Loops:: BEGIN ...
1.32 anton 263: * Counted Loops:: DO
1.67 anton 264: * Arbitrary control structures::
265: * Calls and returns::
1.12 anton 266: * Exception Handling::
267:
268: Defining Words
269:
1.67 anton 270: * CREATE::
1.44 crook 271: * Variables:: Variables and user variables
1.67 anton 272: * Constants::
1.44 crook 273: * Values:: Initialised variables
1.67 anton 274: * Colon Definitions::
1.44 crook 275: * Anonymous Definitions:: Definitions without names
1.71 anton 276: * Supplying names:: Passing definition names as strings
1.67 anton 277: * User-defined Defining Words::
1.170 pazsan 278: * Deferred Words:: Allow forward references
1.67 anton 279: * Aliases::
1.47 crook 280:
1.63 anton 281: User-defined Defining Words
282:
283: * CREATE..DOES> applications::
284: * CREATE..DOES> details::
285: * Advanced does> usage example::
1.155 anton 286: * Const-does>::
1.63 anton 287:
1.47 crook 288: Interpretation and Compilation Semantics
289:
1.67 anton 290: * Combined words::
1.12 anton 291:
1.71 anton 292: Tokens for Words
293:
294: * Execution token:: represents execution/interpretation semantics
295: * Compilation token:: represents compilation semantics
296: * Name token:: represents named words
297:
1.82 anton 298: Compiling words
299:
300: * Literals:: Compiling data values
301: * Macros:: Compiling words
302:
1.21 crook 303: The Text Interpreter
304:
1.67 anton 305: * Input Sources::
306: * Number Conversion::
307: * Interpret/Compile states::
308: * Interpreter Directives::
1.21 crook 309:
1.26 crook 310: Word Lists
311:
1.75 anton 312: * Vocabularies::
1.67 anton 313: * Why use word lists?::
1.75 anton 314: * Word list example::
1.26 crook 315:
316: Files
317:
1.48 anton 318: * Forth source files::
319: * General files::
1.167 anton 320: * Redirection::
1.48 anton 321: * Search Paths::
322:
323: Search Paths
324:
1.75 anton 325: * Source Search Paths::
1.26 crook 326: * General Search Paths::
327:
328: Other I/O
329:
1.32 anton 330: * Simple numeric output:: Predefined formats
331: * Formatted numeric output:: Formatted (pictured) output
332: * String Formats:: How Forth stores strings in memory
1.67 anton 333: * Displaying characters and strings:: Other stuff
1.178 anton 334: * Terminal output:: Cursor positioning etc.
1.181 ! anton 335: * Single-key input::
! 336: * Line input and conversion::
1.112 anton 337: * Pipes:: How to create your own pipes
1.149 pazsan 338: * Xchars and Unicode:: Non-ASCII characters
1.26 crook 339:
340: Locals
341:
342: * Gforth locals::
343: * ANS Forth locals::
344:
345: Gforth locals
346:
347: * Where are locals visible by name?::
348: * How long do locals live?::
1.78 anton 349: * Locals programming style::
350: * Locals implementation::
1.26 crook 351:
1.12 anton 352: Structures
353:
354: * Why explicit structure support?::
355: * Structure Usage::
356: * Structure Naming Convention::
357: * Structure Implementation::
358: * Structure Glossary::
359:
360: Object-oriented Forth
361:
1.48 anton 362: * Why object-oriented programming?::
363: * Object-Oriented Terminology::
364: * Objects::
365: * OOF::
366: * Mini-OOF::
1.23 crook 367: * Comparison with other object models::
1.12 anton 368:
1.24 anton 369: The @file{objects.fs} model
1.12 anton 370:
371: * Properties of the Objects model::
372: * Basic Objects Usage::
1.41 anton 373: * The Objects base class::
1.12 anton 374: * Creating objects::
375: * Object-Oriented Programming Style::
376: * Class Binding::
377: * Method conveniences::
378: * Classes and Scoping::
1.41 anton 379: * Dividing classes::
1.12 anton 380: * Object Interfaces::
381: * Objects Implementation::
382: * Objects Glossary::
383:
1.24 anton 384: The @file{oof.fs} model
1.12 anton 385:
1.67 anton 386: * Properties of the OOF model::
387: * Basic OOF Usage::
388: * The OOF base class::
389: * Class Declaration::
390: * Class Implementation::
1.12 anton 391:
1.24 anton 392: The @file{mini-oof.fs} model
1.23 crook 393:
1.48 anton 394: * Basic Mini-OOF Usage::
395: * Mini-OOF Example::
396: * Mini-OOF Implementation::
1.23 crook 397:
1.78 anton 398: Programming Tools
399:
1.150 anton 400: * Examining:: Data and Code.
401: * Forgetting words:: Usually before reloading.
1.78 anton 402: * Debugging:: Simple and quick.
403: * Assertions:: Making your programs self-checking.
404: * Singlestep Debugger:: Executing your program word by word.
405:
1.155 anton 406: C Interface
407:
408: * Calling C Functions::
409: * Declaring C Functions::
1.180 anton 410: * Calling C function pointers::
1.155 anton 411: * Callbacks::
1.178 anton 412: * C interface internals::
1.155 anton 413: * Low-Level C Interface Words::
414:
1.78 anton 415: Assembler and Code Words
416:
417: * Code and ;code::
418: * Common Assembler:: Assembler Syntax
419: * Common Disassembler::
420: * 386 Assembler:: Deviations and special cases
421: * Alpha Assembler:: Deviations and special cases
422: * MIPS assembler:: Deviations and special cases
1.167 anton 423: * PowerPC assembler:: Deviations and special cases
1.78 anton 424: * Other assemblers:: How to write them
425:
1.12 anton 426: Tools
427:
428: * ANS Report:: Report the words used, sorted by wordset.
1.127 anton 429: * Stack depth changes:: Where does this stack item come from?
1.12 anton 430:
431: ANS conformance
432:
433: * The Core Words::
434: * The optional Block word set::
435: * The optional Double Number word set::
436: * The optional Exception word set::
437: * The optional Facility word set::
438: * The optional File-Access word set::
439: * The optional Floating-Point word set::
440: * The optional Locals word set::
441: * The optional Memory-Allocation word set::
442: * The optional Programming-Tools word set::
443: * The optional Search-Order word set::
444:
445: The Core Words
446:
447: * core-idef:: Implementation Defined Options
448: * core-ambcond:: Ambiguous Conditions
449: * core-other:: Other System Documentation
450:
451: The optional Block word set
452:
453: * block-idef:: Implementation Defined Options
454: * block-ambcond:: Ambiguous Conditions
455: * block-other:: Other System Documentation
456:
457: The optional Double Number word set
458:
459: * double-ambcond:: Ambiguous Conditions
460:
461: The optional Exception word set
462:
463: * exception-idef:: Implementation Defined Options
464:
465: The optional Facility word set
466:
467: * facility-idef:: Implementation Defined Options
468: * facility-ambcond:: Ambiguous Conditions
469:
470: The optional File-Access word set
471:
472: * file-idef:: Implementation Defined Options
473: * file-ambcond:: Ambiguous Conditions
474:
475: The optional Floating-Point word set
476:
477: * floating-idef:: Implementation Defined Options
478: * floating-ambcond:: Ambiguous Conditions
479:
480: The optional Locals word set
481:
482: * locals-idef:: Implementation Defined Options
483: * locals-ambcond:: Ambiguous Conditions
484:
485: The optional Memory-Allocation word set
486:
487: * memory-idef:: Implementation Defined Options
488:
489: The optional Programming-Tools word set
490:
491: * programming-idef:: Implementation Defined Options
492: * programming-ambcond:: Ambiguous Conditions
493:
494: The optional Search-Order word set
495:
496: * search-idef:: Implementation Defined Options
497: * search-ambcond:: Ambiguous Conditions
498:
1.109 anton 499: Emacs and Gforth
500:
501: * Installing gforth.el:: Making Emacs aware of Forth.
502: * Emacs Tags:: Viewing the source of a word in Emacs.
503: * Hilighting:: Making Forth code look prettier.
504: * Auto-Indentation:: Customizing auto-indentation.
505: * Blocks Files:: Reading and writing blocks files.
506:
1.12 anton 507: Image Files
508:
1.24 anton 509: * Image Licensing Issues:: Distribution terms for images.
510: * Image File Background:: Why have image files?
1.67 anton 511: * Non-Relocatable Image Files:: don't always work.
1.24 anton 512: * Data-Relocatable Image Files:: are better.
1.67 anton 513: * Fully Relocatable Image Files:: better yet.
1.24 anton 514: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.32 anton 515: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.24 anton 516: * Modifying the Startup Sequence:: and turnkey applications.
1.12 anton 517:
518: Fully Relocatable Image Files
519:
1.27 crook 520: * gforthmi:: The normal way
1.12 anton 521: * cross.fs:: The hard way
522:
523: Engine
524:
525: * Portability::
526: * Threading::
527: * Primitives::
528: * Performance::
529:
530: Threading
531:
532: * Scheduling::
533: * Direct or Indirect Threaded?::
1.109 anton 534: * Dynamic Superinstructions::
1.12 anton 535: * DOES>::
536:
537: Primitives
538:
539: * Automatic Generation::
540: * TOS Optimization::
541: * Produced code::
1.13 pazsan 542:
543: Cross Compiler
544:
1.67 anton 545: * Using the Cross Compiler::
546: * How the Cross Compiler Works::
1.13 pazsan 547:
1.113 anton 548: Licenses
549:
550: * GNU Free Documentation License:: License for copying this manual.
551: * Copying:: GPL (for copying this software).
552:
1.24 anton 553: @end detailmenu
1.1 anton 554: @end menu
555:
1.113 anton 556: @c ----------------------------------------------------------
1.1 anton 557: @iftex
558: @unnumbered Preface
559: @cindex Preface
1.21 crook 560: This manual documents Gforth. Some introductory material is provided for
561: readers who are unfamiliar with Forth or who are migrating to Gforth
562: from other Forth compilers. However, this manual is primarily a
563: reference manual.
1.1 anton 564: @end iftex
565:
1.28 crook 566: @comment TODO much more blurb here.
1.26 crook 567:
568: @c ******************************************************************
1.113 anton 569: @node Goals, Gforth Environment, Top, Top
1.26 crook 570: @comment node-name, next, previous, up
571: @chapter Goals of Gforth
572: @cindex goals of the Gforth project
573: The goal of the Gforth Project is to develop a standard model for
574: ANS Forth. This can be split into several subgoals:
575:
576: @itemize @bullet
577: @item
578: Gforth should conform to the ANS Forth Standard.
579: @item
580: It should be a model, i.e. it should define all the
581: implementation-dependent things.
582: @item
583: It should become standard, i.e. widely accepted and used. This goal
584: is the most difficult one.
585: @end itemize
586:
587: To achieve these goals Gforth should be
588: @itemize @bullet
589: @item
590: Similar to previous models (fig-Forth, F83)
591: @item
592: Powerful. It should provide for all the things that are considered
593: necessary today and even some that are not yet considered necessary.
594: @item
595: Efficient. It should not get the reputation of being exceptionally
596: slow.
597: @item
598: Free.
599: @item
600: Available on many machines/easy to port.
601: @end itemize
602:
603: Have we achieved these goals? Gforth conforms to the ANS Forth
604: standard. It may be considered a model, but we have not yet documented
605: which parts of the model are stable and which parts we are likely to
606: change. It certainly has not yet become a de facto standard, but it
607: appears to be quite popular. It has some similarities to and some
608: differences from previous models. It has some powerful features, but not
609: yet everything that we envisioned. We certainly have achieved our
1.65 anton 610: execution speed goals (@pxref{Performance})@footnote{However, in 1998
611: the bar was raised when the major commercial Forth vendors switched to
612: native code compilers.}. It is free and available on many machines.
1.29 crook 613:
1.26 crook 614: @c ******************************************************************
1.48 anton 615: @node Gforth Environment, Tutorial, Goals, Top
1.29 crook 616: @chapter Gforth Environment
617: @cindex Gforth environment
1.21 crook 618:
1.45 crook 619: Note: ultimately, the Gforth man page will be auto-generated from the
1.29 crook 620: material in this chapter.
1.21 crook 621:
622: @menu
1.29 crook 623: * Invoking Gforth:: Getting in
624: * Leaving Gforth:: Getting out
625: * Command-line editing::
1.48 anton 626: * Environment variables:: that affect how Gforth starts up
1.29 crook 627: * Gforth Files:: What gets installed and where
1.112 anton 628: * Gforth in pipes::
1.48 anton 629: * Startup speed:: When 35ms is not fast enough ...
1.21 crook 630: @end menu
631:
1.49 anton 632: For related information about the creation of images see @ref{Image Files}.
1.29 crook 633:
1.21 crook 634: @comment ----------------------------------------------
1.48 anton 635: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1.29 crook 636: @section Invoking Gforth
637: @cindex invoking Gforth
638: @cindex running Gforth
639: @cindex command-line options
640: @cindex options on the command line
641: @cindex flags on the command line
1.21 crook 642:
1.30 anton 643: Gforth is made up of two parts; an executable ``engine'' (named
1.109 anton 644: @command{gforth} or @command{gforth-fast}) and an image file. To start it, you
1.30 anton 645: will usually just say @code{gforth} -- this automatically loads the
646: default image file @file{gforth.fi}. In many other cases the default
647: Gforth image will be invoked like this:
1.21 crook 648: @example
1.30 anton 649: gforth [file | -e forth-code] ...
1.21 crook 650: @end example
1.29 crook 651: @noindent
652: This interprets the contents of the files and the Forth code in the order they
653: are given.
1.21 crook 654:
1.109 anton 655: In addition to the @command{gforth} engine, there is also an engine
656: called @command{gforth-fast}, which is faster, but gives less
657: informative error messages (@pxref{Error messages}) and may catch some
1.166 anton 658: errors (in particular, stack underflows and integer division errors)
659: later or not at all. You should use it for debugged,
1.109 anton 660: performance-critical programs.
661:
662: Moreover, there is an engine called @command{gforth-itc}, which is
663: useful in some backwards-compatibility situations (@pxref{Direct or
664: Indirect Threaded?}).
1.30 anton 665:
1.29 crook 666: In general, the command line looks like this:
1.21 crook 667:
668: @example
1.30 anton 669: gforth[-fast] [engine options] [image options]
1.21 crook 670: @end example
671:
1.30 anton 672: The engine options must come before the rest of the command
1.29 crook 673: line. They are:
1.26 crook 674:
1.29 crook 675: @table @code
676: @cindex -i, command-line option
677: @cindex --image-file, command-line option
678: @item --image-file @i{file}
679: @itemx -i @i{file}
680: Loads the Forth image @i{file} instead of the default
681: @file{gforth.fi} (@pxref{Image Files}).
1.21 crook 682:
1.39 anton 683: @cindex --appl-image, command-line option
684: @item --appl-image @i{file}
685: Loads the image @i{file} and leaves all further command-line arguments
1.65 anton 686: to the image (instead of processing them as engine options). This is
687: useful for building executable application images on Unix, built with
1.39 anton 688: @code{gforthmi --application ...}.
689:
1.29 crook 690: @cindex --path, command-line option
691: @cindex -p, command-line option
692: @item --path @i{path}
693: @itemx -p @i{path}
694: Uses @i{path} for searching the image file and Forth source code files
695: instead of the default in the environment variable @code{GFORTHPATH} or
696: the path specified at installation time (e.g.,
697: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
698: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1.21 crook 699:
1.29 crook 700: @cindex --dictionary-size, command-line option
701: @cindex -m, command-line option
702: @cindex @i{size} parameters for command-line options
703: @cindex size of the dictionary and the stacks
704: @item --dictionary-size @i{size}
705: @itemx -m @i{size}
706: Allocate @i{size} space for the Forth dictionary space instead of
707: using the default specified in the image (typically 256K). The
708: @i{size} specification for this and subsequent options consists of
709: an integer and a unit (e.g.,
710: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
711: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
712: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
713: @code{e} is used.
1.21 crook 714:
1.29 crook 715: @cindex --data-stack-size, command-line option
716: @cindex -d, command-line option
717: @item --data-stack-size @i{size}
718: @itemx -d @i{size}
719: Allocate @i{size} space for the data stack instead of using the
720: default specified in the image (typically 16K).
1.21 crook 721:
1.29 crook 722: @cindex --return-stack-size, command-line option
723: @cindex -r, command-line option
724: @item --return-stack-size @i{size}
725: @itemx -r @i{size}
726: Allocate @i{size} space for the return stack instead of using the
727: default specified in the image (typically 15K).
1.21 crook 728:
1.29 crook 729: @cindex --fp-stack-size, command-line option
730: @cindex -f, command-line option
731: @item --fp-stack-size @i{size}
732: @itemx -f @i{size}
733: Allocate @i{size} space for the floating point stack instead of
734: using the default specified in the image (typically 15.5K). In this case
735: the unit specifier @code{e} refers to floating point numbers.
1.21 crook 736:
1.48 anton 737: @cindex --locals-stack-size, command-line option
738: @cindex -l, command-line option
739: @item --locals-stack-size @i{size}
740: @itemx -l @i{size}
741: Allocate @i{size} space for the locals stack instead of using the
742: default specified in the image (typically 14.5K).
743:
1.176 anton 744: @cindex --vm-commit, command-line option
745: @cindex overcommit memory for dictionary and stacks
746: @cindex memory overcommit for dictionary and stacks
747: @item --vm-commit
748: Normally, Gforth tries to start up even if there is not enough virtual
749: memory for the dictionary and the stacks (using @code{MAP_NORESERVE}
750: on OSs that support it); so you can ask for a really big dictionary
751: and/or stacks, and as long as you don't use more virtual memory than
752: is available, everything will be fine (but if you use more, processes
753: get killed). With this option you just use the default allocation
754: policy of the OS; for OSs that don't overcommit (e.g., Solaris), this
755: means that you cannot and should not ask for as big dictionary and
756: stacks, but once Gforth successfully starts up, out-of-memory won't
757: kill it.
758:
1.48 anton 759: @cindex -h, command-line option
760: @cindex --help, command-line option
761: @item --help
762: @itemx -h
763: Print a message about the command-line options
764:
765: @cindex -v, command-line option
766: @cindex --version, command-line option
767: @item --version
768: @itemx -v
769: Print version and exit
770:
771: @cindex --debug, command-line option
772: @item --debug
773: Print some information useful for debugging on startup.
774:
775: @cindex --offset-image, command-line option
776: @item --offset-image
777: Start the dictionary at a slightly different position than would be used
778: otherwise (useful for creating data-relocatable images,
779: @pxref{Data-Relocatable Image Files}).
780:
781: @cindex --no-offset-im, command-line option
782: @item --no-offset-im
783: Start the dictionary at the normal position.
784:
785: @cindex --clear-dictionary, command-line option
786: @item --clear-dictionary
787: Initialize all bytes in the dictionary to 0 before loading the image
788: (@pxref{Data-Relocatable Image Files}).
789:
790: @cindex --die-on-signal, command-line-option
791: @item --die-on-signal
792: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
793: or the segmentation violation SIGSEGV) by translating it into a Forth
794: @code{THROW}. With this option, Gforth exits if it receives such a
795: signal. This option is useful when the engine and/or the image might be
796: severely broken (such that it causes another signal before recovering
797: from the first); this option avoids endless loops in such cases.
1.109 anton 798:
1.119 anton 799: @cindex --no-dynamic, command-line option
800: @cindex --dynamic, command-line option
1.109 anton 801: @item --no-dynamic
802: @item --dynamic
803: Disable or enable dynamic superinstructions with replication
804: (@pxref{Dynamic Superinstructions}).
805:
1.119 anton 806: @cindex --no-super, command-line option
1.109 anton 807: @item --no-super
1.110 anton 808: Disable dynamic superinstructions, use just dynamic replication; this is
809: useful if you want to patch threaded code (@pxref{Dynamic
810: Superinstructions}).
1.119 anton 811:
812: @cindex --ss-number, command-line option
813: @item --ss-number=@var{N}
814: Use only the first @var{N} static superinstructions compiled into the
815: engine (default: use them all; note that only @code{gforth-fast} has
816: any). This option is useful for measuring the performance impact of
817: static superinstructions.
818:
819: @cindex --ss-min-..., command-line options
820: @item --ss-min-codesize
821: @item --ss-min-ls
822: @item --ss-min-lsu
823: @item --ss-min-nexts
824: Use specified metric for determining the cost of a primitive or static
825: superinstruction for static superinstruction selection. @code{Codesize}
826: is the native code size of the primive or static superinstruction,
827: @code{ls} is the number of loads and stores, @code{lsu} is the number of
828: loads, stores, and updates, and @code{nexts} is the number of dispatches
829: (not taking dynamic superinstructions into account), i.e. every
830: primitive or static superinstruction has cost 1. Default:
831: @code{codesize} if you use dynamic code generation, otherwise
832: @code{nexts}.
833:
834: @cindex --ss-greedy, command-line option
835: @item --ss-greedy
836: This option is useful for measuring the performance impact of static
837: superinstructions. By default, an optimal shortest-path algorithm is
838: used for selecting static superinstructions. With @option{--ss-greedy}
839: this algorithm is modified to assume that anything after the static
840: superinstruction currently under consideration is not combined into
841: static superinstructions. With @option{--ss-min-nexts} this produces
842: the same result as a greedy algorithm that always selects the longest
843: superinstruction available at the moment. E.g., if there are
844: superinstructions AB and BCD, then for the sequence A B C D the optimal
845: algorithm will select A BCD and the greedy algorithm will select AB C D.
846:
847: @cindex --print-metrics, command-line option
848: @item --print-metrics
849: Prints some metrics used during static superinstruction selection:
850: @code{code size} is the actual size of the dynamically generated code.
851: @code{Metric codesize} is the sum of the codesize metrics as seen by
852: static superinstruction selection; there is a difference from @code{code
853: size}, because not all primitives and static superinstructions are
854: compiled into dynamically generated code, and because of markers. The
855: other metrics correspond to the @option{ss-min-...} options. This
856: option is useful for evaluating the effects of the @option{--ss-...}
857: options.
1.109 anton 858:
1.48 anton 859: @end table
860:
861: @cindex loading files at startup
862: @cindex executing code on startup
863: @cindex batch processing with Gforth
864: As explained above, the image-specific command-line arguments for the
865: default image @file{gforth.fi} consist of a sequence of filenames and
866: @code{-e @var{forth-code}} options that are interpreted in the sequence
867: in which they are given. The @code{-e @var{forth-code}} or
1.121 anton 868: @code{--evaluate @var{forth-code}} option evaluates the Forth code. This
869: option takes only one argument; if you want to evaluate more Forth
870: words, you have to quote them or use @code{-e} several times. To exit
1.48 anton 871: after processing the command line (instead of entering interactive mode)
1.121 anton 872: append @code{-e bye} to the command line. You can also process the
873: command-line arguments with a Forth program (@pxref{OS command line
874: arguments}).
1.48 anton 875:
876: @cindex versions, invoking other versions of Gforth
877: If you have several versions of Gforth installed, @code{gforth} will
878: invoke the version that was installed last. @code{gforth-@i{version}}
879: invokes a specific version. If your environment contains the variable
880: @code{GFORTHPATH}, you may want to override it by using the
881: @code{--path} option.
882:
883: Not yet implemented:
884: On startup the system first executes the system initialization file
885: (unless the option @code{--no-init-file} is given; note that the system
886: resulting from using this option may not be ANS Forth conformant). Then
887: the user initialization file @file{.gforth.fs} is executed, unless the
1.62 crook 888: option @code{--no-rc} is given; this file is searched for in @file{.},
1.48 anton 889: then in @file{~}, then in the normal path (see above).
890:
891:
892:
893: @comment ----------------------------------------------
894: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
895: @section Leaving Gforth
896: @cindex Gforth - leaving
897: @cindex leaving Gforth
898:
899: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
900: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
901: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1.49 anton 902: data are discarded. For ways of saving the state of the system before
903: leaving Gforth see @ref{Image Files}.
1.48 anton 904:
905: doc-bye
906:
907:
908: @comment ----------------------------------------------
1.65 anton 909: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1.48 anton 910: @section Command-line editing
911: @cindex command-line editing
912:
913: Gforth maintains a history file that records every line that you type to
914: the text interpreter. This file is preserved between sessions, and is
915: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
916: repeatedly you can recall successively older commands from this (or
917: previous) session(s). The full list of command-line editing facilities is:
918:
919: @itemize @bullet
920: @item
921: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
922: commands from the history buffer.
923: @item
924: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
925: from the history buffer.
926: @item
927: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
928: @item
929: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
930: @item
931: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
932: closing up the line.
933: @item
934: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
935: @item
936: @kbd{Ctrl-a} to move the cursor to the start of the line.
937: @item
938: @kbd{Ctrl-e} to move the cursor to the end of the line.
939: @item
940: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
941: line.
942: @item
943: @key{TAB} to step through all possible full-word completions of the word
944: currently being typed.
945: @item
1.65 anton 946: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
947: using @code{bye}).
948: @item
949: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
950: character under the cursor.
1.48 anton 951: @end itemize
952:
953: When editing, displayable characters are inserted to the left of the
954: cursor position; the line is always in ``insert'' (as opposed to
955: ``overstrike'') mode.
956:
957: @cindex history file
958: @cindex @file{.gforth-history}
959: On Unix systems, the history file is @file{~/.gforth-history} by
960: default@footnote{i.e. it is stored in the user's home directory.}. You
961: can find out the name and location of your history file using:
962:
963: @example
964: history-file type \ Unix-class systems
965:
966: history-file type \ Other systems
967: history-dir type
968: @end example
969:
970: If you enter long definitions by hand, you can use a text editor to
971: paste them out of the history file into a Forth source file for reuse at
972: a later time.
973:
974: Gforth never trims the size of the history file, so you should do this
975: periodically, if necessary.
976:
977: @comment this is all defined in history.fs
978: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
979: @comment chosen?
980:
981:
982: @comment ----------------------------------------------
1.65 anton 983: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1.48 anton 984: @section Environment variables
985: @cindex environment variables
986:
987: Gforth uses these environment variables:
988:
989: @itemize @bullet
990: @item
991: @cindex @code{GFORTHHIST} -- environment variable
992: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
993: open/create the history file, @file{.gforth-history}. Default:
994: @code{$HOME}.
995:
996: @item
997: @cindex @code{GFORTHPATH} -- environment variable
998: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
999: for Forth source-code files.
1000:
1001: @item
1.147 anton 1002: @cindex @code{LANG} -- environment variable
1003: @code{LANG} -- see @code{LC_CTYPE}
1004:
1005: @item
1006: @cindex @code{LC_ALL} -- environment variable
1007: @code{LC_ALL} -- see @code{LC_CTYPE}
1008:
1009: @item
1010: @cindex @code{LC_CTYPE} -- environment variable
1011: @code{LC_CTYPE} -- If this variable contains ``UTF-8'' on Gforth
1012: startup, Gforth uses the UTF-8 encoding for strings internally and
1013: expects its input and produces its output in UTF-8 encoding, otherwise
1014: the encoding is 8bit (see @pxref{Xchars and Unicode}). If this
1015: environment variable is unset, Gforth looks in @code{LC_ALL}, and if
1016: that is unset, in @code{LANG}.
1017:
1018: @item
1.129 anton 1019: @cindex @code{GFORTHSYSTEMPREFIX} -- environment variable
1020:
1021: @code{GFORTHSYSTEMPREFIX} -- specifies what to prepend to the argument
1022: of @code{system} before passing it to C's @code{system()}. Default:
1.130 anton 1023: @code{"./$COMSPEC /c "} on Windows, @code{""} on other OSs. The prefix
1.129 anton 1024: and the command are directly concatenated, so if a space between them is
1025: necessary, append it to the prefix.
1026:
1027: @item
1.48 anton 1028: @cindex @code{GFORTH} -- environment variable
1.49 anton 1029: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1030:
1031: @item
1032: @cindex @code{GFORTHD} -- environment variable
1.62 crook 1033: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1.48 anton 1034:
1035: @item
1036: @cindex @code{TMP}, @code{TEMP} - environment variable
1037: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1038: location for the history file.
1039: @end itemize
1040:
1041: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1042: @comment mentioning these.
1043:
1044: All the Gforth environment variables default to sensible values if they
1045: are not set.
1046:
1047:
1048: @comment ----------------------------------------------
1.112 anton 1049: @node Gforth Files, Gforth in pipes, Environment variables, Gforth Environment
1.48 anton 1050: @section Gforth files
1051: @cindex Gforth files
1052:
1053: When you install Gforth on a Unix system, it installs files in these
1054: locations by default:
1055:
1056: @itemize @bullet
1057: @item
1058: @file{/usr/local/bin/gforth}
1059: @item
1060: @file{/usr/local/bin/gforthmi}
1061: @item
1062: @file{/usr/local/man/man1/gforth.1} - man page.
1063: @item
1064: @file{/usr/local/info} - the Info version of this manual.
1065: @item
1066: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1067: @item
1068: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1069: @item
1070: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1071: @item
1072: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1073: @end itemize
1074:
1075: You can select different places for installation by using
1076: @code{configure} options (listed with @code{configure --help}).
1077:
1078: @comment ----------------------------------------------
1.112 anton 1079: @node Gforth in pipes, Startup speed, Gforth Files, Gforth Environment
1080: @section Gforth in pipes
1081: @cindex pipes, Gforth as part of
1082:
1083: Gforth can be used in pipes created elsewhere (described here). It can
1084: also create pipes on its own (@pxref{Pipes}).
1085:
1086: @cindex input from pipes
1087: If you pipe into Gforth, your program should read with @code{read-file}
1088: or @code{read-line} from @code{stdin} (@pxref{General files}).
1089: @code{Key} does not recognize the end of input. Words like
1090: @code{accept} echo the input and are therefore usually not useful for
1091: reading from a pipe. You have to invoke the Forth program with an OS
1092: command-line option, as you have no chance to use the Forth command line
1093: (the text interpreter would try to interpret the pipe input).
1094:
1095: @cindex output in pipes
1096: You can output to a pipe with @code{type}, @code{emit}, @code{cr} etc.
1097:
1098: @cindex silent exiting from Gforth
1099: When you write to a pipe that has been closed at the other end, Gforth
1100: receives a SIGPIPE signal (``pipe broken''). Gforth translates this
1101: into the exception @code{broken-pipe-error}. If your application does
1102: not catch that exception, the system catches it and exits, usually
1103: silently (unless you were working on the Forth command line; then it
1104: prints an error message and exits). This is usually the desired
1105: behaviour.
1106:
1107: If you do not like this behaviour, you have to catch the exception
1108: yourself, and react to it.
1109:
1110: Here's an example of an invocation of Gforth that is usable in a pipe:
1111:
1112: @example
1113: gforth -e ": foo begin pad dup 10 stdin read-file throw dup while \
1114: type repeat ; foo bye"
1115: @end example
1116:
1117: This example just copies the input verbatim to the output. A very
1118: simple pipe containing this example looks like this:
1119:
1120: @example
1121: cat startup.fs |
1122: gforth -e ": foo begin pad dup 80 stdin read-file throw dup while \
1123: type repeat ; foo bye"|
1124: head
1125: @end example
1126:
1127: @cindex stderr and pipes
1128: Pipes involving Gforth's @code{stderr} output do not work.
1129:
1130: @comment ----------------------------------------------
1131: @node Startup speed, , Gforth in pipes, Gforth Environment
1.48 anton 1132: @section Startup speed
1133: @cindex Startup speed
1134: @cindex speed, startup
1135:
1136: If Gforth is used for CGI scripts or in shell scripts, its startup
1137: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1138: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1139: system time.
1140:
1141: If startup speed is a problem, you may consider the following ways to
1142: improve it; or you may consider ways to reduce the number of startups
1.62 crook 1143: (for example, by using Fast-CGI).
1.48 anton 1144:
1.112 anton 1145: An easy step that influences Gforth startup speed is the use of the
1146: @option{--no-dynamic} option; this decreases image loading speed, but
1147: increases compile-time and run-time.
1148:
1149: Another step to improve startup speed is to statically link Gforth, by
1.48 anton 1150: building it with @code{XLDFLAGS=-static}. This requires more memory for
1151: the code and will therefore slow down the first invocation, but
1152: subsequent invocations avoid the dynamic linking overhead. Another
1153: disadvantage is that Gforth won't profit from library upgrades. As a
1154: result, @code{gforth-static -e bye} takes about 17.1ms user and
1155: 8.2ms system time.
1156:
1157: The next step to improve startup speed is to use a non-relocatable image
1.65 anton 1158: (@pxref{Non-Relocatable Image Files}). You can create this image with
1.48 anton 1159: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1160: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1161: and a part of the copy-on-write overhead. The disadvantage is that the
1.62 crook 1162: non-relocatable image does not work if the OS gives Gforth a different
1.48 anton 1163: address for the dictionary, for whatever reason; so you better provide a
1164: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1165: bye} takes about 15.3ms user and 7.5ms system time.
1166:
1167: The final step is to disable dictionary hashing in Gforth. Gforth
1168: builds the hash table on startup, which takes much of the startup
1169: overhead. You can do this by commenting out the @code{include hash.fs}
1170: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1171: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1172: The disadvantages are that functionality like @code{table} and
1173: @code{ekey} is missing and that text interpretation (e.g., compiling)
1174: now takes much longer. So, you should only use this method if there is
1175: no significant text interpretation to perform (the script should be
1.62 crook 1176: compiled into the image, amongst other things). @code{gforth-static -i
1.48 anton 1177: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1178:
1179: @c ******************************************************************
1180: @node Tutorial, Introduction, Gforth Environment, Top
1181: @chapter Forth Tutorial
1182: @cindex Tutorial
1183: @cindex Forth Tutorial
1184:
1.67 anton 1185: @c Topics from nac's Introduction that could be mentioned:
1186: @c press <ret> after each line
1187: @c Prompt
1188: @c numbers vs. words in dictionary on text interpretation
1189: @c what happens on redefinition
1190: @c parsing words (in particular, defining words)
1191:
1.83 anton 1192: The difference of this chapter from the Introduction
1193: (@pxref{Introduction}) is that this tutorial is more fast-paced, should
1194: be used while sitting in front of a computer, and covers much more
1195: material, but does not explain how the Forth system works.
1196:
1.62 crook 1197: This tutorial can be used with any ANS-compliant Forth; any
1198: Gforth-specific features are marked as such and you can skip them if you
1199: work with another Forth. This tutorial does not explain all features of
1200: Forth, just enough to get you started and give you some ideas about the
1201: facilities available in Forth. Read the rest of the manual and the
1202: standard when you are through this.
1.48 anton 1203:
1204: The intended way to use this tutorial is that you work through it while
1205: sitting in front of the console, take a look at the examples and predict
1206: what they will do, then try them out; if the outcome is not as expected,
1207: find out why (e.g., by trying out variations of the example), so you
1208: understand what's going on. There are also some assignments that you
1209: should solve.
1210:
1211: This tutorial assumes that you have programmed before and know what,
1212: e.g., a loop is.
1213:
1214: @c !! explain compat library
1215:
1216: @menu
1217: * Starting Gforth Tutorial::
1218: * Syntax Tutorial::
1219: * Crash Course Tutorial::
1220: * Stack Tutorial::
1221: * Arithmetics Tutorial::
1222: * Stack Manipulation Tutorial::
1223: * Using files for Forth code Tutorial::
1224: * Comments Tutorial::
1225: * Colon Definitions Tutorial::
1226: * Decompilation Tutorial::
1227: * Stack-Effect Comments Tutorial::
1228: * Types Tutorial::
1229: * Factoring Tutorial::
1230: * Designing the stack effect Tutorial::
1231: * Local Variables Tutorial::
1232: * Conditional execution Tutorial::
1233: * Flags and Comparisons Tutorial::
1234: * General Loops Tutorial::
1235: * Counted loops Tutorial::
1236: * Recursion Tutorial::
1237: * Leaving definitions or loops Tutorial::
1238: * Return Stack Tutorial::
1239: * Memory Tutorial::
1240: * Characters and Strings Tutorial::
1241: * Alignment Tutorial::
1.87 anton 1242: * Files Tutorial::
1.48 anton 1243: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1244: * Execution Tokens Tutorial::
1245: * Exceptions Tutorial::
1246: * Defining Words Tutorial::
1247: * Arrays and Records Tutorial::
1248: * POSTPONE Tutorial::
1249: * Literal Tutorial::
1250: * Advanced macros Tutorial::
1251: * Compilation Tokens Tutorial::
1252: * Wordlists and Search Order Tutorial::
1253: @end menu
1254:
1255: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1256: @section Starting Gforth
1.66 anton 1257: @cindex starting Gforth tutorial
1.48 anton 1258: You can start Gforth by typing its name:
1259:
1260: @example
1261: gforth
1262: @end example
1263:
1264: That puts you into interactive mode; you can leave Gforth by typing
1265: @code{bye}. While in Gforth, you can edit the command line and access
1266: the command line history with cursor keys, similar to bash.
1267:
1268:
1269: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1270: @section Syntax
1.66 anton 1271: @cindex syntax tutorial
1.48 anton 1272:
1.171 anton 1273: A @dfn{word} is a sequence of arbitrary characters (except white
1.48 anton 1274: space). Words are separated by white space. E.g., each of the
1275: following lines contains exactly one word:
1276:
1277: @example
1278: word
1279: !@@#$%^&*()
1280: 1234567890
1281: 5!a
1282: @end example
1283:
1284: A frequent beginner's error is to leave away necessary white space,
1285: resulting in an error like @samp{Undefined word}; so if you see such an
1286: error, check if you have put spaces wherever necessary.
1287:
1288: @example
1289: ." hello, world" \ correct
1290: ."hello, world" \ gives an "Undefined word" error
1291: @end example
1292:
1.65 anton 1293: Gforth and most other Forth systems ignore differences in case (they are
1.48 anton 1294: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1295: your system is case-sensitive, you may have to type all the examples
1296: given here in upper case.
1297:
1298:
1299: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1300: @section Crash Course
1301:
1302: Type
1303:
1304: @example
1305: 0 0 !
1306: here execute
1307: ' catch >body 20 erase abort
1308: ' (quit) >body 20 erase
1309: @end example
1310:
1311: The last two examples are guaranteed to destroy parts of Gforth (and
1312: most other systems), so you better leave Gforth afterwards (if it has
1313: not finished by itself). On some systems you may have to kill gforth
1314: from outside (e.g., in Unix with @code{kill}).
1315:
1316: Now that you know how to produce crashes (and that there's not much to
1317: them), let's learn how to produce meaningful programs.
1318:
1319:
1320: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1321: @section Stack
1.66 anton 1322: @cindex stack tutorial
1.48 anton 1323:
1324: The most obvious feature of Forth is the stack. When you type in a
1325: number, it is pushed on the stack. You can display the content of the
1326: stack with @code{.s}.
1327:
1328: @example
1329: 1 2 .s
1330: 3 .s
1331: @end example
1332:
1333: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1334: appear in @code{.s} output as they appeared in the input.
1335:
1336: You can print the top of stack element with @code{.}.
1337:
1338: @example
1339: 1 2 3 . . .
1340: @end example
1341:
1342: In general, words consume their stack arguments (@code{.s} is an
1343: exception).
1344:
1.141 anton 1345: @quotation Assignment
1.48 anton 1346: What does the stack contain after @code{5 6 7 .}?
1.141 anton 1347: @end quotation
1.48 anton 1348:
1349:
1350: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1351: @section Arithmetics
1.66 anton 1352: @cindex arithmetics tutorial
1.48 anton 1353:
1354: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1355: operate on the top two stack items:
1356:
1357: @example
1.67 anton 1358: 2 2 .s
1359: + .s
1360: .
1.48 anton 1361: 2 1 - .
1362: 7 3 mod .
1363: @end example
1364:
1365: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1366: as in the corresponding infix expression (this is generally the case in
1367: Forth).
1368:
1369: Parentheses are superfluous (and not available), because the order of
1370: the words unambiguously determines the order of evaluation and the
1371: operands:
1372:
1373: @example
1374: 3 4 + 5 * .
1375: 3 4 5 * + .
1376: @end example
1377:
1.141 anton 1378: @quotation Assignment
1.48 anton 1379: What are the infix expressions corresponding to the Forth code above?
1380: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1381: known as Postfix or RPN (Reverse Polish Notation).}.
1.141 anton 1382: @end quotation
1.48 anton 1383:
1384: To change the sign, use @code{negate}:
1385:
1386: @example
1387: 2 negate .
1388: @end example
1389:
1.141 anton 1390: @quotation Assignment
1.48 anton 1391: Convert -(-3)*4-5 to Forth.
1.141 anton 1392: @end quotation
1.48 anton 1393:
1394: @code{/mod} performs both @code{/} and @code{mod}.
1395:
1396: @example
1397: 7 3 /mod . .
1398: @end example
1399:
1.66 anton 1400: Reference: @ref{Arithmetic}.
1401:
1402:
1.48 anton 1403: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1404: @section Stack Manipulation
1.66 anton 1405: @cindex stack manipulation tutorial
1.48 anton 1406:
1407: Stack manipulation words rearrange the data on the stack.
1408:
1409: @example
1410: 1 .s drop .s
1411: 1 .s dup .s drop drop .s
1412: 1 2 .s over .s drop drop drop
1413: 1 2 .s swap .s drop drop
1414: 1 2 3 .s rot .s drop drop drop
1415: @end example
1416:
1417: These are the most important stack manipulation words. There are also
1418: variants that manipulate twice as many stack items:
1419:
1420: @example
1421: 1 2 3 4 .s 2swap .s 2drop 2drop
1422: @end example
1423:
1424: Two more stack manipulation words are:
1425:
1426: @example
1427: 1 2 .s nip .s drop
1428: 1 2 .s tuck .s 2drop drop
1429: @end example
1430:
1.141 anton 1431: @quotation Assignment
1.48 anton 1432: Replace @code{nip} and @code{tuck} with combinations of other stack
1433: manipulation words.
1434:
1435: @example
1436: Given: How do you get:
1437: 1 2 3 3 2 1
1438: 1 2 3 1 2 3 2
1439: 1 2 3 1 2 3 3
1440: 1 2 3 1 3 3
1441: 1 2 3 2 1 3
1442: 1 2 3 4 4 3 2 1
1443: 1 2 3 1 2 3 1 2 3
1444: 1 2 3 4 1 2 3 4 1 2
1445: 1 2 3
1446: 1 2 3 1 2 3 4
1447: 1 2 3 1 3
1448: @end example
1.141 anton 1449: @end quotation
1.48 anton 1450:
1451: @example
1452: 5 dup * .
1453: @end example
1454:
1.141 anton 1455: @quotation Assignment
1.48 anton 1456: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1457: Write a piece of Forth code that expects two numbers on the stack
1458: (@var{a} and @var{b}, with @var{b} on top) and computes
1459: @code{(a-b)(a+1)}.
1.141 anton 1460: @end quotation
1.48 anton 1461:
1.66 anton 1462: Reference: @ref{Stack Manipulation}.
1463:
1464:
1.48 anton 1465: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1466: @section Using files for Forth code
1.66 anton 1467: @cindex loading Forth code, tutorial
1468: @cindex files containing Forth code, tutorial
1.48 anton 1469:
1470: While working at the Forth command line is convenient for one-line
1471: examples and short one-off code, you probably want to store your source
1472: code in files for convenient editing and persistence. You can use your
1473: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1.102 anton 1474: Gforth}) to create @var{file.fs} and use
1.48 anton 1475:
1476: @example
1.102 anton 1477: s" @var{file.fs}" included
1.48 anton 1478: @end example
1479:
1480: to load it into your Forth system. The file name extension I use for
1481: Forth files is @samp{.fs}.
1482:
1483: You can easily start Gforth with some files loaded like this:
1484:
1485: @example
1.102 anton 1486: gforth @var{file1.fs} @var{file2.fs}
1.48 anton 1487: @end example
1488:
1489: If an error occurs during loading these files, Gforth terminates,
1490: whereas an error during @code{INCLUDED} within Gforth usually gives you
1491: a Gforth command line. Starting the Forth system every time gives you a
1492: clean start every time, without interference from the results of earlier
1493: tries.
1494:
1495: I often put all the tests in a file, then load the code and run the
1496: tests with
1497:
1498: @example
1.102 anton 1499: gforth @var{code.fs} @var{tests.fs} -e bye
1.48 anton 1500: @end example
1501:
1502: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1503: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1504: restart this command without ado.
1505:
1506: The advantage of this approach is that the tests can be repeated easily
1507: every time the program ist changed, making it easy to catch bugs
1508: introduced by the change.
1509:
1.66 anton 1510: Reference: @ref{Forth source files}.
1511:
1.48 anton 1512:
1513: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1514: @section Comments
1.66 anton 1515: @cindex comments tutorial
1.48 anton 1516:
1517: @example
1518: \ That's a comment; it ends at the end of the line
1519: ( Another comment; it ends here: ) .s
1520: @end example
1521:
1522: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1523: separated with white space from the following text.
1524:
1525: @example
1526: \This gives an "Undefined word" error
1527: @end example
1528:
1529: The first @code{)} ends a comment started with @code{(}, so you cannot
1530: nest @code{(}-comments; and you cannot comment out text containing a
1531: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1532: avoid @code{)} in word names.}.
1533:
1534: I use @code{\}-comments for descriptive text and for commenting out code
1535: of one or more line; I use @code{(}-comments for describing the stack
1536: effect, the stack contents, or for commenting out sub-line pieces of
1537: code.
1538:
1539: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1540: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1541: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1542: with @kbd{M-q}.
1543:
1.66 anton 1544: Reference: @ref{Comments}.
1545:
1.48 anton 1546:
1547: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1548: @section Colon Definitions
1.66 anton 1549: @cindex colon definitions, tutorial
1550: @cindex definitions, tutorial
1551: @cindex procedures, tutorial
1552: @cindex functions, tutorial
1.48 anton 1553:
1554: are similar to procedures and functions in other programming languages.
1555:
1556: @example
1557: : squared ( n -- n^2 )
1558: dup * ;
1559: 5 squared .
1560: 7 squared .
1561: @end example
1562:
1563: @code{:} starts the colon definition; its name is @code{squared}. The
1564: following comment describes its stack effect. The words @code{dup *}
1565: are not executed, but compiled into the definition. @code{;} ends the
1566: colon definition.
1567:
1568: The newly-defined word can be used like any other word, including using
1569: it in other definitions:
1570:
1571: @example
1572: : cubed ( n -- n^3 )
1573: dup squared * ;
1574: -5 cubed .
1575: : fourth-power ( n -- n^4 )
1576: squared squared ;
1577: 3 fourth-power .
1578: @end example
1579:
1.141 anton 1580: @quotation Assignment
1.48 anton 1581: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1582: @code{/mod} in terms of other Forth words, and check if they work (hint:
1583: test your tests on the originals first). Don't let the
1584: @samp{redefined}-Messages spook you, they are just warnings.
1.141 anton 1585: @end quotation
1.48 anton 1586:
1.66 anton 1587: Reference: @ref{Colon Definitions}.
1588:
1.48 anton 1589:
1590: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1591: @section Decompilation
1.66 anton 1592: @cindex decompilation tutorial
1593: @cindex see tutorial
1.48 anton 1594:
1595: You can decompile colon definitions with @code{see}:
1596:
1597: @example
1598: see squared
1599: see cubed
1600: @end example
1601:
1602: In Gforth @code{see} shows you a reconstruction of the source code from
1603: the executable code. Informations that were present in the source, but
1604: not in the executable code, are lost (e.g., comments).
1605:
1.65 anton 1606: You can also decompile the predefined words:
1607:
1608: @example
1609: see .
1610: see +
1611: @end example
1612:
1613:
1.48 anton 1614: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1615: @section Stack-Effect Comments
1.66 anton 1616: @cindex stack-effect comments, tutorial
1617: @cindex --, tutorial
1.48 anton 1618: By convention the comment after the name of a definition describes the
1.171 anton 1619: stack effect: The part in front of the @samp{--} describes the state of
1.48 anton 1620: the stack before the execution of the definition, i.e., the parameters
1621: that are passed into the colon definition; the part behind the @samp{--}
1622: is the state of the stack after the execution of the definition, i.e.,
1623: the results of the definition. The stack comment only shows the top
1624: stack items that the definition accesses and/or changes.
1625:
1626: You should put a correct stack effect on every definition, even if it is
1627: just @code{( -- )}. You should also add some descriptive comment to
1628: more complicated words (I usually do this in the lines following
1629: @code{:}). If you don't do this, your code becomes unreadable (because
1.117 anton 1630: you have to work through every definition before you can understand
1.48 anton 1631: any).
1632:
1.141 anton 1633: @quotation Assignment
1.48 anton 1634: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1635: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1636: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1.65 anton 1637: are done, you can compare your stack effects to those in this manual
1.48 anton 1638: (@pxref{Word Index}).
1.141 anton 1639: @end quotation
1.48 anton 1640:
1641: Sometimes programmers put comments at various places in colon
1642: definitions that describe the contents of the stack at that place (stack
1643: comments); i.e., they are like the first part of a stack-effect
1644: comment. E.g.,
1645:
1646: @example
1647: : cubed ( n -- n^3 )
1648: dup squared ( n n^2 ) * ;
1649: @end example
1650:
1651: In this case the stack comment is pretty superfluous, because the word
1652: is simple enough. If you think it would be a good idea to add such a
1653: comment to increase readability, you should also consider factoring the
1654: word into several simpler words (@pxref{Factoring Tutorial,,
1.60 anton 1655: Factoring}), which typically eliminates the need for the stack comment;
1.48 anton 1656: however, if you decide not to refactor it, then having such a comment is
1657: better than not having it.
1658:
1659: The names of the stack items in stack-effect and stack comments in the
1660: standard, in this manual, and in many programs specify the type through
1661: a type prefix, similar to Fortran and Hungarian notation. The most
1662: frequent prefixes are:
1663:
1664: @table @code
1665: @item n
1666: signed integer
1667: @item u
1668: unsigned integer
1669: @item c
1670: character
1671: @item f
1672: Boolean flags, i.e. @code{false} or @code{true}.
1673: @item a-addr,a-
1674: Cell-aligned address
1675: @item c-addr,c-
1676: Char-aligned address (note that a Char may have two bytes in Windows NT)
1677: @item xt
1678: Execution token, same size as Cell
1679: @item w,x
1680: Cell, can contain an integer or an address. It usually takes 32, 64 or
1681: 16 bits (depending on your platform and Forth system). A cell is more
1682: commonly known as machine word, but the term @emph{word} already means
1683: something different in Forth.
1684: @item d
1685: signed double-cell integer
1686: @item ud
1687: unsigned double-cell integer
1688: @item r
1689: Float (on the FP stack)
1690: @end table
1691:
1692: You can find a more complete list in @ref{Notation}.
1693:
1.141 anton 1694: @quotation Assignment
1.48 anton 1695: Write stack-effect comments for all definitions you have written up to
1696: now.
1.141 anton 1697: @end quotation
1.48 anton 1698:
1699:
1700: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1701: @section Types
1.66 anton 1702: @cindex types tutorial
1.48 anton 1703:
1704: In Forth the names of the operations are not overloaded; so similar
1705: operations on different types need different names; e.g., @code{+} adds
1706: integers, and you have to use @code{f+} to add floating-point numbers.
1707: The following prefixes are often used for related operations on
1708: different types:
1709:
1710: @table @code
1711: @item (none)
1712: signed integer
1713: @item u
1714: unsigned integer
1715: @item c
1716: character
1717: @item d
1718: signed double-cell integer
1719: @item ud, du
1720: unsigned double-cell integer
1721: @item 2
1722: two cells (not-necessarily double-cell numbers)
1723: @item m, um
1724: mixed single-cell and double-cell operations
1725: @item f
1726: floating-point (note that in stack comments @samp{f} represents flags,
1.66 anton 1727: and @samp{r} represents FP numbers).
1.48 anton 1728: @end table
1729:
1730: If there are no differences between the signed and the unsigned variant
1731: (e.g., for @code{+}), there is only the prefix-less variant.
1732:
1733: Forth does not perform type checking, neither at compile time, nor at
1734: run time. If you use the wrong oeration, the data are interpreted
1735: incorrectly:
1736:
1737: @example
1738: -1 u.
1739: @end example
1740:
1741: If you have only experience with type-checked languages until now, and
1742: have heard how important type-checking is, don't panic! In my
1743: experience (and that of other Forthers), type errors in Forth code are
1744: usually easy to find (once you get used to it), the increased vigilance
1745: of the programmer tends to catch some harder errors in addition to most
1746: type errors, and you never have to work around the type system, so in
1747: most situations the lack of type-checking seems to be a win (projects to
1748: add type checking to Forth have not caught on).
1749:
1750:
1751: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1752: @section Factoring
1.66 anton 1753: @cindex factoring tutorial
1.48 anton 1754:
1755: If you try to write longer definitions, you will soon find it hard to
1756: keep track of the stack contents. Therefore, good Forth programmers
1757: tend to write only short definitions (e.g., three lines). The art of
1758: finding meaningful short definitions is known as factoring (as in
1759: factoring polynomials).
1760:
1761: Well-factored programs offer additional advantages: smaller, more
1762: general words, are easier to test and debug and can be reused more and
1763: better than larger, specialized words.
1764:
1765: So, if you run into difficulties with stack management, when writing
1766: code, try to define meaningful factors for the word, and define the word
1767: in terms of those. Even if a factor contains only two words, it is
1768: often helpful.
1769:
1.65 anton 1770: Good factoring is not easy, and it takes some practice to get the knack
1771: for it; but even experienced Forth programmers often don't find the
1772: right solution right away, but only when rewriting the program. So, if
1773: you don't come up with a good solution immediately, keep trying, don't
1774: despair.
1.48 anton 1775:
1776: @c example !!
1777:
1778:
1779: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
1780: @section Designing the stack effect
1.66 anton 1781: @cindex Stack effect design, tutorial
1782: @cindex design of stack effects, tutorial
1.48 anton 1783:
1784: In other languages you can use an arbitrary order of parameters for a
1.65 anton 1785: function; and since there is only one result, you don't have to deal with
1.48 anton 1786: the order of results, either.
1787:
1.117 anton 1788: In Forth (and other stack-based languages, e.g., PostScript) the
1.48 anton 1789: parameter and result order of a definition is important and should be
1790: designed well. The general guideline is to design the stack effect such
1791: that the word is simple to use in most cases, even if that complicates
1792: the implementation of the word. Some concrete rules are:
1793:
1794: @itemize @bullet
1795:
1796: @item
1797: Words consume all of their parameters (e.g., @code{.}).
1798:
1799: @item
1800: If there is a convention on the order of parameters (e.g., from
1801: mathematics or another programming language), stick with it (e.g.,
1802: @code{-}).
1803:
1804: @item
1805: If one parameter usually requires only a short computation (e.g., it is
1806: a constant), pass it on the top of the stack. Conversely, parameters
1807: that usually require a long sequence of code to compute should be passed
1808: as the bottom (i.e., first) parameter. This makes the code easier to
1.171 anton 1809: read, because the reader does not need to keep track of the bottom item
1.48 anton 1810: through a long sequence of code (or, alternatively, through stack
1.49 anton 1811: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1.48 anton 1812: address on top of the stack because it is usually simpler to compute
1813: than the stored value (often the address is just a variable).
1814:
1815: @item
1816: Similarly, results that are usually consumed quickly should be returned
1817: on the top of stack, whereas a result that is often used in long
1818: computations should be passed as bottom result. E.g., the file words
1819: like @code{open-file} return the error code on the top of stack, because
1820: it is usually consumed quickly by @code{throw}; moreover, the error code
1821: has to be checked before doing anything with the other results.
1822:
1823: @end itemize
1824:
1825: These rules are just general guidelines, don't lose sight of the overall
1826: goal to make the words easy to use. E.g., if the convention rule
1827: conflicts with the computation-length rule, you might decide in favour
1828: of the convention if the word will be used rarely, and in favour of the
1829: computation-length rule if the word will be used frequently (because
1830: with frequent use the cost of breaking the computation-length rule would
1831: be quite high, and frequent use makes it easier to remember an
1832: unconventional order).
1833:
1834: @c example !! structure package
1835:
1.65 anton 1836:
1.48 anton 1837: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
1838: @section Local Variables
1.66 anton 1839: @cindex local variables, tutorial
1.48 anton 1840:
1841: You can define local variables (@emph{locals}) in a colon definition:
1842:
1843: @example
1844: : swap @{ a b -- b a @}
1845: b a ;
1846: 1 2 swap .s 2drop
1847: @end example
1848:
1849: (If your Forth system does not support this syntax, include
1850: @file{compat/anslocals.fs} first).
1851:
1852: In this example @code{@{ a b -- b a @}} is the locals definition; it
1853: takes two cells from the stack, puts the top of stack in @code{b} and
1854: the next stack element in @code{a}. @code{--} starts a comment ending
1855: with @code{@}}. After the locals definition, using the name of the
1856: local will push its value on the stack. You can leave the comment
1857: part (@code{-- b a}) away:
1858:
1859: @example
1860: : swap ( x1 x2 -- x2 x1 )
1861: @{ a b @} b a ;
1862: @end example
1863:
1864: In Gforth you can have several locals definitions, anywhere in a colon
1865: definition; in contrast, in a standard program you can have only one
1866: locals definition per colon definition, and that locals definition must
1.163 anton 1867: be outside any control structure.
1.48 anton 1868:
1869: With locals you can write slightly longer definitions without running
1870: into stack trouble. However, I recommend trying to write colon
1871: definitions without locals for exercise purposes to help you gain the
1872: essential factoring skills.
1873:
1.141 anton 1874: @quotation Assignment
1.48 anton 1875: Rewrite your definitions until now with locals
1.141 anton 1876: @end quotation
1.48 anton 1877:
1.66 anton 1878: Reference: @ref{Locals}.
1879:
1.48 anton 1880:
1881: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
1882: @section Conditional execution
1.66 anton 1883: @cindex conditionals, tutorial
1884: @cindex if, tutorial
1.48 anton 1885:
1886: In Forth you can use control structures only inside colon definitions.
1887: An @code{if}-structure looks like this:
1888:
1889: @example
1890: : abs ( n1 -- +n2 )
1891: dup 0 < if
1892: negate
1893: endif ;
1894: 5 abs .
1895: -5 abs .
1896: @end example
1897:
1898: @code{if} takes a flag from the stack. If the flag is non-zero (true),
1899: the following code is performed, otherwise execution continues after the
1.51 pazsan 1900: @code{endif} (or @code{else}). @code{<} compares the top two stack
1.171 anton 1901: elements and produces a flag:
1.48 anton 1902:
1903: @example
1904: 1 2 < .
1905: 2 1 < .
1906: 1 1 < .
1907: @end example
1908:
1909: Actually the standard name for @code{endif} is @code{then}. This
1910: tutorial presents the examples using @code{endif}, because this is often
1911: less confusing for people familiar with other programming languages
1912: where @code{then} has a different meaning. If your system does not have
1913: @code{endif}, define it with
1914:
1915: @example
1916: : endif postpone then ; immediate
1917: @end example
1918:
1919: You can optionally use an @code{else}-part:
1920:
1921: @example
1922: : min ( n1 n2 -- n )
1923: 2dup < if
1924: drop
1925: else
1926: nip
1927: endif ;
1928: 2 3 min .
1929: 3 2 min .
1930: @end example
1931:
1.141 anton 1932: @quotation Assignment
1.48 anton 1933: Write @code{min} without @code{else}-part (hint: what's the definition
1934: of @code{nip}?).
1.141 anton 1935: @end quotation
1.48 anton 1936:
1.66 anton 1937: Reference: @ref{Selection}.
1938:
1.48 anton 1939:
1940: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
1941: @section Flags and Comparisons
1.66 anton 1942: @cindex flags tutorial
1943: @cindex comparison tutorial
1.48 anton 1944:
1945: In a false-flag all bits are clear (0 when interpreted as integer). In
1946: a canonical true-flag all bits are set (-1 as a twos-complement signed
1947: integer); in many contexts (e.g., @code{if}) any non-zero value is
1948: treated as true flag.
1949:
1950: @example
1951: false .
1952: true .
1953: true hex u. decimal
1954: @end example
1955:
1956: Comparison words produce canonical flags:
1957:
1958: @example
1959: 1 1 = .
1960: 1 0= .
1961: 0 1 < .
1962: 0 0 < .
1963: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
1964: -1 1 < .
1965: @end example
1966:
1.66 anton 1967: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
1968: (or none) and the comparisons @code{= <> < > <= >=}. Only a part of
1969: these combinations are standard (for details see the standard,
1970: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
1.48 anton 1971:
1.171 anton 1972: You can use @code{and or xor invert} as operations on canonical flags.
1973: Actually they are bitwise operations:
1.48 anton 1974:
1975: @example
1976: 1 2 and .
1977: 1 2 or .
1978: 1 3 xor .
1979: 1 invert .
1980: @end example
1981:
1982: You can convert a zero/non-zero flag into a canonical flag with
1983: @code{0<>} (and complement it on the way with @code{0=}).
1984:
1985: @example
1986: 1 0= .
1987: 1 0<> .
1988: @end example
1989:
1.65 anton 1990: You can use the all-bits-set feature of canonical flags and the bitwise
1.48 anton 1991: operation of the Boolean operations to avoid @code{if}s:
1992:
1993: @example
1994: : foo ( n1 -- n2 )
1995: 0= if
1996: 14
1997: else
1998: 0
1999: endif ;
2000: 0 foo .
2001: 1 foo .
2002:
2003: : foo ( n1 -- n2 )
2004: 0= 14 and ;
2005: 0 foo .
2006: 1 foo .
2007: @end example
2008:
1.141 anton 2009: @quotation Assignment
1.48 anton 2010: Write @code{min} without @code{if}.
1.141 anton 2011: @end quotation
1.48 anton 2012:
1.66 anton 2013: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
2014: @ref{Bitwise operations}.
2015:
1.48 anton 2016:
2017: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2018: @section General Loops
1.66 anton 2019: @cindex loops, indefinite, tutorial
1.48 anton 2020:
2021: The endless loop is the most simple one:
2022:
2023: @example
2024: : endless ( -- )
2025: 0 begin
2026: dup . 1+
2027: again ;
2028: endless
2029: @end example
2030:
2031: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2032: does nothing at run-time, @code{again} jumps back to @code{begin}.
2033:
2034: A loop with one exit at any place looks like this:
2035:
2036: @example
2037: : log2 ( +n1 -- n2 )
2038: \ logarithmus dualis of n1>0, rounded down to the next integer
2039: assert( dup 0> )
2040: 2/ 0 begin
2041: over 0> while
2042: 1+ swap 2/ swap
2043: repeat
2044: nip ;
2045: 7 log2 .
2046: 8 log2 .
2047: @end example
2048:
2049: At run-time @code{while} consumes a flag; if it is 0, execution
1.51 pazsan 2050: continues behind the @code{repeat}; if the flag is non-zero, execution
1.48 anton 2051: continues behind the @code{while}. @code{Repeat} jumps back to
2052: @code{begin}, just like @code{again}.
2053:
2054: In Forth there are many combinations/abbreviations, like @code{1+}.
1.90 anton 2055: However, @code{2/} is not one of them; it shifts its argument right by
1.48 anton 2056: one bit (arithmetic shift right):
2057:
2058: @example
2059: -5 2 / .
2060: -5 2/ .
2061: @end example
2062:
2063: @code{assert(} is no standard word, but you can get it on systems other
2064: then Gforth by including @file{compat/assert.fs}. You can see what it
2065: does by trying
2066:
2067: @example
2068: 0 log2 .
2069: @end example
2070:
2071: Here's a loop with an exit at the end:
2072:
2073: @example
2074: : log2 ( +n1 -- n2 )
2075: \ logarithmus dualis of n1>0, rounded down to the next integer
2076: assert( dup 0 > )
2077: -1 begin
2078: 1+ swap 2/ swap
2079: over 0 <=
2080: until
2081: nip ;
2082: @end example
2083:
2084: @code{Until} consumes a flag; if it is non-zero, execution continues at
2085: the @code{begin}, otherwise after the @code{until}.
2086:
1.141 anton 2087: @quotation Assignment
1.48 anton 2088: Write a definition for computing the greatest common divisor.
1.141 anton 2089: @end quotation
1.48 anton 2090:
1.66 anton 2091: Reference: @ref{Simple Loops}.
2092:
1.48 anton 2093:
2094: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2095: @section Counted loops
1.66 anton 2096: @cindex loops, counted, tutorial
1.48 anton 2097:
2098: @example
2099: : ^ ( n1 u -- n )
1.171 anton 2100: \ n = the uth power of n1
1.48 anton 2101: 1 swap 0 u+do
2102: over *
2103: loop
2104: nip ;
2105: 3 2 ^ .
2106: 4 3 ^ .
2107: @end example
2108:
2109: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2110: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2111: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2112: times (or not at all, if @code{u3-u4<0}).
2113:
2114: You can see the stack effect design rules at work in the stack effect of
2115: the loop start words: Since the start value of the loop is more
2116: frequently constant than the end value, the start value is passed on
2117: the top-of-stack.
2118:
2119: You can access the counter of a counted loop with @code{i}:
2120:
2121: @example
2122: : fac ( u -- u! )
2123: 1 swap 1+ 1 u+do
2124: i *
2125: loop ;
2126: 5 fac .
2127: 7 fac .
2128: @end example
2129:
2130: There is also @code{+do}, which expects signed numbers (important for
2131: deciding whether to enter the loop).
2132:
1.141 anton 2133: @quotation Assignment
1.48 anton 2134: Write a definition for computing the nth Fibonacci number.
1.141 anton 2135: @end quotation
1.48 anton 2136:
1.65 anton 2137: You can also use increments other than 1:
2138:
2139: @example
2140: : up2 ( n1 n2 -- )
2141: +do
2142: i .
2143: 2 +loop ;
2144: 10 0 up2
2145:
2146: : down2 ( n1 n2 -- )
2147: -do
2148: i .
2149: 2 -loop ;
2150: 0 10 down2
2151: @end example
1.48 anton 2152:
1.66 anton 2153: Reference: @ref{Counted Loops}.
2154:
1.48 anton 2155:
2156: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2157: @section Recursion
1.66 anton 2158: @cindex recursion tutorial
1.48 anton 2159:
2160: Usually the name of a definition is not visible in the definition; but
2161: earlier definitions are usually visible:
2162:
2163: @example
1.166 anton 2164: 1 0 / . \ "Floating-point unidentified fault" in Gforth on some platforms
1.48 anton 2165: : / ( n1 n2 -- n )
2166: dup 0= if
2167: -10 throw \ report division by zero
2168: endif
2169: / \ old version
2170: ;
2171: 1 0 /
2172: @end example
2173:
2174: For recursive definitions you can use @code{recursive} (non-standard) or
2175: @code{recurse}:
2176:
2177: @example
2178: : fac1 ( n -- n! ) recursive
2179: dup 0> if
2180: dup 1- fac1 *
2181: else
2182: drop 1
2183: endif ;
2184: 7 fac1 .
2185:
2186: : fac2 ( n -- n! )
2187: dup 0> if
2188: dup 1- recurse *
2189: else
2190: drop 1
2191: endif ;
2192: 8 fac2 .
2193: @end example
2194:
1.141 anton 2195: @quotation Assignment
1.48 anton 2196: Write a recursive definition for computing the nth Fibonacci number.
1.141 anton 2197: @end quotation
1.48 anton 2198:
1.66 anton 2199: Reference (including indirect recursion): @xref{Calls and returns}.
2200:
1.48 anton 2201:
2202: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2203: @section Leaving definitions or loops
1.66 anton 2204: @cindex leaving definitions, tutorial
2205: @cindex leaving loops, tutorial
1.48 anton 2206:
2207: @code{EXIT} exits the current definition right away. For every counted
2208: loop that is left in this way, an @code{UNLOOP} has to be performed
2209: before the @code{EXIT}:
2210:
2211: @c !! real examples
2212: @example
2213: : ...
2214: ... u+do
2215: ... if
2216: ... unloop exit
2217: endif
2218: ...
2219: loop
2220: ... ;
2221: @end example
2222:
2223: @code{LEAVE} leaves the innermost counted loop right away:
2224:
2225: @example
2226: : ...
2227: ... u+do
2228: ... if
2229: ... leave
2230: endif
2231: ...
2232: loop
2233: ... ;
2234: @end example
2235:
1.65 anton 2236: @c !! example
1.48 anton 2237:
1.66 anton 2238: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
2239:
2240:
1.48 anton 2241: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2242: @section Return Stack
1.66 anton 2243: @cindex return stack tutorial
1.48 anton 2244:
2245: In addition to the data stack Forth also has a second stack, the return
2246: stack; most Forth systems store the return addresses of procedure calls
2247: there (thus its name). Programmers can also use this stack:
2248:
2249: @example
2250: : foo ( n1 n2 -- )
2251: .s
2252: >r .s
1.50 anton 2253: r@@ .
1.48 anton 2254: >r .s
1.50 anton 2255: r@@ .
1.48 anton 2256: r> .
1.50 anton 2257: r@@ .
1.48 anton 2258: r> . ;
2259: 1 2 foo
2260: @end example
2261:
2262: @code{>r} takes an element from the data stack and pushes it onto the
2263: return stack; conversely, @code{r>} moves an elementm from the return to
2264: the data stack; @code{r@@} pushes a copy of the top of the return stack
1.148 anton 2265: on the data stack.
1.48 anton 2266:
2267: Forth programmers usually use the return stack for storing data
2268: temporarily, if using the data stack alone would be too complex, and
2269: factoring and locals are not an option:
2270:
2271: @example
2272: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2273: rot >r rot r> ;
2274: @end example
2275:
2276: The return address of the definition and the loop control parameters of
2277: counted loops usually reside on the return stack, so you have to take
2278: all items, that you have pushed on the return stack in a colon
2279: definition or counted loop, from the return stack before the definition
2280: or loop ends. You cannot access items that you pushed on the return
2281: stack outside some definition or loop within the definition of loop.
2282:
2283: If you miscount the return stack items, this usually ends in a crash:
2284:
2285: @example
2286: : crash ( n -- )
2287: >r ;
2288: 5 crash
2289: @end example
2290:
2291: You cannot mix using locals and using the return stack (according to the
2292: standard; Gforth has no problem). However, they solve the same
2293: problems, so this shouldn't be an issue.
2294:
1.141 anton 2295: @quotation Assignment
1.48 anton 2296: Can you rewrite any of the definitions you wrote until now in a better
2297: way using the return stack?
1.141 anton 2298: @end quotation
1.48 anton 2299:
1.66 anton 2300: Reference: @ref{Return stack}.
2301:
1.48 anton 2302:
2303: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2304: @section Memory
1.66 anton 2305: @cindex memory access/allocation tutorial
1.48 anton 2306:
2307: You can create a global variable @code{v} with
2308:
2309: @example
2310: variable v ( -- addr )
2311: @end example
2312:
2313: @code{v} pushes the address of a cell in memory on the stack. This cell
2314: was reserved by @code{variable}. You can use @code{!} (store) to store
2315: values into this cell and @code{@@} (fetch) to load the value from the
2316: stack into memory:
2317:
2318: @example
2319: v .
2320: 5 v ! .s
1.50 anton 2321: v @@ .
1.48 anton 2322: @end example
2323:
1.65 anton 2324: You can see a raw dump of memory with @code{dump}:
2325:
2326: @example
2327: v 1 cells .s dump
2328: @end example
2329:
2330: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2331: generally, address units (aus)) that @code{n1 cells} occupy. You can
2332: also reserve more memory:
1.48 anton 2333:
2334: @example
2335: create v2 20 cells allot
1.65 anton 2336: v2 20 cells dump
1.48 anton 2337: @end example
2338:
1.65 anton 2339: creates a word @code{v2} and reserves 20 uninitialized cells; the
2340: address pushed by @code{v2} points to the start of these 20 cells. You
2341: can use address arithmetic to access these cells:
1.48 anton 2342:
2343: @example
2344: 3 v2 5 cells + !
1.65 anton 2345: v2 20 cells dump
1.48 anton 2346: @end example
2347:
2348: You can reserve and initialize memory with @code{,}:
2349:
2350: @example
2351: create v3
2352: 5 , 4 , 3 , 2 , 1 ,
1.50 anton 2353: v3 @@ .
2354: v3 cell+ @@ .
2355: v3 2 cells + @@ .
1.65 anton 2356: v3 5 cells dump
1.48 anton 2357: @end example
2358:
1.141 anton 2359: @quotation Assignment
1.48 anton 2360: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2361: @code{u} cells, with the first of these cells at @code{addr}, the next
2362: one at @code{addr cell+} etc.
1.141 anton 2363: @end quotation
1.48 anton 2364:
2365: You can also reserve memory without creating a new word:
2366:
2367: @example
1.60 anton 2368: here 10 cells allot .
2369: here .
1.48 anton 2370: @end example
2371:
2372: @code{Here} pushes the start address of the memory area. You should
2373: store it somewhere, or you will have a hard time finding the memory area
2374: again.
2375:
2376: @code{Allot} manages dictionary memory. The dictionary memory contains
2377: the system's data structures for words etc. on Gforth and most other
2378: Forth systems. It is managed like a stack: You can free the memory that
2379: you have just @code{allot}ed with
2380:
2381: @example
2382: -10 cells allot
1.60 anton 2383: here .
1.48 anton 2384: @end example
2385:
2386: Note that you cannot do this if you have created a new word in the
2387: meantime (because then your @code{allot}ed memory is no longer on the
2388: top of the dictionary ``stack'').
2389:
2390: Alternatively, you can use @code{allocate} and @code{free} which allow
2391: freeing memory in any order:
2392:
2393: @example
2394: 10 cells allocate throw .s
2395: 20 cells allocate throw .s
2396: swap
2397: free throw
2398: free throw
2399: @end example
2400:
2401: The @code{throw}s deal with errors (e.g., out of memory).
2402:
1.65 anton 2403: And there is also a
2404: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2405: garbage collector}, which eliminates the need to @code{free} memory
2406: explicitly.
1.48 anton 2407:
1.66 anton 2408: Reference: @ref{Memory}.
2409:
1.48 anton 2410:
2411: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2412: @section Characters and Strings
1.66 anton 2413: @cindex strings tutorial
2414: @cindex characters tutorial
1.48 anton 2415:
2416: On the stack characters take up a cell, like numbers. In memory they
2417: have their own size (one 8-bit byte on most systems), and therefore
2418: require their own words for memory access:
2419:
2420: @example
2421: create v4
2422: 104 c, 97 c, 108 c, 108 c, 111 c,
1.50 anton 2423: v4 4 chars + c@@ .
1.65 anton 2424: v4 5 chars dump
1.48 anton 2425: @end example
2426:
2427: The preferred representation of strings on the stack is @code{addr
2428: u-count}, where @code{addr} is the address of the first character and
2429: @code{u-count} is the number of characters in the string.
2430:
2431: @example
2432: v4 5 type
2433: @end example
2434:
2435: You get a string constant with
2436:
2437: @example
2438: s" hello, world" .s
2439: type
2440: @end example
2441:
2442: Make sure you have a space between @code{s"} and the string; @code{s"}
2443: is a normal Forth word and must be delimited with white space (try what
2444: happens when you remove the space).
2445:
2446: However, this interpretive use of @code{s"} is quite restricted: the
2447: string exists only until the next call of @code{s"} (some Forth systems
2448: keep more than one of these strings, but usually they still have a
1.62 crook 2449: limited lifetime).
1.48 anton 2450:
2451: @example
2452: s" hello," s" world" .s
2453: type
2454: type
2455: @end example
2456:
1.62 crook 2457: You can also use @code{s"} in a definition, and the resulting
2458: strings then live forever (well, for as long as the definition):
1.48 anton 2459:
2460: @example
2461: : foo s" hello," s" world" ;
2462: foo .s
2463: type
2464: type
2465: @end example
2466:
1.141 anton 2467: @quotation Assignment
1.48 anton 2468: @code{Emit ( c -- )} types @code{c} as character (not a number).
2469: Implement @code{type ( addr u -- )}.
1.141 anton 2470: @end quotation
1.48 anton 2471:
1.66 anton 2472: Reference: @ref{Memory Blocks}.
2473:
2474:
1.84 pazsan 2475: @node Alignment Tutorial, Files Tutorial, Characters and Strings Tutorial, Tutorial
1.48 anton 2476: @section Alignment
1.66 anton 2477: @cindex alignment tutorial
2478: @cindex memory alignment tutorial
1.48 anton 2479:
2480: On many processors cells have to be aligned in memory, if you want to
2481: access them with @code{@@} and @code{!} (and even if the processor does
1.62 crook 2482: not require alignment, access to aligned cells is faster).
1.48 anton 2483:
2484: @code{Create} aligns @code{here} (i.e., the place where the next
2485: allocation will occur, and that the @code{create}d word points to).
2486: Likewise, the memory produced by @code{allocate} starts at an aligned
2487: address. Adding a number of @code{cells} to an aligned address produces
2488: another aligned address.
2489:
2490: However, address arithmetic involving @code{char+} and @code{chars} can
2491: create an address that is not cell-aligned. @code{Aligned ( addr --
2492: a-addr )} produces the next aligned address:
2493:
2494: @example
1.50 anton 2495: v3 char+ aligned .s @@ .
2496: v3 char+ .s @@ .
1.48 anton 2497: @end example
2498:
2499: Similarly, @code{align} advances @code{here} to the next aligned
2500: address:
2501:
2502: @example
2503: create v5 97 c,
2504: here .
2505: align here .
2506: 1000 ,
2507: @end example
2508:
2509: Note that you should use aligned addresses even if your processor does
2510: not require them, if you want your program to be portable.
2511:
1.66 anton 2512: Reference: @ref{Address arithmetic}.
2513:
1.48 anton 2514:
1.84 pazsan 2515: @node Files Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Alignment Tutorial, Tutorial
2516: @section Files
2517: @cindex files tutorial
2518:
2519: This section gives a short introduction into how to use files inside
2520: Forth. It's broken up into five easy steps:
2521:
2522: @enumerate 1
2523: @item Opened an ASCII text file for input
2524: @item Opened a file for output
2525: @item Read input file until string matched (or some other condition matched)
2526: @item Wrote some lines from input ( modified or not) to output
2527: @item Closed the files.
2528: @end enumerate
2529:
1.153 anton 2530: Reference: @ref{General files}.
2531:
1.84 pazsan 2532: @subsection Open file for input
2533:
2534: @example
2535: s" foo.in" r/o open-file throw Value fd-in
2536: @end example
2537:
2538: @subsection Create file for output
2539:
2540: @example
2541: s" foo.out" w/o create-file throw Value fd-out
2542: @end example
2543:
2544: The available file modes are r/o for read-only access, r/w for
2545: read-write access, and w/o for write-only access. You could open both
2546: files with r/w, too, if you like. All file words return error codes; for
2547: most applications, it's best to pass there error codes with @code{throw}
2548: to the outer error handler.
2549:
2550: If you want words for opening and assigning, define them as follows:
2551:
2552: @example
2553: 0 Value fd-in
2554: 0 Value fd-out
2555: : open-input ( addr u -- ) r/o open-file throw to fd-in ;
2556: : open-output ( addr u -- ) w/o create-file throw to fd-out ;
2557: @end example
2558:
2559: Usage example:
2560:
2561: @example
2562: s" foo.in" open-input
2563: s" foo.out" open-output
2564: @end example
2565:
2566: @subsection Scan file for a particular line
2567:
2568: @example
2569: 256 Constant max-line
2570: Create line-buffer max-line 2 + allot
2571:
2572: : scan-file ( addr u -- )
2573: begin
2574: line-buffer max-line fd-in read-line throw
2575: while
2576: >r 2dup line-buffer r> compare 0=
2577: until
2578: else
2579: drop
2580: then
2581: 2drop ;
2582: @end example
2583:
2584: @code{read-line ( addr u1 fd -- u2 flag ior )} reads up to u1 bytes into
1.94 anton 2585: the buffer at addr, and returns the number of bytes read, a flag that is
2586: false when the end of file is reached, and an error code.
1.84 pazsan 2587:
2588: @code{compare ( addr1 u1 addr2 u2 -- n )} compares two strings and
2589: returns zero if both strings are equal. It returns a positive number if
2590: the first string is lexically greater, a negative if the second string
2591: is lexically greater.
2592:
2593: We haven't seen this loop here; it has two exits. Since the @code{while}
2594: exits with the number of bytes read on the stack, we have to clean up
2595: that separately; that's after the @code{else}.
2596:
2597: Usage example:
2598:
2599: @example
2600: s" The text I search is here" scan-file
2601: @end example
2602:
2603: @subsection Copy input to output
2604:
2605: @example
2606: : copy-file ( -- )
2607: begin
2608: line-buffer max-line fd-in read-line throw
2609: while
2610: line-buffer swap fd-out write-file throw
2611: repeat ;
2612: @end example
2613:
2614: @subsection Close files
2615:
2616: @example
2617: fd-in close-file throw
2618: fd-out close-file throw
2619: @end example
2620:
2621: Likewise, you can put that into definitions, too:
2622:
2623: @example
2624: : close-input ( -- ) fd-in close-file throw ;
2625: : close-output ( -- ) fd-out close-file throw ;
2626: @end example
2627:
1.141 anton 2628: @quotation Assignment
1.84 pazsan 2629: How could you modify @code{copy-file} so that it copies until a second line is
2630: matched? Can you write a program that extracts a section of a text file,
2631: given the line that starts and the line that terminates that section?
1.141 anton 2632: @end quotation
1.84 pazsan 2633:
2634: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Files Tutorial, Tutorial
1.48 anton 2635: @section Interpretation and Compilation Semantics and Immediacy
1.66 anton 2636: @cindex semantics tutorial
2637: @cindex interpretation semantics tutorial
2638: @cindex compilation semantics tutorial
2639: @cindex immediate, tutorial
1.48 anton 2640:
2641: When a word is compiled, it behaves differently from being interpreted.
2642: E.g., consider @code{+}:
2643:
2644: @example
2645: 1 2 + .
2646: : foo + ;
2647: @end example
2648:
2649: These two behaviours are known as compilation and interpretation
2650: semantics. For normal words (e.g., @code{+}), the compilation semantics
2651: is to append the interpretation semantics to the currently defined word
2652: (@code{foo} in the example above). I.e., when @code{foo} is executed
2653: later, the interpretation semantics of @code{+} (i.e., adding two
2654: numbers) will be performed.
2655:
2656: However, there are words with non-default compilation semantics, e.g.,
2657: the control-flow words like @code{if}. You can use @code{immediate} to
2658: change the compilation semantics of the last defined word to be equal to
2659: the interpretation semantics:
2660:
2661: @example
2662: : [FOO] ( -- )
2663: 5 . ; immediate
2664:
2665: [FOO]
2666: : bar ( -- )
2667: [FOO] ;
2668: bar
2669: see bar
2670: @end example
2671:
2672: Two conventions to mark words with non-default compilation semnatics are
2673: names with brackets (more frequently used) and to write them all in
2674: upper case (less frequently used).
2675:
2676: In Gforth (and many other systems) you can also remove the
2677: interpretation semantics with @code{compile-only} (the compilation
2678: semantics is derived from the original interpretation semantics):
2679:
2680: @example
2681: : flip ( -- )
2682: 6 . ; compile-only \ but not immediate
2683: flip
2684:
2685: : flop ( -- )
2686: flip ;
2687: flop
2688: @end example
2689:
2690: In this example the interpretation semantics of @code{flop} is equal to
2691: the original interpretation semantics of @code{flip}.
2692:
2693: The text interpreter has two states: in interpret state, it performs the
2694: interpretation semantics of words it encounters; in compile state, it
2695: performs the compilation semantics of these words.
2696:
2697: Among other things, @code{:} switches into compile state, and @code{;}
2698: switches back to interpret state. They contain the factors @code{]}
2699: (switch to compile state) and @code{[} (switch to interpret state), that
2700: do nothing but switch the state.
2701:
2702: @example
2703: : xxx ( -- )
2704: [ 5 . ]
2705: ;
2706:
2707: xxx
2708: see xxx
2709: @end example
2710:
2711: These brackets are also the source of the naming convention mentioned
2712: above.
2713:
1.66 anton 2714: Reference: @ref{Interpretation and Compilation Semantics}.
2715:
1.48 anton 2716:
2717: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2718: @section Execution Tokens
1.66 anton 2719: @cindex execution tokens tutorial
2720: @cindex XT tutorial
1.48 anton 2721:
2722: @code{' word} gives you the execution token (XT) of a word. The XT is a
2723: cell representing the interpretation semantics of a word. You can
2724: execute this semantics with @code{execute}:
2725:
2726: @example
2727: ' + .s
2728: 1 2 rot execute .
2729: @end example
2730:
2731: The XT is similar to a function pointer in C. However, parameter
2732: passing through the stack makes it a little more flexible:
2733:
2734: @example
2735: : map-array ( ... addr u xt -- ... )
1.50 anton 2736: \ executes xt ( ... x -- ... ) for every element of the array starting
2737: \ at addr and containing u elements
1.48 anton 2738: @{ xt @}
2739: cells over + swap ?do
1.50 anton 2740: i @@ xt execute
1.48 anton 2741: 1 cells +loop ;
2742:
2743: create a 3 , 4 , 2 , -1 , 4 ,
2744: a 5 ' . map-array .s
2745: 0 a 5 ' + map-array .
2746: s" max-n" environment? drop .s
2747: a 5 ' min map-array .
2748: @end example
2749:
2750: You can use map-array with the XTs of words that consume one element
2751: more than they produce. In theory you can also use it with other XTs,
2752: but the stack effect then depends on the size of the array, which is
2753: hard to understand.
2754:
1.51 pazsan 2755: Since XTs are cell-sized, you can store them in memory and manipulate
2756: them on the stack like other cells. You can also compile the XT into a
1.48 anton 2757: word with @code{compile,}:
2758:
2759: @example
2760: : foo1 ( n1 n2 -- n )
2761: [ ' + compile, ] ;
2762: see foo
2763: @end example
2764:
2765: This is non-standard, because @code{compile,} has no compilation
2766: semantics in the standard, but it works in good Forth systems. For the
2767: broken ones, use
2768:
2769: @example
2770: : [compile,] compile, ; immediate
2771:
2772: : foo1 ( n1 n2 -- n )
2773: [ ' + ] [compile,] ;
2774: see foo
2775: @end example
2776:
2777: @code{'} is a word with default compilation semantics; it parses the
2778: next word when its interpretation semantics are executed, not during
2779: compilation:
2780:
2781: @example
2782: : foo ( -- xt )
2783: ' ;
2784: see foo
2785: : bar ( ... "word" -- ... )
2786: ' execute ;
2787: see bar
1.60 anton 2788: 1 2 bar + .
1.48 anton 2789: @end example
2790:
2791: You often want to parse a word during compilation and compile its XT so
2792: it will be pushed on the stack at run-time. @code{[']} does this:
2793:
2794: @example
2795: : xt-+ ( -- xt )
2796: ['] + ;
2797: see xt-+
2798: 1 2 xt-+ execute .
2799: @end example
2800:
2801: Many programmers tend to see @code{'} and the word it parses as one
2802: unit, and expect it to behave like @code{[']} when compiled, and are
2803: confused by the actual behaviour. If you are, just remember that the
2804: Forth system just takes @code{'} as one unit and has no idea that it is
2805: a parsing word (attempts to convenience programmers in this issue have
2806: usually resulted in even worse pitfalls, see
1.66 anton 2807: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
2808: @code{State}-smartness---Why it is evil and How to Exorcise it}).
1.48 anton 2809:
2810: Note that the state of the interpreter does not come into play when
1.51 pazsan 2811: creating and executing XTs. I.e., even when you execute @code{'} in
1.48 anton 2812: compile state, it still gives you the interpretation semantics. And
2813: whatever that state is, @code{execute} performs the semantics
1.66 anton 2814: represented by the XT (i.e., for XTs produced with @code{'} the
2815: interpretation semantics).
2816:
2817: Reference: @ref{Tokens for Words}.
1.48 anton 2818:
2819:
2820: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
2821: @section Exceptions
1.66 anton 2822: @cindex exceptions tutorial
1.48 anton 2823:
2824: @code{throw ( n -- )} causes an exception unless n is zero.
2825:
2826: @example
2827: 100 throw .s
2828: 0 throw .s
2829: @end example
2830:
2831: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
2832: it catches exceptions and pushes the number of the exception on the
2833: stack (or 0, if the xt executed without exception). If there was an
2834: exception, the stacks have the same depth as when entering @code{catch}:
2835:
2836: @example
2837: .s
2838: 3 0 ' / catch .s
2839: 3 2 ' / catch .s
2840: @end example
2841:
1.141 anton 2842: @quotation Assignment
1.48 anton 2843: Try the same with @code{execute} instead of @code{catch}.
1.141 anton 2844: @end quotation
1.48 anton 2845:
2846: @code{Throw} always jumps to the dynamically next enclosing
2847: @code{catch}, even if it has to leave several call levels to achieve
2848: this:
2849:
2850: @example
2851: : foo 100 throw ;
2852: : foo1 foo ." after foo" ;
1.51 pazsan 2853: : bar ['] foo1 catch ;
1.60 anton 2854: bar .
1.48 anton 2855: @end example
2856:
2857: It is often important to restore a value upon leaving a definition, even
2858: if the definition is left through an exception. You can ensure this
2859: like this:
2860:
2861: @example
2862: : ...
2863: save-x
1.51 pazsan 2864: ['] word-changing-x catch ( ... n )
1.48 anton 2865: restore-x
2866: ( ... n ) throw ;
2867: @end example
2868:
1.172 anton 2869: However, this is still not safe against, e.g., the user pressing
2870: @kbd{Ctrl-C} when execution is between the @code{catch} and
2871: @code{restore-x}.
2872:
2873: Gforth provides an alternative exception handling syntax that is safe
2874: against such cases: @code{try ... restore ... endtry}. If the code
2875: between @code{try} and @code{endtry} has an exception, the stack
2876: depths are restored, the exception number is pushed on the stack, and
2877: the execution continues right after @code{restore}.
1.48 anton 2878:
1.172 anton 2879: The safer equivalent to the restoration code above is
1.48 anton 2880:
2881: @example
2882: : ...
2883: save-x
2884: try
1.92 anton 2885: word-changing-x 0
1.172 anton 2886: restore
2887: restore-x
2888: endtry
1.48 anton 2889: throw ;
2890: @end example
2891:
1.66 anton 2892: Reference: @ref{Exception Handling}.
2893:
1.48 anton 2894:
2895: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
2896: @section Defining Words
1.66 anton 2897: @cindex defining words tutorial
2898: @cindex does> tutorial
2899: @cindex create...does> tutorial
2900:
2901: @c before semantics?
1.48 anton 2902:
2903: @code{:}, @code{create}, and @code{variable} are definition words: They
2904: define other words. @code{Constant} is another definition word:
2905:
2906: @example
2907: 5 constant foo
2908: foo .
2909: @end example
2910:
2911: You can also use the prefixes @code{2} (double-cell) and @code{f}
2912: (floating point) with @code{variable} and @code{constant}.
2913:
2914: You can also define your own defining words. E.g.:
2915:
2916: @example
2917: : variable ( "name" -- )
2918: create 0 , ;
2919: @end example
2920:
2921: You can also define defining words that create words that do something
2922: other than just producing their address:
2923:
2924: @example
2925: : constant ( n "name" -- )
2926: create ,
2927: does> ( -- n )
1.50 anton 2928: ( addr ) @@ ;
1.48 anton 2929:
2930: 5 constant foo
2931: foo .
2932: @end example
2933:
2934: The definition of @code{constant} above ends at the @code{does>}; i.e.,
2935: @code{does>} replaces @code{;}, but it also does something else: It
2936: changes the last defined word such that it pushes the address of the
2937: body of the word and then performs the code after the @code{does>}
2938: whenever it is called.
2939:
2940: In the example above, @code{constant} uses @code{,} to store 5 into the
2941: body of @code{foo}. When @code{foo} executes, it pushes the address of
2942: the body onto the stack, then (in the code after the @code{does>})
2943: fetches the 5 from there.
2944:
2945: The stack comment near the @code{does>} reflects the stack effect of the
2946: defined word, not the stack effect of the code after the @code{does>}
2947: (the difference is that the code expects the address of the body that
2948: the stack comment does not show).
2949:
2950: You can use these definition words to do factoring in cases that involve
2951: (other) definition words. E.g., a field offset is always added to an
2952: address. Instead of defining
2953:
2954: @example
2955: 2 cells constant offset-field1
2956: @end example
2957:
2958: and using this like
2959:
2960: @example
2961: ( addr ) offset-field1 +
2962: @end example
2963:
2964: you can define a definition word
2965:
2966: @example
2967: : simple-field ( n "name" -- )
2968: create ,
2969: does> ( n1 -- n1+n )
1.50 anton 2970: ( addr ) @@ + ;
1.48 anton 2971: @end example
1.21 crook 2972:
1.48 anton 2973: Definition and use of field offsets now look like this:
1.21 crook 2974:
1.48 anton 2975: @example
2976: 2 cells simple-field field1
1.60 anton 2977: create mystruct 4 cells allot
2978: mystruct .s field1 .s drop
1.48 anton 2979: @end example
1.21 crook 2980:
1.48 anton 2981: If you want to do something with the word without performing the code
2982: after the @code{does>}, you can access the body of a @code{create}d word
2983: with @code{>body ( xt -- addr )}:
1.21 crook 2984:
1.48 anton 2985: @example
2986: : value ( n "name" -- )
2987: create ,
2988: does> ( -- n1 )
1.50 anton 2989: @@ ;
1.48 anton 2990: : to ( n "name" -- )
2991: ' >body ! ;
1.21 crook 2992:
1.48 anton 2993: 5 value foo
2994: foo .
2995: 7 to foo
2996: foo .
2997: @end example
1.21 crook 2998:
1.141 anton 2999: @quotation Assignment
1.48 anton 3000: Define @code{defer ( "name" -- )}, which creates a word that stores an
3001: XT (at the start the XT of @code{abort}), and upon execution
3002: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3003: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3004: recursion is one application of @code{defer}.
1.141 anton 3005: @end quotation
1.29 crook 3006:
1.66 anton 3007: Reference: @ref{User-defined Defining Words}.
3008:
3009:
1.48 anton 3010: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3011: @section Arrays and Records
1.66 anton 3012: @cindex arrays tutorial
3013: @cindex records tutorial
3014: @cindex structs tutorial
1.29 crook 3015:
1.48 anton 3016: Forth has no standard words for defining data structures such as arrays
3017: and records (structs in C terminology), but you can build them yourself
3018: based on address arithmetic. You can also define words for defining
3019: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
1.29 crook 3020:
1.48 anton 3021: One of the first projects a Forth newcomer sets out upon when learning
3022: about defining words is an array defining word (possibly for
3023: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3024: learn something from it. However, don't be disappointed when you later
3025: learn that you have little use for these words (inappropriate use would
3026: be even worse). I have not yet found a set of useful array words yet;
3027: the needs are just too diverse, and named, global arrays (the result of
3028: naive use of defining words) are often not flexible enough (e.g.,
1.66 anton 3029: consider how to pass them as parameters). Another such project is a set
3030: of words to help dealing with strings.
1.29 crook 3031:
1.48 anton 3032: On the other hand, there is a useful set of record words, and it has
3033: been defined in @file{compat/struct.fs}; these words are predefined in
3034: Gforth. They are explained in depth elsewhere in this manual (see
3035: @pxref{Structures}). The @code{simple-field} example above is
3036: simplified variant of fields in this package.
1.21 crook 3037:
3038:
1.48 anton 3039: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3040: @section @code{POSTPONE}
1.66 anton 3041: @cindex postpone tutorial
1.21 crook 3042:
1.48 anton 3043: You can compile the compilation semantics (instead of compiling the
3044: interpretation semantics) of a word with @code{POSTPONE}:
1.21 crook 3045:
1.48 anton 3046: @example
3047: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
1.51 pazsan 3048: POSTPONE + ; immediate
1.48 anton 3049: : foo ( n1 n2 -- n )
3050: MY-+ ;
3051: 1 2 foo .
3052: see foo
3053: @end example
1.21 crook 3054:
1.48 anton 3055: During the definition of @code{foo} the text interpreter performs the
3056: compilation semantics of @code{MY-+}, which performs the compilation
3057: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3058:
3059: This example also displays separate stack comments for the compilation
3060: semantics and for the stack effect of the compiled code. For words with
3061: default compilation semantics these stack effects are usually not
3062: displayed; the stack effect of the compilation semantics is always
3063: @code{( -- )} for these words, the stack effect for the compiled code is
3064: the stack effect of the interpretation semantics.
3065:
3066: Note that the state of the interpreter does not come into play when
3067: performing the compilation semantics in this way. You can also perform
3068: it interpretively, e.g.:
3069:
3070: @example
3071: : foo2 ( n1 n2 -- n )
3072: [ MY-+ ] ;
3073: 1 2 foo .
3074: see foo
3075: @end example
1.21 crook 3076:
1.48 anton 3077: However, there are some broken Forth systems where this does not always
1.62 crook 3078: work, and therefore this practice was been declared non-standard in
1.48 anton 3079: 1999.
3080: @c !! repair.fs
3081:
3082: Here is another example for using @code{POSTPONE}:
1.44 crook 3083:
1.48 anton 3084: @example
3085: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3086: POSTPONE negate POSTPONE + ; immediate compile-only
3087: : bar ( n1 n2 -- n )
3088: MY-- ;
3089: 2 1 bar .
3090: see bar
3091: @end example
1.21 crook 3092:
1.48 anton 3093: You can define @code{ENDIF} in this way:
1.21 crook 3094:
1.48 anton 3095: @example
3096: : ENDIF ( Compilation: orig -- )
3097: POSTPONE then ; immediate
3098: @end example
1.21 crook 3099:
1.141 anton 3100: @quotation Assignment
1.48 anton 3101: Write @code{MY-2DUP} that has compilation semantics equivalent to
3102: @code{2dup}, but compiles @code{over over}.
1.141 anton 3103: @end quotation
1.29 crook 3104:
1.66 anton 3105: @c !! @xref{Macros} for reference
3106:
3107:
1.48 anton 3108: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3109: @section @code{Literal}
1.66 anton 3110: @cindex literal tutorial
1.29 crook 3111:
1.48 anton 3112: You cannot @code{POSTPONE} numbers:
1.21 crook 3113:
1.48 anton 3114: @example
3115: : [FOO] POSTPONE 500 ; immediate
1.21 crook 3116: @end example
3117:
1.48 anton 3118: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
1.29 crook 3119:
1.48 anton 3120: @example
3121: : [FOO] ( compilation: --; run-time: -- n )
3122: 500 POSTPONE literal ; immediate
1.29 crook 3123:
1.60 anton 3124: : flip [FOO] ;
1.48 anton 3125: flip .
3126: see flip
3127: @end example
1.29 crook 3128:
1.48 anton 3129: @code{LITERAL} consumes a number at compile-time (when it's compilation
3130: semantics are executed) and pushes it at run-time (when the code it
3131: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3132: number computed at compile time into the current word:
1.29 crook 3133:
1.48 anton 3134: @example
3135: : bar ( -- n )
3136: [ 2 2 + ] literal ;
3137: see bar
3138: @end example
1.29 crook 3139:
1.141 anton 3140: @quotation Assignment
1.48 anton 3141: Write @code{]L} which allows writing the example above as @code{: bar (
3142: -- n ) [ 2 2 + ]L ;}
1.141 anton 3143: @end quotation
1.48 anton 3144:
1.66 anton 3145: @c !! @xref{Macros} for reference
3146:
1.48 anton 3147:
3148: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3149: @section Advanced macros
1.66 anton 3150: @cindex macros, advanced tutorial
3151: @cindex run-time code generation, tutorial
1.48 anton 3152:
1.66 anton 3153: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
3154: Execution Tokens}. It frequently performs @code{execute}, a relatively
3155: expensive operation in some Forth implementations. You can use
1.48 anton 3156: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3157: and produce a word that contains the word to be performed directly:
3158:
3159: @c use ]] ... [[
3160: @example
3161: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3162: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3163: \ array beginning at addr and containing u elements
3164: @{ xt @}
3165: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
1.50 anton 3166: POSTPONE i POSTPONE @@ xt compile,
1.48 anton 3167: 1 cells POSTPONE literal POSTPONE +loop ;
3168:
3169: : sum-array ( addr u -- n )
3170: 0 rot rot [ ' + compile-map-array ] ;
3171: see sum-array
3172: a 5 sum-array .
3173: @end example
3174:
3175: You can use the full power of Forth for generating the code; here's an
3176: example where the code is generated in a loop:
3177:
3178: @example
3179: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3180: \ n2=n1+(addr1)*n, addr2=addr1+cell
1.50 anton 3181: POSTPONE tuck POSTPONE @@
1.48 anton 3182: POSTPONE literal POSTPONE * POSTPONE +
3183: POSTPONE swap POSTPONE cell+ ;
3184:
3185: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
1.51 pazsan 3186: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
1.48 anton 3187: 0 postpone literal postpone swap
3188: [ ' compile-vmul-step compile-map-array ]
3189: postpone drop ;
3190: see compile-vmul
3191:
3192: : a-vmul ( addr -- n )
1.51 pazsan 3193: \ n=a*v, where v is a vector that's as long as a and starts at addr
1.48 anton 3194: [ a 5 compile-vmul ] ;
3195: see a-vmul
3196: a a-vmul .
3197: @end example
3198:
3199: This example uses @code{compile-map-array} to show off, but you could
1.66 anton 3200: also use @code{map-array} instead (try it now!).
1.48 anton 3201:
3202: You can use this technique for efficient multiplication of large
3203: matrices. In matrix multiplication, you multiply every line of one
3204: matrix with every column of the other matrix. You can generate the code
3205: for one line once, and use it for every column. The only downside of
3206: this technique is that it is cumbersome to recover the memory consumed
3207: by the generated code when you are done (and in more complicated cases
3208: it is not possible portably).
3209:
1.66 anton 3210: @c !! @xref{Macros} for reference
3211:
3212:
1.48 anton 3213: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3214: @section Compilation Tokens
1.66 anton 3215: @cindex compilation tokens, tutorial
3216: @cindex CT, tutorial
1.48 anton 3217:
3218: This section is Gforth-specific. You can skip it.
3219:
3220: @code{' word compile,} compiles the interpretation semantics. For words
3221: with default compilation semantics this is the same as performing the
3222: compilation semantics. To represent the compilation semantics of other
3223: words (e.g., words like @code{if} that have no interpretation
3224: semantics), Gforth has the concept of a compilation token (CT,
3225: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3226: You can perform the compilation semantics represented by a CT with
3227: @code{execute}:
1.29 crook 3228:
1.48 anton 3229: @example
3230: : foo2 ( n1 n2 -- n )
3231: [ comp' + execute ] ;
3232: see foo
3233: @end example
1.29 crook 3234:
1.48 anton 3235: You can compile the compilation semantics represented by a CT with
3236: @code{postpone,}:
1.30 anton 3237:
1.48 anton 3238: @example
3239: : foo3 ( -- )
3240: [ comp' + postpone, ] ;
3241: see foo3
3242: @end example
1.30 anton 3243:
1.51 pazsan 3244: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
1.48 anton 3245: @code{comp'} is particularly useful for words that have no
3246: interpretation semantics:
1.29 crook 3247:
1.30 anton 3248: @example
1.48 anton 3249: ' if
1.60 anton 3250: comp' if .s 2drop
1.30 anton 3251: @end example
3252:
1.66 anton 3253: Reference: @ref{Tokens for Words}.
3254:
1.29 crook 3255:
1.48 anton 3256: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3257: @section Wordlists and Search Order
1.66 anton 3258: @cindex wordlists tutorial
3259: @cindex search order, tutorial
1.48 anton 3260:
3261: The dictionary is not just a memory area that allows you to allocate
3262: memory with @code{allot}, it also contains the Forth words, arranged in
3263: several wordlists. When searching for a word in a wordlist,
3264: conceptually you start searching at the youngest and proceed towards
3265: older words (in reality most systems nowadays use hash-tables); i.e., if
3266: you define a word with the same name as an older word, the new word
3267: shadows the older word.
3268:
3269: Which wordlists are searched in which order is determined by the search
3270: order. You can display the search order with @code{order}. It displays
3271: first the search order, starting with the wordlist searched first, then
3272: it displays the wordlist that will contain newly defined words.
1.21 crook 3273:
1.48 anton 3274: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
1.21 crook 3275:
1.48 anton 3276: @example
3277: wordlist constant mywords
3278: @end example
1.21 crook 3279:
1.48 anton 3280: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3281: defined words (the @emph{current} wordlist):
1.21 crook 3282:
1.48 anton 3283: @example
3284: mywords set-current
3285: order
3286: @end example
1.26 crook 3287:
1.48 anton 3288: Gforth does not display a name for the wordlist in @code{mywords}
3289: because this wordlist was created anonymously with @code{wordlist}.
1.21 crook 3290:
1.48 anton 3291: You can get the current wordlist with @code{get-current ( -- wid)}. If
3292: you want to put something into a specific wordlist without overall
3293: effect on the current wordlist, this typically looks like this:
1.21 crook 3294:
1.48 anton 3295: @example
3296: get-current mywords set-current ( wid )
3297: create someword
3298: ( wid ) set-current
3299: @end example
1.21 crook 3300:
1.48 anton 3301: You can write the search order with @code{set-order ( wid1 .. widn n --
3302: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3303: searched wordlist is topmost.
1.21 crook 3304:
1.48 anton 3305: @example
3306: get-order mywords swap 1+ set-order
3307: order
3308: @end example
1.21 crook 3309:
1.48 anton 3310: Yes, the order of wordlists in the output of @code{order} is reversed
3311: from stack comments and the output of @code{.s} and thus unintuitive.
1.21 crook 3312:
1.141 anton 3313: @quotation Assignment
1.48 anton 3314: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3315: wordlist to the search order. Define @code{previous ( -- )}, which
3316: removes the first searched wordlist from the search order. Experiment
3317: with boundary conditions (you will see some crashes or situations that
3318: are hard or impossible to leave).
1.141 anton 3319: @end quotation
1.21 crook 3320:
1.48 anton 3321: The search order is a powerful foundation for providing features similar
3322: to Modula-2 modules and C++ namespaces. However, trying to modularize
3323: programs in this way has disadvantages for debugging and reuse/factoring
3324: that overcome the advantages in my experience (I don't do huge projects,
1.55 anton 3325: though). These disadvantages are not so clear in other
1.82 anton 3326: languages/programming environments, because these languages are not so
1.48 anton 3327: strong in debugging and reuse.
1.21 crook 3328:
1.66 anton 3329: @c !! example
3330:
3331: Reference: @ref{Word Lists}.
1.21 crook 3332:
1.29 crook 3333: @c ******************************************************************
1.48 anton 3334: @node Introduction, Words, Tutorial, Top
1.29 crook 3335: @comment node-name, next, previous, up
3336: @chapter An Introduction to ANS Forth
3337: @cindex Forth - an introduction
1.21 crook 3338:
1.83 anton 3339: The difference of this chapter from the Tutorial (@pxref{Tutorial}) is
3340: that it is slower-paced in its examples, but uses them to dive deep into
3341: explaining Forth internals (not covered by the Tutorial). Apart from
3342: that, this chapter covers far less material. It is suitable for reading
3343: without using a computer.
3344:
1.29 crook 3345: The primary purpose of this manual is to document Gforth. However, since
3346: Forth is not a widely-known language and there is a lack of up-to-date
3347: teaching material, it seems worthwhile to provide some introductory
1.49 anton 3348: material. For other sources of Forth-related
3349: information, see @ref{Forth-related information}.
1.21 crook 3350:
1.29 crook 3351: The examples in this section should work on any ANS Forth; the
3352: output shown was produced using Gforth. Each example attempts to
3353: reproduce the exact output that Gforth produces. If you try out the
3354: examples (and you should), what you should type is shown @kbd{like this}
3355: and Gforth's response is shown @code{like this}. The single exception is
1.30 anton 3356: that, where the example shows @key{RET} it means that you should
1.29 crook 3357: press the ``carriage return'' key. Unfortunately, some output formats for
3358: this manual cannot show the difference between @kbd{this} and
3359: @code{this} which will make trying out the examples harder (but not
3360: impossible).
1.21 crook 3361:
1.29 crook 3362: Forth is an unusual language. It provides an interactive development
3363: environment which includes both an interpreter and compiler. Forth
3364: programming style encourages you to break a problem down into many
3365: @cindex factoring
3366: small fragments (@dfn{factoring}), and then to develop and test each
3367: fragment interactively. Forth advocates assert that breaking the
3368: edit-compile-test cycle used by conventional programming languages can
3369: lead to great productivity improvements.
1.21 crook 3370:
1.29 crook 3371: @menu
1.67 anton 3372: * Introducing the Text Interpreter::
3373: * Stacks and Postfix notation::
3374: * Your first definition::
3375: * How does that work?::
3376: * Forth is written in Forth::
3377: * Review - elements of a Forth system::
3378: * Where to go next::
3379: * Exercises::
1.29 crook 3380: @end menu
1.21 crook 3381:
1.29 crook 3382: @comment ----------------------------------------------
3383: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3384: @section Introducing the Text Interpreter
3385: @cindex text interpreter
3386: @cindex outer interpreter
1.21 crook 3387:
1.30 anton 3388: @c IMO this is too detailed and the pace is too slow for
3389: @c an introduction. If you know German, take a look at
3390: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3391: @c to see how I do it - anton
3392:
1.44 crook 3393: @c nac-> Where I have accepted your comments 100% and modified the text
3394: @c accordingly, I have deleted your comments. Elsewhere I have added a
3395: @c response like this to attempt to rationalise what I have done. Of
3396: @c course, this is a very clumsy mechanism for something that would be
3397: @c done far more efficiently over a beer. Please delete any dialogue
3398: @c you consider closed.
3399:
1.29 crook 3400: When you invoke the Forth image, you will see a startup banner printed
3401: and nothing else (if you have Gforth installed on your system, try
1.30 anton 3402: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
1.29 crook 3403: its command line interpreter, which is called the @dfn{Text Interpreter}
3404: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
1.49 anton 3405: about the text interpreter as you read through this chapter, for more
3406: detail @pxref{The Text Interpreter}).
1.21 crook 3407:
1.29 crook 3408: Although it's not obvious, Forth is actually waiting for your
1.30 anton 3409: input. Type a number and press the @key{RET} key:
1.21 crook 3410:
1.26 crook 3411: @example
1.30 anton 3412: @kbd{45@key{RET}} ok
1.26 crook 3413: @end example
1.21 crook 3414:
1.29 crook 3415: Rather than give you a prompt to invite you to input something, the text
3416: interpreter prints a status message @i{after} it has processed a line
3417: of input. The status message in this case (``@code{ ok}'' followed by
3418: carriage-return) indicates that the text interpreter was able to process
3419: all of your input successfully. Now type something illegal:
3420:
3421: @example
1.30 anton 3422: @kbd{qwer341@key{RET}}
1.134 anton 3423: *the terminal*:2: Undefined word
3424: >>>qwer341<<<
3425: Backtrace:
3426: $2A95B42A20 throw
3427: $2A95B57FB8 no.extensions
1.29 crook 3428: @end example
1.23 crook 3429:
1.134 anton 3430: The exact text, other than the ``Undefined word'' may differ slightly
3431: on your system, but the effect is the same; when the text interpreter
1.29 crook 3432: detects an error, it discards any remaining text on a line, resets
1.134 anton 3433: certain internal state and prints an error message. For a detailed
3434: description of error messages see @ref{Error messages}.
1.23 crook 3435:
1.29 crook 3436: The text interpreter waits for you to press carriage-return, and then
3437: processes your input line. Starting at the beginning of the line, it
3438: breaks the line into groups of characters separated by spaces. For each
3439: group of characters in turn, it makes two attempts to do something:
1.23 crook 3440:
1.29 crook 3441: @itemize @bullet
3442: @item
1.44 crook 3443: @cindex name dictionary
1.29 crook 3444: It tries to treat it as a command. It does this by searching a @dfn{name
3445: dictionary}. If the group of characters matches an entry in the name
3446: dictionary, the name dictionary provides the text interpreter with
3447: information that allows the text interpreter perform some actions. In
3448: Forth jargon, we say that the group
3449: @cindex word
3450: @cindex definition
3451: @cindex execution token
3452: @cindex xt
3453: of characters names a @dfn{word}, that the dictionary search returns an
3454: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3455: word, and that the text interpreter executes the xt. Often, the terms
3456: @dfn{word} and @dfn{definition} are used interchangeably.
3457: @item
3458: If the text interpreter fails to find a match in the name dictionary, it
3459: tries to treat the group of characters as a number in the current number
3460: base (when you start up Forth, the current number base is base 10). If
3461: the group of characters legitimately represents a number, the text
3462: interpreter pushes the number onto a stack (we'll learn more about that
3463: in the next section).
3464: @end itemize
1.23 crook 3465:
1.29 crook 3466: If the text interpreter is unable to do either of these things with any
3467: group of characters, it discards the group of characters and the rest of
3468: the line, then prints an error message. If the text interpreter reaches
3469: the end of the line without error, it prints the status message ``@code{ ok}''
3470: followed by carriage-return.
1.21 crook 3471:
1.29 crook 3472: This is the simplest command we can give to the text interpreter:
1.23 crook 3473:
3474: @example
1.30 anton 3475: @key{RET} ok
1.23 crook 3476: @end example
1.21 crook 3477:
1.29 crook 3478: The text interpreter did everything we asked it to do (nothing) without
3479: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3480: command:
1.21 crook 3481:
1.23 crook 3482: @example
1.30 anton 3483: @kbd{12 dup fred dup@key{RET}}
1.134 anton 3484: *the terminal*:3: Undefined word
3485: 12 dup >>>fred<<< dup
3486: Backtrace:
3487: $2A95B42A20 throw
3488: $2A95B57FB8 no.extensions
1.23 crook 3489: @end example
1.21 crook 3490:
1.29 crook 3491: When you press the carriage-return key, the text interpreter starts to
3492: work its way along the line:
1.21 crook 3493:
1.29 crook 3494: @itemize @bullet
3495: @item
3496: When it gets to the space after the @code{2}, it takes the group of
3497: characters @code{12} and looks them up in the name
3498: dictionary@footnote{We can't tell if it found them or not, but assume
3499: for now that it did not}. There is no match for this group of characters
3500: in the name dictionary, so it tries to treat them as a number. It is
3501: able to do this successfully, so it puts the number, 12, ``on the stack''
3502: (whatever that means).
3503: @item
3504: The text interpreter resumes scanning the line and gets the next group
3505: of characters, @code{dup}. It looks it up in the name dictionary and
3506: (you'll have to take my word for this) finds it, and executes the word
3507: @code{dup} (whatever that means).
3508: @item
3509: Once again, the text interpreter resumes scanning the line and gets the
3510: group of characters @code{fred}. It looks them up in the name
3511: dictionary, but can't find them. It tries to treat them as a number, but
3512: they don't represent any legal number.
3513: @end itemize
1.21 crook 3514:
1.29 crook 3515: At this point, the text interpreter gives up and prints an error
3516: message. The error message shows exactly how far the text interpreter
3517: got in processing the line. In particular, it shows that the text
3518: interpreter made no attempt to do anything with the final character
3519: group, @code{dup}, even though we have good reason to believe that the
3520: text interpreter would have no problem looking that word up and
3521: executing it a second time.
1.21 crook 3522:
3523:
1.29 crook 3524: @comment ----------------------------------------------
3525: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3526: @section Stacks, postfix notation and parameter passing
3527: @cindex text interpreter
3528: @cindex outer interpreter
1.21 crook 3529:
1.29 crook 3530: In procedural programming languages (like C and Pascal), the
3531: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3532: functions or procedures are called with @dfn{explicit parameters}. For
3533: example, in C we might write:
1.21 crook 3534:
1.23 crook 3535: @example
1.29 crook 3536: total = total + new_volume(length,height,depth);
1.23 crook 3537: @end example
1.21 crook 3538:
1.23 crook 3539: @noindent
1.29 crook 3540: where new_volume is a function-call to another piece of code, and total,
3541: length, height and depth are all variables. length, height and depth are
3542: parameters to the function-call.
1.21 crook 3543:
1.29 crook 3544: In Forth, the equivalent of the function or procedure is the
3545: @dfn{definition} and parameters are implicitly passed between
3546: definitions using a shared stack that is visible to the
3547: programmer. Although Forth does support variables, the existence of the
3548: stack means that they are used far less often than in most other
3549: programming languages. When the text interpreter encounters a number, it
3550: will place (@dfn{push}) it on the stack. There are several stacks (the
1.30 anton 3551: actual number is implementation-dependent ...) and the particular stack
1.29 crook 3552: used for any operation is implied unambiguously by the operation being
3553: performed. The stack used for all integer operations is called the @dfn{data
3554: stack} and, since this is the stack used most commonly, references to
3555: ``the data stack'' are often abbreviated to ``the stack''.
1.21 crook 3556:
1.29 crook 3557: The stacks have a last-in, first-out (LIFO) organisation. If you type:
1.21 crook 3558:
1.23 crook 3559: @example
1.30 anton 3560: @kbd{1 2 3@key{RET}} ok
1.23 crook 3561: @end example
1.21 crook 3562:
1.29 crook 3563: Then this instructs the text interpreter to placed three numbers on the
3564: (data) stack. An analogy for the behaviour of the stack is to take a
3565: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3566: the table. The 3 was the last card onto the pile (``last-in'') and if
3567: you take a card off the pile then, unless you're prepared to fiddle a
3568: bit, the card that you take off will be the 3 (``first-out''). The
3569: number that will be first-out of the stack is called the @dfn{top of
3570: stack}, which
3571: @cindex TOS definition
3572: is often abbreviated to @dfn{TOS}.
1.21 crook 3573:
1.29 crook 3574: To understand how parameters are passed in Forth, consider the
3575: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3576: be surprised to learn that this definition performs addition. More
3577: precisely, it adds two number together and produces a result. Where does
3578: it get the two numbers from? It takes the top two numbers off the
3579: stack. Where does it place the result? On the stack. You can act-out the
3580: behaviour of @code{+} with your playing cards like this:
1.21 crook 3581:
3582: @itemize @bullet
3583: @item
1.29 crook 3584: Pick up two cards from the stack on the table
1.21 crook 3585: @item
1.29 crook 3586: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3587: numbers''
1.21 crook 3588: @item
1.29 crook 3589: Decide that the answer is 5
1.21 crook 3590: @item
1.29 crook 3591: Shuffle the two cards back into the pack and find a 5
1.21 crook 3592: @item
1.29 crook 3593: Put a 5 on the remaining ace that's on the table.
1.21 crook 3594: @end itemize
3595:
1.29 crook 3596: If you don't have a pack of cards handy but you do have Forth running,
3597: you can use the definition @code{.s} to show the current state of the stack,
3598: without affecting the stack. Type:
1.21 crook 3599:
3600: @example
1.124 anton 3601: @kbd{clearstacks 1 2 3@key{RET}} ok
1.30 anton 3602: @kbd{.s@key{RET}} <3> 1 2 3 ok
1.23 crook 3603: @end example
3604:
1.124 anton 3605: The text interpreter looks up the word @code{clearstacks} and executes
3606: it; it tidies up the stacks and removes any entries that may have been
1.29 crook 3607: left on it by earlier examples. The text interpreter pushes each of the
3608: three numbers in turn onto the stack. Finally, the text interpreter
3609: looks up the word @code{.s} and executes it. The effect of executing
3610: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3611: followed by a list of all the items on the stack; the item on the far
3612: right-hand side is the TOS.
1.21 crook 3613:
1.29 crook 3614: You can now type:
1.21 crook 3615:
1.29 crook 3616: @example
1.30 anton 3617: @kbd{+ .s@key{RET}} <2> 1 5 ok
1.29 crook 3618: @end example
1.21 crook 3619:
1.29 crook 3620: @noindent
3621: which is correct; there are now 2 items on the stack and the result of
3622: the addition is 5.
1.23 crook 3623:
1.29 crook 3624: If you're playing with cards, try doing a second addition: pick up the
3625: two cards, work out that their sum is 6, shuffle them into the pack,
3626: look for a 6 and place that on the table. You now have just one item on
3627: the stack. What happens if you try to do a third addition? Pick up the
3628: first card, pick up the second card -- ah! There is no second card. This
3629: is called a @dfn{stack underflow} and consitutes an error. If you try to
1.95 anton 3630: do the same thing with Forth it often reports an error (probably a Stack
1.29 crook 3631: Underflow or an Invalid Memory Address error).
1.23 crook 3632:
1.29 crook 3633: The opposite situation to a stack underflow is a @dfn{stack overflow},
3634: which simply accepts that there is a finite amount of storage space
3635: reserved for the stack. To stretch the playing card analogy, if you had
3636: enough packs of cards and you piled the cards up on the table, you would
3637: eventually be unable to add another card; you'd hit the ceiling. Gforth
3638: allows you to set the maximum size of the stacks. In general, the only
3639: time that you will get a stack overflow is because a definition has a
3640: bug in it and is generating data on the stack uncontrollably.
1.23 crook 3641:
1.29 crook 3642: There's one final use for the playing card analogy. If you model your
3643: stack using a pack of playing cards, the maximum number of items on
3644: your stack will be 52 (I assume you didn't use the Joker). The maximum
3645: @i{value} of any item on the stack is 13 (the King). In fact, the only
3646: possible numbers are positive integer numbers 1 through 13; you can't
3647: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3648: think about some of the cards, you can accommodate different
3649: numbers. For example, you could think of the Jack as representing 0,
3650: the Queen as representing -1 and the King as representing -2. Your
1.45 crook 3651: @i{range} remains unchanged (you can still only represent a total of 13
1.29 crook 3652: numbers) but the numbers that you can represent are -2 through 10.
1.28 crook 3653:
1.29 crook 3654: In that analogy, the limit was the amount of information that a single
3655: stack entry could hold, and Forth has a similar limit. In Forth, the
3656: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3657: implementation dependent and affects the maximum value that a stack
3658: entry can hold. A Standard Forth provides a cell size of at least
3659: 16-bits, and most desktop systems use a cell size of 32-bits.
1.21 crook 3660:
1.29 crook 3661: Forth does not do any type checking for you, so you are free to
3662: manipulate and combine stack items in any way you wish. A convenient way
3663: of treating stack items is as 2's complement signed integers, and that
3664: is what Standard words like @code{+} do. Therefore you can type:
1.21 crook 3665:
1.29 crook 3666: @example
1.30 anton 3667: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
1.29 crook 3668: @end example
1.21 crook 3669:
1.29 crook 3670: If you use numbers and definitions like @code{+} in order to turn Forth
3671: into a great big pocket calculator, you will realise that it's rather
3672: different from a normal calculator. Rather than typing 2 + 3 = you had
3673: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3674: result). The terminology used to describe this difference is to say that
3675: your calculator uses @dfn{Infix Notation} (parameters and operators are
3676: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3677: operators are separate), also called @dfn{Reverse Polish Notation}.
1.21 crook 3678:
1.29 crook 3679: Whilst postfix notation might look confusing to begin with, it has
3680: several important advantages:
1.21 crook 3681:
1.23 crook 3682: @itemize @bullet
3683: @item
1.29 crook 3684: it is unambiguous
1.23 crook 3685: @item
1.29 crook 3686: it is more concise
1.23 crook 3687: @item
1.29 crook 3688: it fits naturally with a stack-based system
1.23 crook 3689: @end itemize
1.21 crook 3690:
1.29 crook 3691: To examine these claims in more detail, consider these sums:
1.21 crook 3692:
1.29 crook 3693: @example
3694: 6 + 5 * 4 =
3695: 4 * 5 + 6 =
3696: @end example
1.21 crook 3697:
1.29 crook 3698: If you're just learning maths or your maths is very rusty, you will
3699: probably come up with the answer 44 for the first and 26 for the
3700: second. If you are a bit of a whizz at maths you will remember the
3701: @i{convention} that multiplication takes precendence over addition, and
3702: you'd come up with the answer 26 both times. To explain the answer 26
3703: to someone who got the answer 44, you'd probably rewrite the first sum
3704: like this:
1.21 crook 3705:
1.29 crook 3706: @example
3707: 6 + (5 * 4) =
3708: @end example
1.21 crook 3709:
1.29 crook 3710: If what you really wanted was to perform the addition before the
3711: multiplication, you would have to use parentheses to force it.
1.21 crook 3712:
1.29 crook 3713: If you did the first two sums on a pocket calculator you would probably
3714: get the right answers, unless you were very cautious and entered them using
3715: these keystroke sequences:
1.21 crook 3716:
1.29 crook 3717: 6 + 5 = * 4 =
3718: 4 * 5 = + 6 =
1.21 crook 3719:
1.29 crook 3720: Postfix notation is unambiguous because the order that the operators
3721: are applied is always explicit; that also means that parentheses are
3722: never required. The operators are @i{active} (the act of quoting the
3723: operator makes the operation occur) which removes the need for ``=''.
1.28 crook 3724:
1.29 crook 3725: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3726: equivalent ways:
1.26 crook 3727:
3728: @example
1.29 crook 3729: 6 5 4 * + or:
3730: 5 4 * 6 +
1.26 crook 3731: @end example
1.23 crook 3732:
1.29 crook 3733: An important thing that you should notice about this notation is that
3734: the @i{order} of the numbers does not change; if you want to subtract
3735: 2 from 10 you type @code{10 2 -}.
1.1 anton 3736:
1.29 crook 3737: The reason that Forth uses postfix notation is very simple to explain: it
3738: makes the implementation extremely simple, and it follows naturally from
3739: using the stack as a mechanism for passing parameters. Another way of
3740: thinking about this is to realise that all Forth definitions are
3741: @i{active}; they execute as they are encountered by the text
3742: interpreter. The result of this is that the syntax of Forth is trivially
3743: simple.
1.1 anton 3744:
3745:
3746:
1.29 crook 3747: @comment ----------------------------------------------
3748: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3749: @section Your first Forth definition
3750: @cindex first definition
1.1 anton 3751:
1.29 crook 3752: Until now, the examples we've seen have been trivial; we've just been
3753: using Forth as a bigger-than-pocket calculator. Also, each calculation
3754: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3755: again@footnote{That's not quite true. If you press the up-arrow key on
3756: your keyboard you should be able to scroll back to any earlier command,
3757: edit it and re-enter it.} In this section we'll see how to add new
3758: words to Forth's vocabulary.
1.1 anton 3759:
1.29 crook 3760: The easiest way to create a new word is to use a @dfn{colon
3761: definition}. We'll define a few and try them out before worrying too
3762: much about how they work. Try typing in these examples; be careful to
3763: copy the spaces accurately:
1.1 anton 3764:
1.29 crook 3765: @example
3766: : add-two 2 + . ;
3767: : greet ." Hello and welcome" ;
3768: : demo 5 add-two ;
3769: @end example
1.1 anton 3770:
1.29 crook 3771: @noindent
3772: Now try them out:
1.1 anton 3773:
1.29 crook 3774: @example
1.30 anton 3775: @kbd{greet@key{RET}} Hello and welcome ok
3776: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
3777: @kbd{4 add-two@key{RET}} 6 ok
3778: @kbd{demo@key{RET}} 7 ok
3779: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
1.29 crook 3780: @end example
1.1 anton 3781:
1.29 crook 3782: The first new thing that we've introduced here is the pair of words
3783: @code{:} and @code{;}. These are used to start and terminate a new
3784: definition, respectively. The first word after the @code{:} is the name
3785: for the new definition.
1.1 anton 3786:
1.29 crook 3787: As you can see from the examples, a definition is built up of words that
3788: have already been defined; Forth makes no distinction between
3789: definitions that existed when you started the system up, and those that
3790: you define yourself.
1.1 anton 3791:
1.29 crook 3792: The examples also introduce the words @code{.} (dot), @code{."}
3793: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
3794: the stack and displays it. It's like @code{.s} except that it only
3795: displays the top item of the stack and it is destructive; after it has
3796: executed, the number is no longer on the stack. There is always one
3797: space printed after the number, and no spaces before it. Dot-quote
3798: defines a string (a sequence of characters) that will be printed when
3799: the word is executed. The string can contain any printable characters
3800: except @code{"}. A @code{"} has a special function; it is not a Forth
3801: word but it acts as a delimiter (the way that delimiters work is
3802: described in the next section). Finally, @code{dup} duplicates the value
3803: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
1.1 anton 3804:
1.29 crook 3805: We already know that the text interpreter searches through the
3806: dictionary to locate names. If you've followed the examples earlier, you
3807: will already have a definition called @code{add-two}. Lets try modifying
3808: it by typing in a new definition:
1.1 anton 3809:
1.29 crook 3810: @example
1.30 anton 3811: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
1.29 crook 3812: @end example
1.5 anton 3813:
1.29 crook 3814: Forth recognised that we were defining a word that already exists, and
3815: printed a message to warn us of that fact. Let's try out the new
3816: definition:
1.5 anton 3817:
1.29 crook 3818: @example
1.30 anton 3819: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
1.29 crook 3820: @end example
1.1 anton 3821:
1.29 crook 3822: @noindent
3823: All that we've actually done here, though, is to create a new
3824: definition, with a particular name. The fact that there was already a
3825: definition with the same name did not make any difference to the way
3826: that the new definition was created (except that Forth printed a warning
3827: message). The old definition of add-two still exists (try @code{demo}
3828: again to see that this is true). Any new definition will use the new
3829: definition of @code{add-two}, but old definitions continue to use the
3830: version that already existed at the time that they were @code{compiled}.
1.1 anton 3831:
1.29 crook 3832: Before you go on to the next section, try defining and redefining some
3833: words of your own.
1.1 anton 3834:
1.29 crook 3835: @comment ----------------------------------------------
3836: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
3837: @section How does that work?
3838: @cindex parsing words
1.1 anton 3839:
1.30 anton 3840: @c That's pretty deep (IMO way too deep) for an introduction. - anton
3841:
3842: @c Is it a good idea to talk about the interpretation semantics of a
3843: @c number? We don't have an xt to go along with it. - anton
3844:
3845: @c Now that I have eliminated execution semantics, I wonder if it would not
3846: @c be better to keep them (or add run-time semantics), to make it easier to
3847: @c explain what compilation semantics usually does. - anton
3848:
1.44 crook 3849: @c nac-> I removed the term ``default compilation sematics'' from the
3850: @c introductory chapter. Removing ``execution semantics'' was making
3851: @c everything simpler to explain, then I think the use of this term made
3852: @c everything more complex again. I replaced it with ``default
3853: @c semantics'' (which is used elsewhere in the manual) by which I mean
3854: @c ``a definition that has neither the immediate nor the compile-only
1.83 anton 3855: @c flag set''.
3856:
3857: @c anton: I have eliminated default semantics (except in one place where it
3858: @c means "default interpretation and compilation semantics"), because it
3859: @c makes no sense in the presence of combined words. I reverted to
3860: @c "execution semantics" where necessary.
3861:
3862: @c nac-> I reworded big chunks of the ``how does that work''
1.44 crook 3863: @c section (and, unusually for me, I think I even made it shorter!). See
3864: @c what you think -- I know I have not addressed your primary concern
3865: @c that it is too heavy-going for an introduction. From what I understood
3866: @c of your course notes it looks as though they might be a good framework.
3867: @c Things that I've tried to capture here are some things that came as a
3868: @c great revelation here when I first understood them. Also, I like the
3869: @c fact that a very simple code example shows up almost all of the issues
3870: @c that you need to understand to see how Forth works. That's unique and
3871: @c worthwhile to emphasise.
3872:
1.83 anton 3873: @c anton: I think it's a good idea to present the details, especially those
3874: @c that you found to be a revelation, and probably the tutorial tries to be
3875: @c too superficial and does not get some of the things across that make
3876: @c Forth special. I do believe that most of the time these things should
3877: @c be discussed at the end of a section or in separate sections instead of
3878: @c in the middle of a section (e.g., the stuff you added in "User-defined
3879: @c defining words" leads in a completely different direction from the rest
3880: @c of the section).
3881:
1.29 crook 3882: Now we're going to take another look at the definition of @code{add-two}
3883: from the previous section. From our knowledge of the way that the text
3884: interpreter works, we would have expected this result when we tried to
3885: define @code{add-two}:
1.21 crook 3886:
1.29 crook 3887: @example
1.44 crook 3888: @kbd{: add-two 2 + . ;@key{RET}}
1.134 anton 3889: *the terminal*:4: Undefined word
3890: : >>>add-two<<< 2 + . ;
1.29 crook 3891: @end example
1.28 crook 3892:
1.29 crook 3893: The reason that this didn't happen is bound up in the way that @code{:}
3894: works. The word @code{:} does two special things. The first special
3895: thing that it does prevents the text interpreter from ever seeing the
3896: characters @code{add-two}. The text interpreter uses a variable called
3897: @cindex modifying >IN
1.44 crook 3898: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
1.29 crook 3899: input line. When it encounters the word @code{:} it behaves in exactly
3900: the same way as it does for any other word; it looks it up in the name
3901: dictionary, finds its xt and executes it. When @code{:} executes, it
3902: looks at the input buffer, finds the word @code{add-two} and advances the
3903: value of @code{>IN} to point past it. It then does some other stuff
3904: associated with creating the new definition (including creating an entry
3905: for @code{add-two} in the name dictionary). When the execution of @code{:}
3906: completes, control returns to the text interpreter, which is oblivious
3907: to the fact that it has been tricked into ignoring part of the input
3908: line.
1.21 crook 3909:
1.29 crook 3910: @cindex parsing words
3911: Words like @code{:} -- words that advance the value of @code{>IN} and so
3912: prevent the text interpreter from acting on the whole of the input line
3913: -- are called @dfn{parsing words}.
1.21 crook 3914:
1.29 crook 3915: @cindex @code{state} - effect on the text interpreter
3916: @cindex text interpreter - effect of state
3917: The second special thing that @code{:} does is change the value of a
3918: variable called @code{state}, which affects the way that the text
3919: interpreter behaves. When Gforth starts up, @code{state} has the value
3920: 0, and the text interpreter is said to be @dfn{interpreting}. During a
3921: colon definition (started with @code{:}), @code{state} is set to -1 and
1.44 crook 3922: the text interpreter is said to be @dfn{compiling}.
3923:
3924: In this example, the text interpreter is compiling when it processes the
3925: string ``@code{2 + . ;}''. It still breaks the string down into
3926: character sequences in the same way. However, instead of pushing the
3927: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
3928: into the definition of @code{add-two} that will make the number @code{2} get
3929: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
3930: the behaviours of @code{+} and @code{.} are also compiled into the
3931: definition.
3932:
3933: One category of words don't get compiled. These so-called @dfn{immediate
3934: words} get executed (performed @i{now}) regardless of whether the text
3935: interpreter is interpreting or compiling. The word @code{;} is an
3936: immediate word. Rather than being compiled into the definition, it
3937: executes. Its effect is to terminate the current definition, which
3938: includes changing the value of @code{state} back to 0.
3939:
3940: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
3941: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
3942: definition.
1.28 crook 3943:
1.30 anton 3944: In Forth, every word or number can be described in terms of two
1.29 crook 3945: properties:
1.28 crook 3946:
3947: @itemize @bullet
3948: @item
1.29 crook 3949: @cindex interpretation semantics
1.44 crook 3950: Its @dfn{interpretation semantics} describe how it will behave when the
3951: text interpreter encounters it in @dfn{interpret} state. The
3952: interpretation semantics of a word are represented by an @dfn{execution
3953: token}.
1.28 crook 3954: @item
1.29 crook 3955: @cindex compilation semantics
1.44 crook 3956: Its @dfn{compilation semantics} describe how it will behave when the
3957: text interpreter encounters it in @dfn{compile} state. The compilation
3958: semantics of a word are represented in an implementation-dependent way;
3959: Gforth uses a @dfn{compilation token}.
1.29 crook 3960: @end itemize
3961:
3962: @noindent
3963: Numbers are always treated in a fixed way:
3964:
3965: @itemize @bullet
1.28 crook 3966: @item
1.44 crook 3967: When the number is @dfn{interpreted}, its behaviour is to push the
3968: number onto the stack.
1.28 crook 3969: @item
1.30 anton 3970: When the number is @dfn{compiled}, a piece of code is appended to the
3971: current definition that pushes the number when it runs. (In other words,
3972: the compilation semantics of a number are to postpone its interpretation
3973: semantics until the run-time of the definition that it is being compiled
3974: into.)
1.29 crook 3975: @end itemize
3976:
1.44 crook 3977: Words don't behave in such a regular way, but most have @i{default
3978: semantics} which means that they behave like this:
1.29 crook 3979:
3980: @itemize @bullet
1.28 crook 3981: @item
1.30 anton 3982: The @dfn{interpretation semantics} of the word are to do something useful.
3983: @item
1.29 crook 3984: The @dfn{compilation semantics} of the word are to append its
1.30 anton 3985: @dfn{interpretation semantics} to the current definition (so that its
3986: run-time behaviour is to do something useful).
1.28 crook 3987: @end itemize
3988:
1.30 anton 3989: @cindex immediate words
1.44 crook 3990: The actual behaviour of any particular word can be controlled by using
3991: the words @code{immediate} and @code{compile-only} when the word is
3992: defined. These words set flags in the name dictionary entry of the most
3993: recently defined word, and these flags are retrieved by the text
3994: interpreter when it finds the word in the name dictionary.
3995:
3996: A word that is marked as @dfn{immediate} has compilation semantics that
3997: are identical to its interpretation semantics. In other words, it
3998: behaves like this:
1.29 crook 3999:
4000: @itemize @bullet
4001: @item
1.30 anton 4002: The @dfn{interpretation semantics} of the word are to do something useful.
1.29 crook 4003: @item
1.30 anton 4004: The @dfn{compilation semantics} of the word are to do something useful
4005: (and actually the same thing); i.e., it is executed during compilation.
1.29 crook 4006: @end itemize
1.28 crook 4007:
1.44 crook 4008: Marking a word as @dfn{compile-only} prohibits the text interpreter from
4009: performing the interpretation semantics of the word directly; an attempt
4010: to do so will generate an error. It is never necessary to use
4011: @code{compile-only} (and it is not even part of ANS Forth, though it is
4012: provided by many implementations) but it is good etiquette to apply it
4013: to a word that will not behave correctly (and might have unexpected
4014: side-effects) in interpret state. For example, it is only legal to use
4015: the conditional word @code{IF} within a definition. If you forget this
4016: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4017: @code{compile-only} allows the text interpreter to generate a helpful
4018: error message rather than subjecting you to the consequences of your
4019: folly.
4020:
1.29 crook 4021: This example shows the difference between an immediate and a
4022: non-immediate word:
1.28 crook 4023:
1.29 crook 4024: @example
4025: : show-state state @@ . ;
4026: : show-state-now show-state ; immediate
4027: : word1 show-state ;
4028: : word2 show-state-now ;
1.28 crook 4029: @end example
1.23 crook 4030:
1.29 crook 4031: The word @code{immediate} after the definition of @code{show-state-now}
4032: makes that word an immediate word. These definitions introduce a new
4033: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4034: variable, and leaves it on the stack. Therefore, the behaviour of
4035: @code{show-state} is to print a number that represents the current value
4036: of @code{state}.
1.28 crook 4037:
1.29 crook 4038: When you execute @code{word1}, it prints the number 0, indicating that
4039: the system is interpreting. When the text interpreter compiled the
4040: definition of @code{word1}, it encountered @code{show-state} whose
1.30 anton 4041: compilation semantics are to append its interpretation semantics to the
1.29 crook 4042: current definition. When you execute @code{word1}, it performs the
1.30 anton 4043: interpretation semantics of @code{show-state}. At the time that @code{word1}
1.29 crook 4044: (and therefore @code{show-state}) are executed, the system is
4045: interpreting.
1.28 crook 4046:
1.30 anton 4047: When you pressed @key{RET} after entering the definition of @code{word2},
1.29 crook 4048: you should have seen the number -1 printed, followed by ``@code{
4049: ok}''. When the text interpreter compiled the definition of
4050: @code{word2}, it encountered @code{show-state-now}, an immediate word,
1.30 anton 4051: whose compilation semantics are therefore to perform its interpretation
1.29 crook 4052: semantics. It is executed straight away (even before the text
4053: interpreter has moved on to process another group of characters; the
4054: @code{;} in this example). The effect of executing it are to display the
4055: value of @code{state} @i{at the time that the definition of}
4056: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4057: system is compiling at this time. If you execute @code{word2} it does
4058: nothing at all.
1.28 crook 4059:
1.29 crook 4060: @cindex @code{."}, how it works
4061: Before leaving the subject of immediate words, consider the behaviour of
4062: @code{."} in the definition of @code{greet}, in the previous
4063: section. This word is both a parsing word and an immediate word. Notice
4064: that there is a space between @code{."} and the start of the text
4065: @code{Hello and welcome}, but that there is no space between the last
4066: letter of @code{welcome} and the @code{"} character. The reason for this
4067: is that @code{."} is a Forth word; it must have a space after it so that
4068: the text interpreter can identify it. The @code{"} is not a Forth word;
4069: it is a @dfn{delimiter}. The examples earlier show that, when the string
4070: is displayed, there is neither a space before the @code{H} nor after the
4071: @code{e}. Since @code{."} is an immediate word, it executes at the time
4072: that @code{greet} is defined. When it executes, its behaviour is to
4073: search forward in the input line looking for the delimiter. When it
4074: finds the delimiter, it updates @code{>IN} to point past the
4075: delimiter. It also compiles some magic code into the definition of
4076: @code{greet}; the xt of a run-time routine that prints a text string. It
4077: compiles the string @code{Hello and welcome} into memory so that it is
4078: available to be printed later. When the text interpreter gains control,
4079: the next word it finds in the input stream is @code{;} and so it
4080: terminates the definition of @code{greet}.
1.28 crook 4081:
4082:
4083: @comment ----------------------------------------------
1.29 crook 4084: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4085: @section Forth is written in Forth
4086: @cindex structure of Forth programs
4087:
4088: When you start up a Forth compiler, a large number of definitions
4089: already exist. In Forth, you develop a new application using bottom-up
4090: programming techniques to create new definitions that are defined in
4091: terms of existing definitions. As you create each definition you can
4092: test and debug it interactively.
4093:
4094: If you have tried out the examples in this section, you will probably
4095: have typed them in by hand; when you leave Gforth, your definitions will
4096: be lost. You can avoid this by using a text editor to enter Forth source
4097: code into a file, and then loading code from the file using
1.49 anton 4098: @code{include} (@pxref{Forth source files}). A Forth source file is
1.29 crook 4099: processed by the text interpreter, just as though you had typed it in by
4100: hand@footnote{Actually, there are some subtle differences -- see
4101: @ref{The Text Interpreter}.}.
4102:
4103: Gforth also supports the traditional Forth alternative to using text
1.49 anton 4104: files for program entry (@pxref{Blocks}).
1.28 crook 4105:
1.29 crook 4106: In common with many, if not most, Forth compilers, most of Gforth is
4107: actually written in Forth. All of the @file{.fs} files in the
4108: installation directory@footnote{For example,
1.30 anton 4109: @file{/usr/local/share/gforth...}} are Forth source files, which you can
1.29 crook 4110: study to see examples of Forth programming.
1.28 crook 4111:
1.29 crook 4112: Gforth maintains a history file that records every line that you type to
4113: the text interpreter. This file is preserved between sessions, and is
4114: used to provide a command-line recall facility. If you enter long
4115: definitions by hand, you can use a text editor to paste them out of the
4116: history file into a Forth source file for reuse at a later time
1.49 anton 4117: (for more information @pxref{Command-line editing}).
1.28 crook 4118:
4119:
4120: @comment ----------------------------------------------
1.29 crook 4121: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4122: @section Review - elements of a Forth system
4123: @cindex elements of a Forth system
1.28 crook 4124:
1.29 crook 4125: To summarise this chapter:
1.28 crook 4126:
4127: @itemize @bullet
4128: @item
1.29 crook 4129: Forth programs use @dfn{factoring} to break a problem down into small
4130: fragments called @dfn{words} or @dfn{definitions}.
4131: @item
4132: Forth program development is an interactive process.
4133: @item
4134: The main command loop that accepts input, and controls both
4135: interpretation and compilation, is called the @dfn{text interpreter}
4136: (also known as the @dfn{outer interpreter}).
4137: @item
4138: Forth has a very simple syntax, consisting of words and numbers
4139: separated by spaces or carriage-return characters. Any additional syntax
4140: is imposed by @dfn{parsing words}.
4141: @item
4142: Forth uses a stack to pass parameters between words. As a result, it
4143: uses postfix notation.
4144: @item
4145: To use a word that has previously been defined, the text interpreter
4146: searches for the word in the @dfn{name dictionary}.
4147: @item
1.30 anton 4148: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
1.28 crook 4149: @item
1.29 crook 4150: The text interpreter uses the value of @code{state} to select between
4151: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4152: semantics} of a word that it encounters.
1.28 crook 4153: @item
1.30 anton 4154: The relationship between the @dfn{interpretation semantics} and
4155: @dfn{compilation semantics} for a word
1.29 crook 4156: depend upon the way in which the word was defined (for example, whether
4157: it is an @dfn{immediate} word).
1.28 crook 4158: @item
1.29 crook 4159: Forth definitions can be implemented in Forth (called @dfn{high-level
4160: definitions}) or in some other way (usually a lower-level language and
4161: as a result often called @dfn{low-level definitions}, @dfn{code
4162: definitions} or @dfn{primitives}).
1.28 crook 4163: @item
1.29 crook 4164: Many Forth systems are implemented mainly in Forth.
1.28 crook 4165: @end itemize
4166:
4167:
1.29 crook 4168: @comment ----------------------------------------------
1.48 anton 4169: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
1.29 crook 4170: @section Where To Go Next
4171: @cindex where to go next
1.28 crook 4172:
1.29 crook 4173: Amazing as it may seem, if you have read (and understood) this far, you
4174: know almost all the fundamentals about the inner workings of a Forth
4175: system. You certainly know enough to be able to read and understand the
4176: rest of this manual and the ANS Forth document, to learn more about the
4177: facilities that Forth in general and Gforth in particular provide. Even
4178: scarier, you know almost enough to implement your own Forth system.
1.30 anton 4179: However, that's not a good idea just yet... better to try writing some
1.29 crook 4180: programs in Gforth.
1.28 crook 4181:
1.29 crook 4182: Forth has such a rich vocabulary that it can be hard to know where to
4183: start in learning it. This section suggests a few sets of words that are
4184: enough to write small but useful programs. Use the word index in this
4185: document to learn more about each word, then try it out and try to write
4186: small definitions using it. Start by experimenting with these words:
1.28 crook 4187:
4188: @itemize @bullet
4189: @item
1.29 crook 4190: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4191: @item
4192: Comparison: @code{MIN MAX =}
4193: @item
4194: Logic: @code{AND OR XOR NOT}
4195: @item
4196: Stack manipulation: @code{DUP DROP SWAP OVER}
1.28 crook 4197: @item
1.29 crook 4198: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
1.28 crook 4199: @item
1.29 crook 4200: Input/Output: @code{. ." EMIT CR KEY}
1.28 crook 4201: @item
1.29 crook 4202: Defining words: @code{: ; CREATE}
1.28 crook 4203: @item
1.29 crook 4204: Memory allocation words: @code{ALLOT ,}
1.28 crook 4205: @item
1.29 crook 4206: Tools: @code{SEE WORDS .S MARKER}
4207: @end itemize
4208:
4209: When you have mastered those, go on to:
4210:
4211: @itemize @bullet
1.28 crook 4212: @item
1.29 crook 4213: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
1.28 crook 4214: @item
1.29 crook 4215: Memory access: @code{@@ !}
1.28 crook 4216: @end itemize
1.23 crook 4217:
1.29 crook 4218: When you have mastered these, there's nothing for it but to read through
4219: the whole of this manual and find out what you've missed.
4220:
4221: @comment ----------------------------------------------
1.48 anton 4222: @node Exercises, , Where to go next, Introduction
1.29 crook 4223: @section Exercises
4224: @cindex exercises
4225:
4226: TODO: provide a set of programming excercises linked into the stuff done
4227: already and into other sections of the manual. Provide solutions to all
4228: the exercises in a .fs file in the distribution.
4229:
4230: @c Get some inspiration from Starting Forth and Kelly&Spies.
4231:
4232: @c excercises:
4233: @c 1. take inches and convert to feet and inches.
4234: @c 2. take temperature and convert from fahrenheight to celcius;
4235: @c may need to care about symmetric vs floored??
4236: @c 3. take input line and do character substitution
4237: @c to encipher or decipher
4238: @c 4. as above but work on a file for in and out
4239: @c 5. take input line and convert to pig-latin
4240: @c
4241: @c thing of sets of things to exercise then come up with
4242: @c problems that need those things.
4243:
4244:
1.26 crook 4245: @c ******************************************************************
1.29 crook 4246: @node Words, Error messages, Introduction, Top
1.1 anton 4247: @chapter Forth Words
1.26 crook 4248: @cindex words
1.1 anton 4249:
4250: @menu
4251: * Notation::
1.65 anton 4252: * Case insensitivity::
4253: * Comments::
4254: * Boolean Flags::
1.1 anton 4255: * Arithmetic::
4256: * Stack Manipulation::
1.5 anton 4257: * Memory::
1.1 anton 4258: * Control Structures::
4259: * Defining Words::
1.65 anton 4260: * Interpretation and Compilation Semantics::
1.47 crook 4261: * Tokens for Words::
1.81 anton 4262: * Compiling words::
1.65 anton 4263: * The Text Interpreter::
1.111 anton 4264: * The Input Stream::
1.65 anton 4265: * Word Lists::
4266: * Environmental Queries::
1.12 anton 4267: * Files::
4268: * Blocks::
4269: * Other I/O::
1.121 anton 4270: * OS command line arguments::
1.78 anton 4271: * Locals::
4272: * Structures::
4273: * Object-oriented Forth::
1.12 anton 4274: * Programming Tools::
1.150 anton 4275: * C Interface::
1.12 anton 4276: * Assembler and Code Words::
4277: * Threading Words::
1.65 anton 4278: * Passing Commands to the OS::
4279: * Keeping track of Time::
4280: * Miscellaneous Words::
1.1 anton 4281: @end menu
4282:
1.65 anton 4283: @node Notation, Case insensitivity, Words, Words
1.1 anton 4284: @section Notation
4285: @cindex notation of glossary entries
4286: @cindex format of glossary entries
4287: @cindex glossary notation format
4288: @cindex word glossary entry format
4289:
4290: The Forth words are described in this section in the glossary notation
1.67 anton 4291: that has become a de-facto standard for Forth texts:
1.1 anton 4292:
4293: @format
1.29 crook 4294: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
1.1 anton 4295: @end format
1.29 crook 4296: @i{Description}
1.1 anton 4297:
4298: @table @var
4299: @item word
1.28 crook 4300: The name of the word.
1.1 anton 4301:
4302: @item Stack effect
4303: @cindex stack effect
1.29 crook 4304: The stack effect is written in the notation @code{@i{before} --
4305: @i{after}}, where @i{before} and @i{after} describe the top of
1.1 anton 4306: stack entries before and after the execution of the word. The rest of
4307: the stack is not touched by the word. The top of stack is rightmost,
4308: i.e., a stack sequence is written as it is typed in. Note that Gforth
4309: uses a separate floating point stack, but a unified stack
1.29 crook 4310: notation. Also, return stack effects are not shown in @i{stack
4311: effect}, but in @i{Description}. The name of a stack item describes
1.1 anton 4312: the type and/or the function of the item. See below for a discussion of
4313: the types.
4314:
4315: All words have two stack effects: A compile-time stack effect and a
4316: run-time stack effect. The compile-time stack-effect of most words is
1.29 crook 4317: @i{ -- }. If the compile-time stack-effect of a word deviates from
1.1 anton 4318: this standard behaviour, or the word does other unusual things at
4319: compile time, both stack effects are shown; otherwise only the run-time
4320: stack effect is shown.
4321:
4322: @cindex pronounciation of words
4323: @item pronunciation
4324: How the word is pronounced.
4325:
4326: @cindex wordset
1.67 anton 4327: @cindex environment wordset
1.1 anton 4328: @item wordset
1.21 crook 4329: The ANS Forth standard is divided into several word sets. A standard
4330: system need not support all of them. Therefore, in theory, the fewer
4331: word sets your program uses the more portable it will be. However, we
4332: suspect that most ANS Forth systems on personal machines will feature
1.26 crook 4333: all word sets. Words that are not defined in ANS Forth have
1.21 crook 4334: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
1.1 anton 4335: describes words that will work in future releases of Gforth;
4336: @code{gforth-internal} words are more volatile. Environmental query
4337: strings are also displayed like words; you can recognize them by the
1.21 crook 4338: @code{environment} in the word set field.
1.1 anton 4339:
4340: @item Description
4341: A description of the behaviour of the word.
4342: @end table
4343:
4344: @cindex types of stack items
4345: @cindex stack item types
4346: The type of a stack item is specified by the character(s) the name
4347: starts with:
4348:
4349: @table @code
4350: @item f
4351: @cindex @code{f}, stack item type
4352: Boolean flags, i.e. @code{false} or @code{true}.
4353: @item c
4354: @cindex @code{c}, stack item type
4355: Char
4356: @item w
4357: @cindex @code{w}, stack item type
4358: Cell, can contain an integer or an address
4359: @item n
4360: @cindex @code{n}, stack item type
4361: signed integer
4362: @item u
4363: @cindex @code{u}, stack item type
4364: unsigned integer
4365: @item d
4366: @cindex @code{d}, stack item type
4367: double sized signed integer
4368: @item ud
4369: @cindex @code{ud}, stack item type
4370: double sized unsigned integer
4371: @item r
4372: @cindex @code{r}, stack item type
4373: Float (on the FP stack)
1.21 crook 4374: @item a-
1.1 anton 4375: @cindex @code{a_}, stack item type
4376: Cell-aligned address
1.21 crook 4377: @item c-
1.1 anton 4378: @cindex @code{c_}, stack item type
4379: Char-aligned address (note that a Char may have two bytes in Windows NT)
1.21 crook 4380: @item f-
1.1 anton 4381: @cindex @code{f_}, stack item type
4382: Float-aligned address
1.21 crook 4383: @item df-
1.1 anton 4384: @cindex @code{df_}, stack item type
4385: Address aligned for IEEE double precision float
1.21 crook 4386: @item sf-
1.1 anton 4387: @cindex @code{sf_}, stack item type
4388: Address aligned for IEEE single precision float
4389: @item xt
4390: @cindex @code{xt}, stack item type
4391: Execution token, same size as Cell
4392: @item wid
4393: @cindex @code{wid}, stack item type
1.21 crook 4394: Word list ID, same size as Cell
1.68 anton 4395: @item ior, wior
4396: @cindex ior type description
4397: @cindex wior type description
4398: I/O result code, cell-sized. In Gforth, you can @code{throw} iors.
1.1 anton 4399: @item f83name
4400: @cindex @code{f83name}, stack item type
4401: Pointer to a name structure
4402: @item "
4403: @cindex @code{"}, stack item type
1.12 anton 4404: string in the input stream (not on the stack). The terminating character
4405: is a blank by default. If it is not a blank, it is shown in @code{<>}
1.1 anton 4406: quotes.
4407: @end table
4408:
1.65 anton 4409: @comment ----------------------------------------------
4410: @node Case insensitivity, Comments, Notation, Words
4411: @section Case insensitivity
4412: @cindex case sensitivity
4413: @cindex upper and lower case
4414:
4415: Gforth is case-insensitive; you can enter definitions and invoke
4416: Standard words using upper, lower or mixed case (however,
4417: @pxref{core-idef, Implementation-defined options, Implementation-defined
4418: options}).
4419:
4420: ANS Forth only @i{requires} implementations to recognise Standard words
4421: when they are typed entirely in upper case. Therefore, a Standard
4422: program must use upper case for all Standard words. You can use whatever
4423: case you like for words that you define, but in a Standard program you
4424: have to use the words in the same case that you defined them.
4425:
4426: Gforth supports case sensitivity through @code{table}s (case-sensitive
4427: wordlists, @pxref{Word Lists}).
4428:
4429: Two people have asked how to convert Gforth to be case-sensitive; while
4430: we think this is a bad idea, you can change all wordlists into tables
4431: like this:
4432:
4433: @example
4434: ' table-find forth-wordlist wordlist-map @ !
4435: @end example
4436:
4437: Note that you now have to type the predefined words in the same case
4438: that we defined them, which are varying. You may want to convert them
4439: to your favourite case before doing this operation (I won't explain how,
4440: because if you are even contemplating doing this, you'd better have
4441: enough knowledge of Forth systems to know this already).
4442:
4443: @node Comments, Boolean Flags, Case insensitivity, Words
1.21 crook 4444: @section Comments
1.26 crook 4445: @cindex comments
1.21 crook 4446:
1.29 crook 4447: Forth supports two styles of comment; the traditional @i{in-line} comment,
4448: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
1.21 crook 4449:
1.44 crook 4450:
1.23 crook 4451: doc-(
1.21 crook 4452: doc-\
1.23 crook 4453: doc-\G
1.21 crook 4454:
1.44 crook 4455:
1.21 crook 4456: @node Boolean Flags, Arithmetic, Comments, Words
4457: @section Boolean Flags
1.26 crook 4458: @cindex Boolean flags
1.21 crook 4459:
4460: A Boolean flag is cell-sized. A cell with all bits clear represents the
4461: flag @code{false} and a flag with all bits set represents the flag
1.26 crook 4462: @code{true}. Words that check a flag (for example, @code{IF}) will treat
1.29 crook 4463: a cell that has @i{any} bit set as @code{true}.
1.67 anton 4464: @c on and off to Memory?
4465: @c true and false to "Bitwise operations" or "Numeric comparison"?
1.44 crook 4466:
1.21 crook 4467: doc-true
4468: doc-false
1.29 crook 4469: doc-on
4470: doc-off
1.21 crook 4471:
1.44 crook 4472:
1.21 crook 4473: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
1.1 anton 4474: @section Arithmetic
4475: @cindex arithmetic words
4476:
4477: @cindex division with potentially negative operands
4478: Forth arithmetic is not checked, i.e., you will not hear about integer
4479: overflow on addition or multiplication, you may hear about division by
4480: zero if you are lucky. The operator is written after the operands, but
4481: the operands are still in the original order. I.e., the infix @code{2-1}
4482: corresponds to @code{2 1 -}. Forth offers a variety of division
4483: operators. If you perform division with potentially negative operands,
4484: you do not want to use @code{/} or @code{/mod} with its undefined
4485: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4486: former, @pxref{Mixed precision}).
1.26 crook 4487: @comment TODO discuss the different division forms and the std approach
1.1 anton 4488:
4489: @menu
4490: * Single precision::
1.67 anton 4491: * Double precision:: Double-cell integer arithmetic
1.1 anton 4492: * Bitwise operations::
1.67 anton 4493: * Numeric comparison::
1.29 crook 4494: * Mixed precision:: Operations with single and double-cell integers
1.1 anton 4495: * Floating Point::
4496: @end menu
4497:
1.67 anton 4498: @node Single precision, Double precision, Arithmetic, Arithmetic
1.1 anton 4499: @subsection Single precision
4500: @cindex single precision arithmetic words
4501:
1.67 anton 4502: @c !! cell undefined
4503:
4504: By default, numbers in Forth are single-precision integers that are one
1.26 crook 4505: cell in size. They can be signed or unsigned, depending upon how you
1.49 anton 4506: treat them. For the rules used by the text interpreter for recognising
4507: single-precision integers see @ref{Number Conversion}.
1.21 crook 4508:
1.67 anton 4509: These words are all defined for signed operands, but some of them also
4510: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
4511: @code{*}.
1.44 crook 4512:
1.1 anton 4513: doc-+
1.21 crook 4514: doc-1+
1.128 anton 4515: doc-under+
1.1 anton 4516: doc--
1.21 crook 4517: doc-1-
1.1 anton 4518: doc-*
4519: doc-/
4520: doc-mod
4521: doc-/mod
4522: doc-negate
4523: doc-abs
4524: doc-min
4525: doc-max
1.27 crook 4526: doc-floored
1.1 anton 4527:
1.44 crook 4528:
1.67 anton 4529: @node Double precision, Bitwise operations, Single precision, Arithmetic
1.21 crook 4530: @subsection Double precision
4531: @cindex double precision arithmetic words
4532:
1.49 anton 4533: For the rules used by the text interpreter for
4534: recognising double-precision integers, see @ref{Number Conversion}.
1.21 crook 4535:
4536: A double precision number is represented by a cell pair, with the most
1.67 anton 4537: significant cell at the TOS. It is trivial to convert an unsigned single
4538: to a double: simply push a @code{0} onto the TOS. Since numbers are
4539: represented by Gforth using 2's complement arithmetic, converting a
4540: signed single to a (signed) double requires sign-extension across the
4541: most significant cell. This can be achieved using @code{s>d}. The moral
4542: of the story is that you cannot convert a number without knowing whether
4543: it represents an unsigned or a signed number.
1.21 crook 4544:
1.67 anton 4545: These words are all defined for signed operands, but some of them also
4546: work for unsigned numbers: @code{d+}, @code{d-}.
1.44 crook 4547:
1.21 crook 4548: doc-s>d
1.67 anton 4549: doc-d>s
1.21 crook 4550: doc-d+
4551: doc-d-
4552: doc-dnegate
4553: doc-dabs
4554: doc-dmin
4555: doc-dmax
4556:
1.44 crook 4557:
1.67 anton 4558: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
4559: @subsection Bitwise operations
4560: @cindex bitwise operation words
4561:
4562:
4563: doc-and
4564: doc-or
4565: doc-xor
4566: doc-invert
4567: doc-lshift
4568: doc-rshift
4569: doc-2*
4570: doc-d2*
4571: doc-2/
4572: doc-d2/
4573:
4574:
4575: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
1.21 crook 4576: @subsection Numeric comparison
4577: @cindex numeric comparison words
4578:
1.67 anton 4579: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
4580: d0= d0<>}) work for for both signed and unsigned numbers.
1.44 crook 4581:
1.28 crook 4582: doc-<
4583: doc-<=
4584: doc-<>
4585: doc-=
4586: doc->
4587: doc->=
4588:
1.21 crook 4589: doc-0<
1.23 crook 4590: doc-0<=
1.21 crook 4591: doc-0<>
4592: doc-0=
1.23 crook 4593: doc-0>
4594: doc-0>=
1.28 crook 4595:
4596: doc-u<
4597: doc-u<=
1.44 crook 4598: @c u<> and u= exist but are the same as <> and =
1.31 anton 4599: @c doc-u<>
4600: @c doc-u=
1.28 crook 4601: doc-u>
4602: doc-u>=
4603:
4604: doc-within
4605:
4606: doc-d<
4607: doc-d<=
4608: doc-d<>
4609: doc-d=
4610: doc-d>
4611: doc-d>=
1.23 crook 4612:
1.21 crook 4613: doc-d0<
1.23 crook 4614: doc-d0<=
4615: doc-d0<>
1.21 crook 4616: doc-d0=
1.23 crook 4617: doc-d0>
4618: doc-d0>=
4619:
1.21 crook 4620: doc-du<
1.28 crook 4621: doc-du<=
1.44 crook 4622: @c du<> and du= exist but are the same as d<> and d=
1.31 anton 4623: @c doc-du<>
4624: @c doc-du=
1.28 crook 4625: doc-du>
4626: doc-du>=
1.1 anton 4627:
1.44 crook 4628:
1.21 crook 4629: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
1.1 anton 4630: @subsection Mixed precision
4631: @cindex mixed precision arithmetic words
4632:
1.44 crook 4633:
1.1 anton 4634: doc-m+
4635: doc-*/
4636: doc-*/mod
4637: doc-m*
4638: doc-um*
4639: doc-m*/
4640: doc-um/mod
4641: doc-fm/mod
4642: doc-sm/rem
4643:
1.44 crook 4644:
1.21 crook 4645: @node Floating Point, , Mixed precision, Arithmetic
1.1 anton 4646: @subsection Floating Point
4647: @cindex floating point arithmetic words
4648:
1.49 anton 4649: For the rules used by the text interpreter for
4650: recognising floating-point numbers see @ref{Number Conversion}.
1.1 anton 4651:
1.67 anton 4652: Gforth has a separate floating point stack, but the documentation uses
4653: the unified notation.@footnote{It's easy to generate the separate
4654: notation from that by just separating the floating-point numbers out:
4655: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
4656: r3 )}.}
1.1 anton 4657:
4658: @cindex floating-point arithmetic, pitfalls
4659: Floating point numbers have a number of unpleasant surprises for the
4660: unwary (e.g., floating point addition is not associative) and even a few
4661: for the wary. You should not use them unless you know what you are doing
4662: or you don't care that the results you get are totally bogus. If you
4663: want to learn about the problems of floating point numbers (and how to
1.66 anton 4664: avoid them), you might start with @cite{David Goldberg,
4665: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
4666: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
4667: Surveys 23(1):5@minus{}48, March 1991}.
1.1 anton 4668:
1.44 crook 4669:
1.21 crook 4670: doc-d>f
4671: doc-f>d
1.1 anton 4672: doc-f+
4673: doc-f-
4674: doc-f*
4675: doc-f/
4676: doc-fnegate
4677: doc-fabs
4678: doc-fmax
4679: doc-fmin
4680: doc-floor
4681: doc-fround
4682: doc-f**
4683: doc-fsqrt
4684: doc-fexp
4685: doc-fexpm1
4686: doc-fln
4687: doc-flnp1
4688: doc-flog
4689: doc-falog
1.32 anton 4690: doc-f2*
4691: doc-f2/
4692: doc-1/f
4693: doc-precision
4694: doc-set-precision
4695:
4696: @cindex angles in trigonometric operations
4697: @cindex trigonometric operations
4698: Angles in floating point operations are given in radians (a full circle
4699: has 2 pi radians).
4700:
1.1 anton 4701: doc-fsin
4702: doc-fcos
4703: doc-fsincos
4704: doc-ftan
4705: doc-fasin
4706: doc-facos
4707: doc-fatan
4708: doc-fatan2
4709: doc-fsinh
4710: doc-fcosh
4711: doc-ftanh
4712: doc-fasinh
4713: doc-facosh
4714: doc-fatanh
1.21 crook 4715: doc-pi
1.28 crook 4716:
1.32 anton 4717: @cindex equality of floats
4718: @cindex floating-point comparisons
1.31 anton 4719: One particular problem with floating-point arithmetic is that comparison
4720: for equality often fails when you would expect it to succeed. For this
4721: reason approximate equality is often preferred (but you still have to
1.67 anton 4722: know what you are doing). Also note that IEEE NaNs may compare
1.68 anton 4723: differently from what you might expect. The comparison words are:
1.31 anton 4724:
4725: doc-f~rel
4726: doc-f~abs
1.68 anton 4727: doc-f~
1.31 anton 4728: doc-f=
4729: doc-f<>
4730:
4731: doc-f<
4732: doc-f<=
4733: doc-f>
4734: doc-f>=
4735:
1.21 crook 4736: doc-f0<
1.28 crook 4737: doc-f0<=
4738: doc-f0<>
1.21 crook 4739: doc-f0=
1.28 crook 4740: doc-f0>
4741: doc-f0>=
4742:
1.1 anton 4743:
4744: @node Stack Manipulation, Memory, Arithmetic, Words
4745: @section Stack Manipulation
4746: @cindex stack manipulation words
4747:
4748: @cindex floating-point stack in the standard
1.21 crook 4749: Gforth maintains a number of separate stacks:
4750:
1.29 crook 4751: @cindex data stack
4752: @cindex parameter stack
1.21 crook 4753: @itemize @bullet
4754: @item
1.29 crook 4755: A data stack (also known as the @dfn{parameter stack}) -- for
4756: characters, cells, addresses, and double cells.
1.21 crook 4757:
1.29 crook 4758: @cindex floating-point stack
1.21 crook 4759: @item
1.44 crook 4760: A floating point stack -- for holding floating point (FP) numbers.
1.21 crook 4761:
1.29 crook 4762: @cindex return stack
1.21 crook 4763: @item
1.44 crook 4764: A return stack -- for holding the return addresses of colon
1.32 anton 4765: definitions and other (non-FP) data.
1.21 crook 4766:
1.29 crook 4767: @cindex locals stack
1.21 crook 4768: @item
1.44 crook 4769: A locals stack -- for holding local variables.
1.21 crook 4770: @end itemize
4771:
1.1 anton 4772: @menu
4773: * Data stack::
4774: * Floating point stack::
4775: * Return stack::
4776: * Locals stack::
4777: * Stack pointer manipulation::
4778: @end menu
4779:
4780: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
4781: @subsection Data stack
4782: @cindex data stack manipulation words
4783: @cindex stack manipulations words, data stack
4784:
1.44 crook 4785:
1.1 anton 4786: doc-drop
4787: doc-nip
4788: doc-dup
4789: doc-over
4790: doc-tuck
4791: doc-swap
1.21 crook 4792: doc-pick
1.1 anton 4793: doc-rot
4794: doc--rot
4795: doc-?dup
4796: doc-roll
4797: doc-2drop
4798: doc-2nip
4799: doc-2dup
4800: doc-2over
4801: doc-2tuck
4802: doc-2swap
4803: doc-2rot
4804:
1.44 crook 4805:
1.1 anton 4806: @node Floating point stack, Return stack, Data stack, Stack Manipulation
4807: @subsection Floating point stack
4808: @cindex floating-point stack manipulation words
4809: @cindex stack manipulation words, floating-point stack
4810:
1.32 anton 4811: Whilst every sane Forth has a separate floating-point stack, it is not
4812: strictly required; an ANS Forth system could theoretically keep
4813: floating-point numbers on the data stack. As an additional difficulty,
4814: you don't know how many cells a floating-point number takes. It is
4815: reportedly possible to write words in a way that they work also for a
4816: unified stack model, but we do not recommend trying it. Instead, just
4817: say that your program has an environmental dependency on a separate
4818: floating-point stack.
4819:
4820: doc-floating-stack
4821:
1.1 anton 4822: doc-fdrop
4823: doc-fnip
4824: doc-fdup
4825: doc-fover
4826: doc-ftuck
4827: doc-fswap
1.21 crook 4828: doc-fpick
1.1 anton 4829: doc-frot
4830:
1.44 crook 4831:
1.1 anton 4832: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
4833: @subsection Return stack
4834: @cindex return stack manipulation words
4835: @cindex stack manipulation words, return stack
4836:
1.32 anton 4837: @cindex return stack and locals
4838: @cindex locals and return stack
4839: A Forth system is allowed to keep local variables on the
4840: return stack. This is reasonable, as local variables usually eliminate
4841: the need to use the return stack explicitly. So, if you want to produce
4842: a standard compliant program and you are using local variables in a
4843: word, forget about return stack manipulations in that word (refer to the
4844: standard document for the exact rules).
4845:
1.1 anton 4846: doc->r
4847: doc-r>
4848: doc-r@
4849: doc-rdrop
4850: doc-2>r
4851: doc-2r>
4852: doc-2r@
4853: doc-2rdrop
4854:
1.44 crook 4855:
1.1 anton 4856: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
4857: @subsection Locals stack
4858:
1.78 anton 4859: Gforth uses an extra locals stack. It is described, along with the
4860: reasons for its existence, in @ref{Locals implementation}.
1.21 crook 4861:
1.1 anton 4862: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
4863: @subsection Stack pointer manipulation
4864: @cindex stack pointer manipulation words
4865:
1.44 crook 4866: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
1.21 crook 4867: doc-sp0
1.1 anton 4868: doc-sp@
4869: doc-sp!
1.21 crook 4870: doc-fp0
1.1 anton 4871: doc-fp@
4872: doc-fp!
1.21 crook 4873: doc-rp0
1.1 anton 4874: doc-rp@
4875: doc-rp!
1.21 crook 4876: doc-lp0
1.1 anton 4877: doc-lp@
4878: doc-lp!
4879:
1.44 crook 4880:
1.1 anton 4881: @node Memory, Control Structures, Stack Manipulation, Words
4882: @section Memory
1.26 crook 4883: @cindex memory words
1.1 anton 4884:
1.32 anton 4885: @menu
4886: * Memory model::
4887: * Dictionary allocation::
4888: * Heap Allocation::
4889: * Memory Access::
4890: * Address arithmetic::
4891: * Memory Blocks::
4892: @end menu
4893:
1.67 anton 4894: In addition to the standard Forth memory allocation words, there is also
4895: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
4896: garbage collector}.
4897:
1.32 anton 4898: @node Memory model, Dictionary allocation, Memory, Memory
4899: @subsection ANS Forth and Gforth memory models
4900:
4901: @c The ANS Forth description is a mess (e.g., is the heap part of
4902: @c the dictionary?), so let's not stick to closely with it.
4903:
1.67 anton 4904: ANS Forth considers a Forth system as consisting of several address
4905: spaces, of which only @dfn{data space} is managed and accessible with
4906: the memory words. Memory not necessarily in data space includes the
4907: stacks, the code (called code space) and the headers (called name
4908: space). In Gforth everything is in data space, but the code for the
4909: primitives is usually read-only.
1.32 anton 4910:
4911: Data space is divided into a number of areas: The (data space portion of
4912: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
4913: refer to the search data structure embodied in word lists and headers,
4914: because it is used for looking up names, just as you would in a
4915: conventional dictionary.}, the heap, and a number of system-allocated
4916: buffers.
4917:
1.68 anton 4918: @cindex address arithmetic restrictions, ANS vs. Gforth
4919: @cindex contiguous regions, ANS vs. Gforth
1.32 anton 4920: In ANS Forth data space is also divided into contiguous regions. You
4921: can only use address arithmetic within a contiguous region, not between
4922: them. Usually each allocation gives you one contiguous region, but the
1.33 anton 4923: dictionary allocation words have additional rules (@pxref{Dictionary
1.32 anton 4924: allocation}).
4925:
4926: Gforth provides one big address space, and address arithmetic can be
4927: performed between any addresses. However, in the dictionary headers or
4928: code are interleaved with data, so almost the only contiguous data space
4929: regions there are those described by ANS Forth as contiguous; but you
4930: can be sure that the dictionary is allocated towards increasing
4931: addresses even between contiguous regions. The memory order of
4932: allocations in the heap is platform-dependent (and possibly different
4933: from one run to the next).
4934:
1.27 crook 4935:
1.32 anton 4936: @node Dictionary allocation, Heap Allocation, Memory model, Memory
4937: @subsection Dictionary allocation
1.27 crook 4938: @cindex reserving data space
4939: @cindex data space - reserving some
4940:
1.32 anton 4941: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
4942: you want to deallocate X, you also deallocate everything
4943: allocated after X.
4944:
1.68 anton 4945: @cindex contiguous regions in dictionary allocation
1.32 anton 4946: The allocations using the words below are contiguous and grow the region
4947: towards increasing addresses. Other words that allocate dictionary
4948: memory of any kind (i.e., defining words including @code{:noname}) end
4949: the contiguous region and start a new one.
4950:
4951: In ANS Forth only @code{create}d words are guaranteed to produce an
4952: address that is the start of the following contiguous region. In
4953: particular, the cell allocated by @code{variable} is not guaranteed to
4954: be contiguous with following @code{allot}ed memory.
4955:
4956: You can deallocate memory by using @code{allot} with a negative argument
4957: (with some restrictions, see @code{allot}). For larger deallocations use
4958: @code{marker}.
1.27 crook 4959:
1.29 crook 4960:
1.27 crook 4961: doc-here
4962: doc-unused
4963: doc-allot
4964: doc-c,
1.29 crook 4965: doc-f,
1.27 crook 4966: doc-,
4967: doc-2,
4968:
1.32 anton 4969: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
4970: course you should allocate memory in an aligned way, too. I.e., before
4971: allocating allocating a cell, @code{here} must be cell-aligned, etc.
4972: The words below align @code{here} if it is not already. Basically it is
4973: only already aligned for a type, if the last allocation was a multiple
4974: of the size of this type and if @code{here} was aligned for this type
4975: before.
4976:
4977: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
4978: ANS Forth (@code{maxalign}ed in Gforth).
4979:
4980: doc-align
4981: doc-falign
4982: doc-sfalign
4983: doc-dfalign
4984: doc-maxalign
4985: doc-cfalign
4986:
4987:
4988: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
4989: @subsection Heap allocation
4990: @cindex heap allocation
4991: @cindex dynamic allocation of memory
4992: @cindex memory-allocation word set
4993:
1.68 anton 4994: @cindex contiguous regions and heap allocation
1.32 anton 4995: Heap allocation supports deallocation of allocated memory in any
4996: order. Dictionary allocation is not affected by it (i.e., it does not
4997: end a contiguous region). In Gforth, these words are implemented using
4998: the standard C library calls malloc(), free() and resize().
4999:
1.68 anton 5000: The memory region produced by one invocation of @code{allocate} or
5001: @code{resize} is internally contiguous. There is no contiguity between
5002: such a region and any other region (including others allocated from the
5003: heap).
5004:
1.32 anton 5005: doc-allocate
5006: doc-free
5007: doc-resize
5008:
1.27 crook 5009:
1.32 anton 5010: @node Memory Access, Address arithmetic, Heap Allocation, Memory
1.1 anton 5011: @subsection Memory Access
5012: @cindex memory access words
5013:
5014: doc-@
5015: doc-!
5016: doc-+!
5017: doc-c@
5018: doc-c!
5019: doc-2@
5020: doc-2!
5021: doc-f@
5022: doc-f!
5023: doc-sf@
5024: doc-sf!
5025: doc-df@
5026: doc-df!
1.144 anton 5027: doc-sw@
5028: doc-uw@
5029: doc-w!
5030: doc-sl@
5031: doc-ul@
5032: doc-l!
1.68 anton 5033:
1.32 anton 5034: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5035: @subsection Address arithmetic
1.1 anton 5036: @cindex address arithmetic words
5037:
1.67 anton 5038: Address arithmetic is the foundation on which you can build data
5039: structures like arrays, records (@pxref{Structures}) and objects
5040: (@pxref{Object-oriented Forth}).
1.32 anton 5041:
1.68 anton 5042: @cindex address unit
5043: @cindex au (address unit)
1.1 anton 5044: ANS Forth does not specify the sizes of the data types. Instead, it
5045: offers a number of words for computing sizes and doing address
1.29 crook 5046: arithmetic. Address arithmetic is performed in terms of address units
5047: (aus); on most systems the address unit is one byte. Note that a
5048: character may have more than one au, so @code{chars} is no noop (on
1.68 anton 5049: platforms where it is a noop, it compiles to nothing).
1.1 anton 5050:
1.67 anton 5051: The basic address arithmetic words are @code{+} and @code{-}. E.g., if
5052: you have the address of a cell, perform @code{1 cells +}, and you will
5053: have the address of the next cell.
5054:
1.68 anton 5055: @cindex contiguous regions and address arithmetic
1.67 anton 5056: In ANS Forth you can perform address arithmetic only within a contiguous
5057: region, i.e., if you have an address into one region, you can only add
5058: and subtract such that the result is still within the region; you can
5059: only subtract or compare addresses from within the same contiguous
5060: region. Reasons: several contiguous regions can be arranged in memory
5061: in any way; on segmented systems addresses may have unusual
5062: representations, such that address arithmetic only works within a
5063: region. Gforth provides a few more guarantees (linear address space,
5064: dictionary grows upwards), but in general I have found it easy to stay
5065: within contiguous regions (exception: computing and comparing to the
5066: address just beyond the end of an array).
5067:
1.1 anton 5068: @cindex alignment of addresses for types
5069: ANS Forth also defines words for aligning addresses for specific
5070: types. Many computers require that accesses to specific data types
5071: must only occur at specific addresses; e.g., that cells may only be
5072: accessed at addresses divisible by 4. Even if a machine allows unaligned
5073: accesses, it can usually perform aligned accesses faster.
5074:
5075: For the performance-conscious: alignment operations are usually only
5076: necessary during the definition of a data structure, not during the
5077: (more frequent) accesses to it.
5078:
5079: ANS Forth defines no words for character-aligning addresses. This is not
5080: an oversight, but reflects the fact that addresses that are not
5081: char-aligned have no use in the standard and therefore will not be
5082: created.
5083:
5084: @cindex @code{CREATE} and alignment
1.29 crook 5085: ANS Forth guarantees that addresses returned by @code{CREATE}d words
1.1 anton 5086: are cell-aligned; in addition, Gforth guarantees that these addresses
5087: are aligned for all purposes.
5088:
1.26 crook 5089: Note that the ANS Forth word @code{char} has nothing to do with address
5090: arithmetic.
1.1 anton 5091:
1.44 crook 5092:
1.1 anton 5093: doc-chars
5094: doc-char+
5095: doc-cells
5096: doc-cell+
5097: doc-cell
5098: doc-aligned
5099: doc-floats
5100: doc-float+
5101: doc-float
5102: doc-faligned
5103: doc-sfloats
5104: doc-sfloat+
5105: doc-sfaligned
5106: doc-dfloats
5107: doc-dfloat+
5108: doc-dfaligned
5109: doc-maxaligned
5110: doc-cfaligned
5111: doc-address-unit-bits
1.145 anton 5112: doc-/w
5113: doc-/l
1.44 crook 5114:
1.32 anton 5115: @node Memory Blocks, , Address arithmetic, Memory
1.1 anton 5116: @subsection Memory Blocks
5117: @cindex memory block words
1.27 crook 5118: @cindex character strings - moving and copying
5119:
1.49 anton 5120: Memory blocks often represent character strings; For ways of storing
5121: character strings in memory see @ref{String Formats}. For other
5122: string-processing words see @ref{Displaying characters and strings}.
1.1 anton 5123:
1.67 anton 5124: A few of these words work on address unit blocks. In that case, you
5125: usually have to insert @code{CHARS} before the word when working on
5126: character strings. Most words work on character blocks, and expect a
5127: char-aligned address.
5128:
5129: When copying characters between overlapping memory regions, use
5130: @code{chars move} or choose carefully between @code{cmove} and
5131: @code{cmove>}.
1.44 crook 5132:
1.1 anton 5133: doc-move
5134: doc-erase
5135: doc-cmove
5136: doc-cmove>
5137: doc-fill
5138: doc-blank
1.21 crook 5139: doc-compare
1.111 anton 5140: doc-str=
5141: doc-str<
5142: doc-string-prefix?
1.21 crook 5143: doc-search
1.27 crook 5144: doc--trailing
5145: doc-/string
1.82 anton 5146: doc-bounds
1.141 anton 5147: doc-pad
1.111 anton 5148:
1.27 crook 5149: @comment TODO examples
5150:
1.1 anton 5151:
1.26 crook 5152: @node Control Structures, Defining Words, Memory, Words
1.1 anton 5153: @section Control Structures
5154: @cindex control structures
5155:
1.33 anton 5156: Control structures in Forth cannot be used interpretively, only in a
5157: colon definition@footnote{To be precise, they have no interpretation
5158: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5159: not like this limitation, but have not seen a satisfying way around it
5160: yet, although many schemes have been proposed.
1.1 anton 5161:
5162: @menu
1.33 anton 5163: * Selection:: IF ... ELSE ... ENDIF
5164: * Simple Loops:: BEGIN ...
1.29 crook 5165: * Counted Loops:: DO
1.67 anton 5166: * Arbitrary control structures::
5167: * Calls and returns::
1.1 anton 5168: * Exception Handling::
5169: @end menu
5170:
5171: @node Selection, Simple Loops, Control Structures, Control Structures
5172: @subsection Selection
5173: @cindex selection control structures
5174: @cindex control structures for selection
5175:
5176: @cindex @code{IF} control structure
5177: @example
1.29 crook 5178: @i{flag}
1.1 anton 5179: IF
1.29 crook 5180: @i{code}
1.1 anton 5181: ENDIF
5182: @end example
1.21 crook 5183: @noindent
1.33 anton 5184:
1.44 crook 5185: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5186: with any bit set represents truth) @i{code} is executed.
1.33 anton 5187:
1.1 anton 5188: @example
1.29 crook 5189: @i{flag}
1.1 anton 5190: IF
1.29 crook 5191: @i{code1}
1.1 anton 5192: ELSE
1.29 crook 5193: @i{code2}
1.1 anton 5194: ENDIF
5195: @end example
5196:
1.44 crook 5197: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5198: executed.
1.33 anton 5199:
1.1 anton 5200: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5201: standard, and @code{ENDIF} is not, although it is quite popular. We
5202: recommend using @code{ENDIF}, because it is less confusing for people
5203: who also know other languages (and is not prone to reinforcing negative
5204: prejudices against Forth in these people). Adding @code{ENDIF} to a
5205: system that only supplies @code{THEN} is simple:
5206: @example
1.82 anton 5207: : ENDIF POSTPONE then ; immediate
1.1 anton 5208: @end example
5209:
5210: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5211: (adv.)} has the following meanings:
5212: @quotation
5213: ... 2b: following next after in order ... 3d: as a necessary consequence
5214: (if you were there, then you saw them).
5215: @end quotation
5216: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5217: and many other programming languages has the meaning 3d.]
5218:
1.21 crook 5219: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
1.1 anton 5220: you can avoid using @code{?dup}. Using these alternatives is also more
1.26 crook 5221: efficient than using @code{?dup}. Definitions in ANS Forth
1.1 anton 5222: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5223: @file{compat/control.fs}.
5224:
5225: @cindex @code{CASE} control structure
5226: @example
1.29 crook 5227: @i{n}
1.1 anton 5228: CASE
1.29 crook 5229: @i{n1} OF @i{code1} ENDOF
5230: @i{n2} OF @i{code2} ENDOF
1.1 anton 5231: @dots{}
1.68 anton 5232: ( n ) @i{default-code} ( n )
1.131 anton 5233: ENDCASE ( )
1.1 anton 5234: @end example
5235:
1.131 anton 5236: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}. If
5237: no @i{ni} matches, the optional @i{default-code} is executed. The
5238: optional default case can be added by simply writing the code after
5239: the last @code{ENDOF}. It may use @i{n}, which is on top of the stack,
5240: but must not consume it. The value @i{n} is consumed by this
5241: construction (either by a OF that matches, or by the ENDCASE, if no OF
5242: matches).
1.1 anton 5243:
1.69 anton 5244: @progstyle
1.131 anton 5245: To keep the code understandable, you should ensure that you change the
5246: stack in the same way (wrt. number and types of stack items consumed
5247: and pushed) on all paths through a selection construct.
1.69 anton 5248:
1.1 anton 5249: @node Simple Loops, Counted Loops, Selection, Control Structures
5250: @subsection Simple Loops
5251: @cindex simple loops
5252: @cindex loops without count
5253:
5254: @cindex @code{WHILE} loop
5255: @example
5256: BEGIN
1.29 crook 5257: @i{code1}
5258: @i{flag}
1.1 anton 5259: WHILE
1.29 crook 5260: @i{code2}
1.1 anton 5261: REPEAT
5262: @end example
5263:
1.29 crook 5264: @i{code1} is executed and @i{flag} is computed. If it is true,
5265: @i{code2} is executed and the loop is restarted; If @i{flag} is
1.1 anton 5266: false, execution continues after the @code{REPEAT}.
5267:
5268: @cindex @code{UNTIL} loop
5269: @example
5270: BEGIN
1.29 crook 5271: @i{code}
5272: @i{flag}
1.1 anton 5273: UNTIL
5274: @end example
5275:
1.29 crook 5276: @i{code} is executed. The loop is restarted if @code{flag} is false.
1.1 anton 5277:
1.69 anton 5278: @progstyle
5279: To keep the code understandable, a complete iteration of the loop should
5280: not change the number and types of the items on the stacks.
5281:
1.1 anton 5282: @cindex endless loop
5283: @cindex loops, endless
5284: @example
5285: BEGIN
1.29 crook 5286: @i{code}
1.1 anton 5287: AGAIN
5288: @end example
5289:
5290: This is an endless loop.
5291:
5292: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5293: @subsection Counted Loops
5294: @cindex counted loops
5295: @cindex loops, counted
5296: @cindex @code{DO} loops
5297:
5298: The basic counted loop is:
5299: @example
1.29 crook 5300: @i{limit} @i{start}
1.1 anton 5301: ?DO
1.29 crook 5302: @i{body}
1.1 anton 5303: LOOP
5304: @end example
5305:
1.29 crook 5306: This performs one iteration for every integer, starting from @i{start}
5307: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
1.21 crook 5308: accessed with @code{i}. For example, the loop:
1.1 anton 5309: @example
5310: 10 0 ?DO
5311: i .
5312: LOOP
5313: @end example
1.21 crook 5314: @noindent
5315: prints @code{0 1 2 3 4 5 6 7 8 9}
5316:
1.1 anton 5317: The index of the innermost loop can be accessed with @code{i}, the index
5318: of the next loop with @code{j}, and the index of the third loop with
5319: @code{k}.
5320:
1.44 crook 5321:
1.1 anton 5322: doc-i
5323: doc-j
5324: doc-k
5325:
1.44 crook 5326:
1.1 anton 5327: The loop control data are kept on the return stack, so there are some
1.21 crook 5328: restrictions on mixing return stack accesses and counted loop words. In
5329: particuler, if you put values on the return stack outside the loop, you
5330: cannot read them inside the loop@footnote{well, not in a way that is
5331: portable.}. If you put values on the return stack within a loop, you
5332: have to remove them before the end of the loop and before accessing the
5333: index of the loop.
1.1 anton 5334:
5335: There are several variations on the counted loop:
5336:
1.21 crook 5337: @itemize @bullet
5338: @item
5339: @code{LEAVE} leaves the innermost counted loop immediately; execution
5340: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5341:
5342: @example
5343: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5344: @end example
5345: prints @code{0 1 2 3}
5346:
1.1 anton 5347:
1.21 crook 5348: @item
5349: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5350: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5351: return stack so @code{EXIT} can get to its return address. For example:
5352:
5353: @example
5354: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5355: @end example
5356: prints @code{0 1 2 3}
5357:
5358:
5359: @item
1.29 crook 5360: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
1.1 anton 5361: (and @code{LOOP} iterates until they become equal by wrap-around
5362: arithmetic). This behaviour is usually not what you want. Therefore,
5363: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
1.29 crook 5364: @code{?DO}), which do not enter the loop if @i{start} is greater than
5365: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
1.1 anton 5366: unsigned loop parameters.
5367:
1.21 crook 5368: @item
5369: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5370: the loop, independent of the loop parameters. Do not use @code{DO}, even
5371: if you know that the loop is entered in any case. Such knowledge tends
5372: to become invalid during maintenance of a program, and then the
5373: @code{DO} will make trouble.
5374:
5375: @item
1.29 crook 5376: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5377: index by @i{n} instead of by 1. The loop is terminated when the border
5378: between @i{limit-1} and @i{limit} is crossed. E.g.:
1.1 anton 5379:
1.21 crook 5380: @example
5381: 4 0 +DO i . 2 +LOOP
5382: @end example
5383: @noindent
5384: prints @code{0 2}
5385:
5386: @example
5387: 4 1 +DO i . 2 +LOOP
5388: @end example
5389: @noindent
5390: prints @code{1 3}
1.1 anton 5391:
1.68 anton 5392: @item
1.1 anton 5393: @cindex negative increment for counted loops
5394: @cindex counted loops with negative increment
1.29 crook 5395: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
1.1 anton 5396:
1.21 crook 5397: @example
5398: -1 0 ?DO i . -1 +LOOP
5399: @end example
5400: @noindent
5401: prints @code{0 -1}
1.1 anton 5402:
1.21 crook 5403: @example
5404: 0 0 ?DO i . -1 +LOOP
5405: @end example
5406: prints nothing.
1.1 anton 5407:
1.29 crook 5408: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5409: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5410: index by @i{u} each iteration. The loop is terminated when the border
5411: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
1.1 anton 5412: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5413:
1.21 crook 5414: @example
5415: -2 0 -DO i . 1 -LOOP
5416: @end example
5417: @noindent
5418: prints @code{0 -1}
1.1 anton 5419:
1.21 crook 5420: @example
5421: -1 0 -DO i . 1 -LOOP
5422: @end example
5423: @noindent
5424: prints @code{0}
5425:
5426: @example
5427: 0 0 -DO i . 1 -LOOP
5428: @end example
5429: @noindent
5430: prints nothing.
1.1 anton 5431:
1.21 crook 5432: @end itemize
1.1 anton 5433:
5434: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
1.26 crook 5435: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5436: for these words that uses only standard words is provided in
5437: @file{compat/loops.fs}.
1.1 anton 5438:
5439:
5440: @cindex @code{FOR} loops
1.26 crook 5441: Another counted loop is:
1.1 anton 5442: @example
1.29 crook 5443: @i{n}
1.1 anton 5444: FOR
1.29 crook 5445: @i{body}
1.1 anton 5446: NEXT
5447: @end example
5448: This is the preferred loop of native code compiler writers who are too
1.26 crook 5449: lazy to optimize @code{?DO} loops properly. This loop structure is not
1.29 crook 5450: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5451: @code{i} produces values starting with @i{n} and ending with 0. Other
1.26 crook 5452: Forth systems may behave differently, even if they support @code{FOR}
5453: loops. To avoid problems, don't use @code{FOR} loops.
1.1 anton 5454:
5455: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5456: @subsection Arbitrary control structures
5457: @cindex control structures, user-defined
5458:
5459: @cindex control-flow stack
5460: ANS Forth permits and supports using control structures in a non-nested
5461: way. Information about incomplete control structures is stored on the
5462: control-flow stack. This stack may be implemented on the Forth data
5463: stack, and this is what we have done in Gforth.
5464:
5465: @cindex @code{orig}, control-flow stack item
5466: @cindex @code{dest}, control-flow stack item
5467: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5468: entry represents a backward branch target. A few words are the basis for
5469: building any control structure possible (except control structures that
5470: need storage, like calls, coroutines, and backtracking).
5471:
1.44 crook 5472:
1.1 anton 5473: doc-if
5474: doc-ahead
5475: doc-then
5476: doc-begin
5477: doc-until
5478: doc-again
5479: doc-cs-pick
5480: doc-cs-roll
5481:
1.44 crook 5482:
1.21 crook 5483: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5484: manipulate the control-flow stack in a portable way. Without them, you
5485: would need to know how many stack items are occupied by a control-flow
5486: entry (many systems use one cell. In Gforth they currently take three,
5487: but this may change in the future).
5488:
1.1 anton 5489: Some standard control structure words are built from these words:
5490:
1.44 crook 5491:
1.1 anton 5492: doc-else
5493: doc-while
5494: doc-repeat
5495:
1.44 crook 5496:
5497: @noindent
1.1 anton 5498: Gforth adds some more control-structure words:
5499:
1.44 crook 5500:
1.1 anton 5501: doc-endif
5502: doc-?dup-if
5503: doc-?dup-0=-if
5504:
1.44 crook 5505:
5506: @noindent
1.1 anton 5507: Counted loop words constitute a separate group of words:
5508:
1.44 crook 5509:
1.1 anton 5510: doc-?do
5511: doc-+do
5512: doc-u+do
5513: doc--do
5514: doc-u-do
5515: doc-do
5516: doc-for
5517: doc-loop
5518: doc-+loop
5519: doc--loop
5520: doc-next
5521: doc-leave
5522: doc-?leave
5523: doc-unloop
5524: doc-done
5525:
1.44 crook 5526:
1.21 crook 5527: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5528: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
1.1 anton 5529: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5530: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5531: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5532: resolved (by using one of the loop-ending words or @code{DONE}).
5533:
1.44 crook 5534: @noindent
1.26 crook 5535: Another group of control structure words are:
1.1 anton 5536:
1.44 crook 5537:
1.1 anton 5538: doc-case
5539: doc-endcase
5540: doc-of
5541: doc-endof
5542:
1.44 crook 5543:
1.21 crook 5544: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5545: @code{CS-ROLL}.
1.1 anton 5546:
5547: @subsubsection Programming Style
1.47 crook 5548: @cindex control structures programming style
5549: @cindex programming style, arbitrary control structures
1.1 anton 5550:
5551: In order to ensure readability we recommend that you do not create
5552: arbitrary control structures directly, but define new control structure
5553: words for the control structure you want and use these words in your
1.26 crook 5554: program. For example, instead of writing:
1.1 anton 5555:
5556: @example
1.26 crook 5557: BEGIN
1.1 anton 5558: ...
1.26 crook 5559: IF [ 1 CS-ROLL ]
1.1 anton 5560: ...
1.26 crook 5561: AGAIN THEN
1.1 anton 5562: @end example
5563:
1.21 crook 5564: @noindent
1.1 anton 5565: we recommend defining control structure words, e.g.,
5566:
5567: @example
1.26 crook 5568: : WHILE ( DEST -- ORIG DEST )
5569: POSTPONE IF
5570: 1 CS-ROLL ; immediate
5571:
5572: : REPEAT ( orig dest -- )
5573: POSTPONE AGAIN
5574: POSTPONE THEN ; immediate
1.1 anton 5575: @end example
5576:
1.21 crook 5577: @noindent
1.1 anton 5578: and then using these to create the control structure:
5579:
5580: @example
1.26 crook 5581: BEGIN
1.1 anton 5582: ...
1.26 crook 5583: WHILE
1.1 anton 5584: ...
1.26 crook 5585: REPEAT
1.1 anton 5586: @end example
5587:
5588: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5589: @code{WHILE} are predefined, so in this example it would not be
5590: necessary to define them.
5591:
5592: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5593: @subsection Calls and returns
5594: @cindex calling a definition
5595: @cindex returning from a definition
5596:
1.3 anton 5597: @cindex recursive definitions
5598: A definition can be called simply be writing the name of the definition
1.26 crook 5599: to be called. Normally a definition is invisible during its own
1.3 anton 5600: definition. If you want to write a directly recursive definition, you
1.26 crook 5601: can use @code{recursive} to make the current definition visible, or
5602: @code{recurse} to call the current definition directly.
1.3 anton 5603:
1.44 crook 5604:
1.3 anton 5605: doc-recursive
5606: doc-recurse
5607:
1.44 crook 5608:
1.21 crook 5609: @comment TODO add example of the two recursion methods
1.12 anton 5610: @quotation
5611: @progstyle
5612: I prefer using @code{recursive} to @code{recurse}, because calling the
5613: definition by name is more descriptive (if the name is well-chosen) than
5614: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5615: implementation, it is much better to read (and think) ``now sort the
5616: partitions'' than to read ``now do a recursive call''.
5617: @end quotation
1.3 anton 5618:
1.29 crook 5619: For mutual recursion, use @code{Defer}red words, like this:
1.3 anton 5620:
5621: @example
1.28 crook 5622: Defer foo
1.3 anton 5623:
5624: : bar ( ... -- ... )
5625: ... foo ... ;
5626:
5627: :noname ( ... -- ... )
5628: ... bar ... ;
5629: IS foo
5630: @end example
5631:
1.170 pazsan 5632: Deferred words are discussed in more detail in @ref{Deferred Words}.
1.33 anton 5633:
1.26 crook 5634: The current definition returns control to the calling definition when
1.33 anton 5635: the end of the definition is reached or @code{EXIT} is encountered.
1.1 anton 5636:
5637: doc-exit
5638: doc-;s
5639:
1.44 crook 5640:
1.1 anton 5641: @node Exception Handling, , Calls and returns, Control Structures
5642: @subsection Exception Handling
1.26 crook 5643: @cindex exceptions
1.1 anton 5644:
1.68 anton 5645: @c quit is a very bad idea for error handling,
5646: @c because it does not translate into a THROW
5647: @c it also does not belong into this chapter
5648:
5649: If a word detects an error condition that it cannot handle, it can
5650: @code{throw} an exception. In the simplest case, this will terminate
5651: your program, and report an appropriate error.
1.21 crook 5652:
1.68 anton 5653: doc-throw
1.1 anton 5654:
1.69 anton 5655: @code{Throw} consumes a cell-sized error number on the stack. There are
5656: some predefined error numbers in ANS Forth (see @file{errors.fs}). In
5657: Gforth (and most other systems) you can use the iors produced by various
5658: words as error numbers (e.g., a typical use of @code{allocate} is
5659: @code{allocate throw}). Gforth also provides the word @code{exception}
5660: to define your own error numbers (with decent error reporting); an ANS
5661: Forth version of this word (but without the error messages) is available
5662: in @code{compat/except.fs}. And finally, you can use your own error
1.68 anton 5663: numbers (anything outside the range -4095..0), but won't get nice error
5664: messages, only numbers. For example, try:
5665:
5666: @example
1.69 anton 5667: -10 throw \ ANS defined
5668: -267 throw \ system defined
5669: s" my error" exception throw \ user defined
5670: 7 throw \ arbitrary number
1.68 anton 5671: @end example
5672:
5673: doc---exception-exception
1.1 anton 5674:
1.69 anton 5675: A common idiom to @code{THROW} a specific error if a flag is true is
5676: this:
5677:
5678: @example
5679: @code{( flag ) 0<> @i{errno} and throw}
5680: @end example
5681:
5682: Your program can provide exception handlers to catch exceptions. An
5683: exception handler can be used to correct the problem, or to clean up
5684: some data structures and just throw the exception to the next exception
5685: handler. Note that @code{throw} jumps to the dynamically innermost
5686: exception handler. The system's exception handler is outermost, and just
5687: prints an error and restarts command-line interpretation (or, in batch
5688: mode (i.e., while processing the shell command line), leaves Gforth).
1.1 anton 5689:
1.68 anton 5690: The ANS Forth way to catch exceptions is @code{catch}:
1.1 anton 5691:
1.68 anton 5692: doc-catch
1.160 anton 5693: doc-nothrow
1.68 anton 5694:
5695: The most common use of exception handlers is to clean up the state when
5696: an error happens. E.g.,
1.1 anton 5697:
1.26 crook 5698: @example
1.68 anton 5699: base @ >r hex \ actually the hex should be inside foo, or we h
5700: ['] foo catch ( nerror|0 )
5701: r> base !
1.69 anton 5702: ( nerror|0 ) throw \ pass it on
1.26 crook 5703: @end example
1.1 anton 5704:
1.69 anton 5705: A use of @code{catch} for handling the error @code{myerror} might look
5706: like this:
1.44 crook 5707:
1.68 anton 5708: @example
1.69 anton 5709: ['] foo catch
5710: CASE
1.160 anton 5711: myerror OF ... ( do something about it ) nothrow ENDOF
1.69 anton 5712: dup throw \ default: pass other errors on, do nothing on non-errors
5713: ENDCASE
1.68 anton 5714: @end example
1.44 crook 5715:
1.68 anton 5716: Having to wrap the code into a separate word is often cumbersome,
5717: therefore Gforth provides an alternative syntax:
1.1 anton 5718:
5719: @example
1.69 anton 5720: TRY
1.68 anton 5721: @i{code1}
1.172 anton 5722: IFERROR
5723: @i{code2}
5724: THEN
5725: @i{code3}
1.69 anton 5726: ENDTRY
1.1 anton 5727: @end example
5728:
1.172 anton 5729: This performs @i{code1}. If @i{code1} completes normally, execution
5730: continues with @i{code3}. If @i{code1} or there is an exception
5731: before @code{endtry}, the stacks are reset to the state during
5732: @code{try}, the throw value is pushed on the data stack, and execution
5733: constinues at @i{code2}, and finally falls through the @i{code3}.
1.26 crook 5734:
1.68 anton 5735: doc-try
5736: doc-endtry
1.172 anton 5737: doc-iferror
5738:
5739: If you don't need @i{code2}, you can write @code{restore} instead of
5740: @code{iferror then}:
5741:
5742: @example
5743: TRY
5744: @i{code1}
5745: RESTORE
5746: @i{code3}
5747: ENDTRY
5748: @end example
1.26 crook 5749:
1.172 anton 5750: @cindex unwind-protect
1.69 anton 5751: The cleanup example from above in this syntax:
1.26 crook 5752:
1.68 anton 5753: @example
1.174 anton 5754: base @@ @{ oldbase @}
1.172 anton 5755: TRY
1.68 anton 5756: hex foo \ now the hex is placed correctly
1.69 anton 5757: 0 \ value for throw
1.172 anton 5758: RESTORE
5759: oldbase base !
5760: ENDTRY
5761: throw
1.1 anton 5762: @end example
5763:
1.172 anton 5764: An additional advantage of this variant is that an exception between
5765: @code{restore} and @code{endtry} (e.g., from the user pressing
5766: @kbd{Ctrl-C}) restarts the execution of the code after @code{restore},
5767: so the base will be restored under all circumstances.
5768:
5769: However, you have to ensure that this code does not cause an exception
5770: itself, otherwise the @code{iferror}/@code{restore} code will loop.
5771: Moreover, you should also make sure that the stack contents needed by
5772: the @code{iferror}/@code{restore} code exist everywhere between
5773: @code{try} and @code{endtry}; in our example this is achived by
5774: putting the data in a local before the @code{try} (you cannot use the
5775: return stack because the exception frame (@i{sys1}) is in the way
5776: there).
5777:
5778: This kind of usage corresponds to Lisp's @code{unwind-protect}.
5779:
5780: @cindex @code{recover} (old Gforth versions)
5781: If you do not want this exception-restarting behaviour, you achieve
5782: this as follows:
5783:
5784: @example
5785: TRY
5786: @i{code1}
5787: ENDTRY-IFERROR
5788: @i{code2}
5789: THEN
5790: @end example
5791:
5792: If there is an exception in @i{code1}, then @i{code2} is executed,
5793: otherwise execution continues behind the @code{then} (or in a possible
5794: @code{else} branch). This corresponds to the construct
5795:
5796: @example
5797: TRY
5798: @i{code1}
5799: RECOVER
5800: @i{code2}
5801: ENDTRY
5802: @end example
5803:
5804: in Gforth before version 0.7. So you can directly replace
5805: @code{recover}-using code; however, we recommend that you check if it
5806: would not be better to use one of the other @code{try} variants while
5807: you are at it.
5808:
1.173 anton 5809: To ease the transition, Gforth provides two compatibility files:
5810: @file{endtry-iferror.fs} provides the @code{try ... endtry-iferror
5811: ... then} syntax (but not @code{iferror} or @code{restore}) for old
5812: systems; @file{recover-endtry.fs} provides the @code{try ... recover
5813: ... endtry} syntax on new systems, so you can use that file as a
5814: stopgap to run old programs. Both files work on any system (they just
5815: do nothing if the system already has the syntax it implements), so you
5816: can unconditionally @code{require} one of these files, even if you use
5817: a mix old and new systems.
5818:
1.172 anton 5819: doc-restore
5820: doc-endtry-iferror
5821:
5822: Here's the error handling example:
1.1 anton 5823:
1.68 anton 5824: @example
1.69 anton 5825: TRY
1.68 anton 5826: foo
1.172 anton 5827: ENDTRY-IFERROR
1.69 anton 5828: CASE
1.160 anton 5829: myerror OF ... ( do something about it ) nothrow ENDOF
1.69 anton 5830: throw \ pass other errors on
5831: ENDCASE
1.172 anton 5832: THEN
1.68 anton 5833: @end example
1.1 anton 5834:
1.69 anton 5835: @progstyle
5836: As usual, you should ensure that the stack depth is statically known at
5837: the end: either after the @code{throw} for passing on errors, or after
5838: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
5839: selection construct for handling the error).
5840:
1.68 anton 5841: There are two alternatives to @code{throw}: @code{Abort"} is conditional
5842: and you can provide an error message. @code{Abort} just produces an
5843: ``Aborted'' error.
1.1 anton 5844:
1.68 anton 5845: The problem with these words is that exception handlers cannot
5846: differentiate between different @code{abort"}s; they just look like
5847: @code{-2 throw} to them (the error message cannot be accessed by
5848: standard programs). Similar @code{abort} looks like @code{-1 throw} to
5849: exception handlers.
1.44 crook 5850:
1.68 anton 5851: doc-abort"
1.26 crook 5852: doc-abort
1.29 crook 5853:
5854:
1.44 crook 5855:
1.29 crook 5856: @c -------------------------------------------------------------
1.47 crook 5857: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
1.29 crook 5858: @section Defining Words
5859: @cindex defining words
5860:
1.47 crook 5861: Defining words are used to extend Forth by creating new entries in the dictionary.
5862:
1.29 crook 5863: @menu
1.67 anton 5864: * CREATE::
1.44 crook 5865: * Variables:: Variables and user variables
1.67 anton 5866: * Constants::
1.44 crook 5867: * Values:: Initialised variables
1.67 anton 5868: * Colon Definitions::
1.44 crook 5869: * Anonymous Definitions:: Definitions without names
1.69 anton 5870: * Supplying names:: Passing definition names as strings
1.67 anton 5871: * User-defined Defining Words::
1.170 pazsan 5872: * Deferred Words:: Allow forward references
1.67 anton 5873: * Aliases::
1.29 crook 5874: @end menu
5875:
1.44 crook 5876: @node CREATE, Variables, Defining Words, Defining Words
5877: @subsection @code{CREATE}
1.29 crook 5878: @cindex simple defining words
5879: @cindex defining words, simple
5880:
5881: Defining words are used to create new entries in the dictionary. The
5882: simplest defining word is @code{CREATE}. @code{CREATE} is used like
5883: this:
5884:
5885: @example
5886: CREATE new-word1
5887: @end example
5888:
1.69 anton 5889: @code{CREATE} is a parsing word, i.e., it takes an argument from the
5890: input stream (@code{new-word1} in our example). It generates a
5891: dictionary entry for @code{new-word1}. When @code{new-word1} is
5892: executed, all that it does is leave an address on the stack. The address
5893: represents the value of the data space pointer (@code{HERE}) at the time
5894: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
5895: associating a name with the address of a region of memory.
1.29 crook 5896:
1.34 anton 5897: doc-create
5898:
1.69 anton 5899: Note that in ANS Forth guarantees only for @code{create} that its body
5900: is in dictionary data space (i.e., where @code{here}, @code{allot}
5901: etc. work, @pxref{Dictionary allocation}). Also, in ANS Forth only
5902: @code{create}d words can be modified with @code{does>}
5903: (@pxref{User-defined Defining Words}). And in ANS Forth @code{>body}
5904: can only be applied to @code{create}d words.
5905:
1.29 crook 5906: By extending this example to reserve some memory in data space, we end
1.69 anton 5907: up with something like a @i{variable}. Here are two different ways to do
5908: it:
1.29 crook 5909:
5910: @example
5911: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
5912: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
5913: @end example
5914:
5915: The variable can be examined and modified using @code{@@} (``fetch'') and
5916: @code{!} (``store'') like this:
5917:
5918: @example
5919: new-word2 @@ . \ get address, fetch from it and display
5920: 1234 new-word2 ! \ new value, get address, store to it
5921: @end example
5922:
1.44 crook 5923: @cindex arrays
5924: A similar mechanism can be used to create arrays. For example, an
5925: 80-character text input buffer:
1.29 crook 5926:
5927: @example
1.44 crook 5928: CREATE text-buf 80 chars allot
5929:
1.168 anton 5930: text-buf 0 chars + c@@ \ the 1st character (offset 0)
5931: text-buf 3 chars + c@@ \ the 4th character (offset 3)
1.44 crook 5932: @end example
1.29 crook 5933:
1.44 crook 5934: You can build arbitrarily complex data structures by allocating
1.49 anton 5935: appropriate areas of memory. For further discussions of this, and to
1.66 anton 5936: learn about some Gforth tools that make it easier,
1.49 anton 5937: @xref{Structures}.
1.44 crook 5938:
5939:
5940: @node Variables, Constants, CREATE, Defining Words
5941: @subsection Variables
5942: @cindex variables
5943:
5944: The previous section showed how a sequence of commands could be used to
5945: generate a variable. As a final refinement, the whole code sequence can
5946: be wrapped up in a defining word (pre-empting the subject of the next
5947: section), making it easier to create new variables:
5948:
5949: @example
5950: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
5951: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
5952:
5953: myvariableX foo \ variable foo starts off with an unknown value
5954: myvariable0 joe \ whilst joe is initialised to 0
1.29 crook 5955:
5956: 45 3 * foo ! \ set foo to 135
5957: 1234 joe ! \ set joe to 1234
5958: 3 joe +! \ increment joe by 3.. to 1237
5959: @end example
5960:
5961: Not surprisingly, there is no need to define @code{myvariable}, since
1.44 crook 5962: Forth already has a definition @code{Variable}. ANS Forth does not
1.69 anton 5963: guarantee that a @code{Variable} is initialised when it is created
5964: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
5965: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
5966: like @code{myvariable0}). Forth also provides @code{2Variable} and
1.47 crook 5967: @code{fvariable} for double and floating-point variables, respectively
1.69 anton 5968: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
1.47 crook 5969: store a boolean, you can use @code{on} and @code{off} to toggle its
5970: state.
1.29 crook 5971:
1.34 anton 5972: doc-variable
5973: doc-2variable
5974: doc-fvariable
5975:
1.29 crook 5976: @cindex user variables
5977: @cindex user space
5978: The defining word @code{User} behaves in the same way as @code{Variable}.
5979: The difference is that it reserves space in @i{user (data) space} rather
5980: than normal data space. In a Forth system that has a multi-tasker, each
5981: task has its own set of user variables.
5982:
1.34 anton 5983: doc-user
1.67 anton 5984: @c doc-udp
5985: @c doc-uallot
1.34 anton 5986:
1.29 crook 5987: @comment TODO is that stuff about user variables strictly correct? Is it
5988: @comment just terminal tasks that have user variables?
5989: @comment should document tasker.fs (with some examples) elsewhere
5990: @comment in this manual, then expand on user space and user variables.
5991:
1.44 crook 5992: @node Constants, Values, Variables, Defining Words
5993: @subsection Constants
5994: @cindex constants
5995:
5996: @code{Constant} allows you to declare a fixed value and refer to it by
5997: name. For example:
1.29 crook 5998:
5999: @example
6000: 12 Constant INCHES-PER-FOOT
6001: 3E+08 fconstant SPEED-O-LIGHT
6002: @end example
6003:
6004: A @code{Variable} can be both read and written, so its run-time
6005: behaviour is to supply an address through which its current value can be
6006: manipulated. In contrast, the value of a @code{Constant} cannot be
6007: changed once it has been declared@footnote{Well, often it can be -- but
6008: not in a Standard, portable way. It's safer to use a @code{Value} (read
6009: on).} so it's not necessary to supply the address -- it is more
6010: efficient to return the value of the constant directly. That's exactly
6011: what happens; the run-time effect of a constant is to put its value on
1.49 anton 6012: the top of the stack (You can find one
6013: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
1.29 crook 6014:
1.69 anton 6015: Forth also provides @code{2Constant} and @code{fconstant} for defining
1.29 crook 6016: double and floating-point constants, respectively.
6017:
1.34 anton 6018: doc-constant
6019: doc-2constant
6020: doc-fconstant
6021:
6022: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
1.44 crook 6023: @c nac-> How could that not be true in an ANS Forth? You can't define a
6024: @c constant, use it and then delete the definition of the constant..
1.69 anton 6025:
6026: @c anton->An ANS Forth system can compile a constant to a literal; On
6027: @c decompilation you would see only the number, just as if it had been used
6028: @c in the first place. The word will stay, of course, but it will only be
6029: @c used by the text interpreter (no run-time duties, except when it is
6030: @c POSTPONEd or somesuch).
6031:
6032: @c nac:
1.44 crook 6033: @c I agree that it's rather deep, but IMO it is an important difference
6034: @c relative to other programming languages.. often it's annoying: it
6035: @c certainly changes my programming style relative to C.
6036:
1.69 anton 6037: @c anton: In what way?
6038:
1.29 crook 6039: Constants in Forth behave differently from their equivalents in other
6040: programming languages. In other languages, a constant (such as an EQU in
6041: assembler or a #define in C) only exists at compile-time; in the
6042: executable program the constant has been translated into an absolute
6043: number and, unless you are using a symbolic debugger, it's impossible to
6044: know what abstract thing that number represents. In Forth a constant has
1.44 crook 6045: an entry in the header space and remains there after the code that uses
6046: it has been defined. In fact, it must remain in the dictionary since it
6047: has run-time duties to perform. For example:
1.29 crook 6048:
6049: @example
6050: 12 Constant INCHES-PER-FOOT
6051: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
6052: @end example
6053:
6054: @cindex in-lining of constants
6055: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
6056: associated with the constant @code{INCHES-PER-FOOT}. If you use
6057: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
6058: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
6059: attempt to optimise constants by in-lining them where they are used. You
6060: can force Gforth to in-line a constant like this:
6061:
6062: @example
6063: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
6064: @end example
6065:
6066: If you use @code{see} to decompile @i{this} version of
6067: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
1.49 anton 6068: longer present. To understand how this works, read
6069: @ref{Interpret/Compile states}, and @ref{Literals}.
1.29 crook 6070:
6071: In-lining constants in this way might improve execution time
6072: fractionally, and can ensure that a constant is now only referenced at
6073: compile-time. However, the definition of the constant still remains in
6074: the dictionary. Some Forth compilers provide a mechanism for controlling
6075: a second dictionary for holding transient words such that this second
6076: dictionary can be deleted later in order to recover memory
6077: space. However, there is no standard way of doing this.
6078:
6079:
1.44 crook 6080: @node Values, Colon Definitions, Constants, Defining Words
6081: @subsection Values
6082: @cindex values
1.34 anton 6083:
1.69 anton 6084: A @code{Value} behaves like a @code{Constant}, but it can be changed.
6085: @code{TO} is a parsing word that changes a @code{Values}. In Gforth
6086: (not in ANS Forth) you can access (and change) a @code{value} also with
6087: @code{>body}.
6088:
6089: Here are some
6090: examples:
1.29 crook 6091:
6092: @example
1.69 anton 6093: 12 Value APPLES \ Define APPLES with an initial value of 12
6094: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6095: 1 ' APPLES >body +! \ Increment APPLES. Non-standard usage.
6096: APPLES \ puts 35 on the top of the stack.
1.29 crook 6097: @end example
6098:
1.44 crook 6099: doc-value
6100: doc-to
1.29 crook 6101:
1.35 anton 6102:
1.69 anton 6103:
1.44 crook 6104: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6105: @subsection Colon Definitions
6106: @cindex colon definitions
1.35 anton 6107:
6108: @example
1.44 crook 6109: : name ( ... -- ... )
6110: word1 word2 word3 ;
1.29 crook 6111: @end example
6112:
1.44 crook 6113: @noindent
6114: Creates a word called @code{name} that, upon execution, executes
6115: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
1.29 crook 6116:
1.49 anton 6117: The explanation above is somewhat superficial. For simple examples of
6118: colon definitions see @ref{Your first definition}. For an in-depth
1.66 anton 6119: discussion of some of the issues involved, @xref{Interpretation and
1.49 anton 6120: Compilation Semantics}.
1.29 crook 6121:
1.44 crook 6122: doc-:
6123: doc-;
1.1 anton 6124:
1.34 anton 6125:
1.69 anton 6126: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
1.44 crook 6127: @subsection Anonymous Definitions
6128: @cindex colon definitions
6129: @cindex defining words without name
1.34 anton 6130:
1.44 crook 6131: Sometimes you want to define an @dfn{anonymous word}; a word without a
6132: name. You can do this with:
1.1 anton 6133:
1.44 crook 6134: doc-:noname
1.1 anton 6135:
1.44 crook 6136: This leaves the execution token for the word on the stack after the
6137: closing @code{;}. Here's an example in which a deferred word is
6138: initialised with an @code{xt} from an anonymous colon definition:
1.1 anton 6139:
1.29 crook 6140: @example
1.44 crook 6141: Defer deferred
6142: :noname ( ... -- ... )
6143: ... ;
6144: IS deferred
1.29 crook 6145: @end example
1.26 crook 6146:
1.44 crook 6147: @noindent
6148: Gforth provides an alternative way of doing this, using two separate
6149: words:
1.27 crook 6150:
1.44 crook 6151: doc-noname
6152: @cindex execution token of last defined word
1.116 anton 6153: doc-latestxt
1.1 anton 6154:
1.44 crook 6155: @noindent
6156: The previous example can be rewritten using @code{noname} and
1.116 anton 6157: @code{latestxt}:
1.1 anton 6158:
1.26 crook 6159: @example
1.44 crook 6160: Defer deferred
6161: noname : ( ... -- ... )
6162: ... ;
1.116 anton 6163: latestxt IS deferred
1.26 crook 6164: @end example
1.1 anton 6165:
1.29 crook 6166: @noindent
1.44 crook 6167: @code{noname} works with any defining word, not just @code{:}.
6168:
1.116 anton 6169: @code{latestxt} also works when the last word was not defined as
1.71 anton 6170: @code{noname}. It does not work for combined words, though. It also has
6171: the useful property that is is valid as soon as the header for a
6172: definition has been built. Thus:
1.44 crook 6173:
6174: @example
1.116 anton 6175: latestxt . : foo [ latestxt . ] ; ' foo .
1.44 crook 6176: @end example
1.1 anton 6177:
1.44 crook 6178: @noindent
6179: prints 3 numbers; the last two are the same.
1.26 crook 6180:
1.69 anton 6181: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
6182: @subsection Supplying the name of a defined word
6183: @cindex names for defined words
6184: @cindex defining words, name given in a string
6185:
6186: By default, a defining word takes the name for the defined word from the
6187: input stream. Sometimes you want to supply the name from a string. You
6188: can do this with:
6189:
6190: doc-nextname
6191:
6192: For example:
6193:
6194: @example
6195: s" foo" nextname create
6196: @end example
6197:
6198: @noindent
6199: is equivalent to:
6200:
6201: @example
6202: create foo
6203: @end example
6204:
6205: @noindent
6206: @code{nextname} works with any defining word.
6207:
1.1 anton 6208:
1.170 pazsan 6209: @node User-defined Defining Words, Deferred Words, Supplying names, Defining Words
1.26 crook 6210: @subsection User-defined Defining Words
6211: @cindex user-defined defining words
6212: @cindex defining words, user-defined
1.1 anton 6213:
1.29 crook 6214: You can create a new defining word by wrapping defining-time code around
6215: an existing defining word and putting the sequence in a colon
1.69 anton 6216: definition.
6217:
6218: @c anton: This example is very complex and leads in a quite different
6219: @c direction from the CREATE-DOES> stuff that follows. It should probably
6220: @c be done elsewhere, or as a subsubsection of this subsection (or as a
6221: @c subsection of Defining Words)
6222:
6223: For example, suppose that you have a word @code{stats} that
1.29 crook 6224: gathers statistics about colon definitions given the @i{xt} of the
6225: definition, and you want every colon definition in your application to
6226: make a call to @code{stats}. You can define and use a new version of
6227: @code{:} like this:
6228:
6229: @example
6230: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6231: ... ; \ other code
6232:
1.116 anton 6233: : my: : latestxt postpone literal ['] stats compile, ;
1.29 crook 6234:
6235: my: foo + - ;
6236: @end example
6237:
6238: When @code{foo} is defined using @code{my:} these steps occur:
6239:
6240: @itemize @bullet
6241: @item
6242: @code{my:} is executed.
6243: @item
6244: The @code{:} within the definition (the one between @code{my:} and
1.116 anton 6245: @code{latestxt}) is executed, and does just what it always does; it parses
1.29 crook 6246: the input stream for a name, builds a dictionary header for the name
6247: @code{foo} and switches @code{state} from interpret to compile.
6248: @item
1.116 anton 6249: The word @code{latestxt} is executed. It puts the @i{xt} for the word that is
1.29 crook 6250: being defined -- @code{foo} -- onto the stack.
6251: @item
6252: The code that was produced by @code{postpone literal} is executed; this
6253: causes the value on the stack to be compiled as a literal in the code
6254: area of @code{foo}.
6255: @item
6256: The code @code{['] stats} compiles a literal into the definition of
6257: @code{my:}. When @code{compile,} is executed, that literal -- the
6258: execution token for @code{stats} -- is layed down in the code area of
6259: @code{foo} , following the literal@footnote{Strictly speaking, the
6260: mechanism that @code{compile,} uses to convert an @i{xt} into something
6261: in the code area is implementation-dependent. A threaded implementation
6262: might spit out the execution token directly whilst another
6263: implementation might spit out a native code sequence.}.
6264: @item
6265: At this point, the execution of @code{my:} is complete, and control
6266: returns to the text interpreter. The text interpreter is in compile
6267: state, so subsequent text @code{+ -} is compiled into the definition of
6268: @code{foo} and the @code{;} terminates the definition as always.
6269: @end itemize
6270:
6271: You can use @code{see} to decompile a word that was defined using
6272: @code{my:} and see how it is different from a normal @code{:}
6273: definition. For example:
6274:
6275: @example
6276: : bar + - ; \ like foo but using : rather than my:
6277: see bar
6278: : bar
6279: + - ;
6280: see foo
6281: : foo
6282: 107645672 stats + - ;
6283:
1.140 anton 6284: \ use ' foo . to show that 107645672 is the xt for foo
1.29 crook 6285: @end example
6286:
6287: You can use techniques like this to make new defining words in terms of
6288: @i{any} existing defining word.
1.1 anton 6289:
6290:
1.29 crook 6291: @cindex defining defining words
1.26 crook 6292: @cindex @code{CREATE} ... @code{DOES>}
6293: If you want the words defined with your defining words to behave
6294: differently from words defined with standard defining words, you can
6295: write your defining word like this:
1.1 anton 6296:
6297: @example
1.26 crook 6298: : def-word ( "name" -- )
1.29 crook 6299: CREATE @i{code1}
1.26 crook 6300: DOES> ( ... -- ... )
1.29 crook 6301: @i{code2} ;
1.26 crook 6302:
6303: def-word name
1.1 anton 6304: @end example
6305:
1.29 crook 6306: @cindex child words
6307: This fragment defines a @dfn{defining word} @code{def-word} and then
6308: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6309: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6310: is not executed at this time. The word @code{name} is sometimes called a
6311: @dfn{child} of @code{def-word}.
6312:
6313: When you execute @code{name}, the address of the body of @code{name} is
6314: put on the data stack and @i{code2} is executed (the address of the body
6315: of @code{name} is the address @code{HERE} returns immediately after the
1.69 anton 6316: @code{CREATE}, i.e., the address a @code{create}d word returns by
6317: default).
6318:
6319: @c anton:
6320: @c www.dictionary.com says:
6321: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
6322: @c several generations of absence, usually caused by the chance
6323: @c recombination of genes. 2.An individual or a part that exhibits
6324: @c atavism. Also called throwback. 3.The return of a trait or recurrence
6325: @c of previous behavior after a period of absence.
6326: @c
6327: @c Doesn't seem to fit.
1.29 crook 6328:
1.69 anton 6329: @c @cindex atavism in child words
1.33 anton 6330: You can use @code{def-word} to define a set of child words that behave
1.69 anton 6331: similarly; they all have a common run-time behaviour determined by
6332: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
6333: body of the child word. The structure of the data is common to all
6334: children of @code{def-word}, but the data values are specific -- and
6335: private -- to each child word. When a child word is executed, the
6336: address of its private data area is passed as a parameter on TOS to be
6337: used and manipulated@footnote{It is legitimate both to read and write to
6338: this data area.} by @i{code2}.
1.29 crook 6339:
6340: The two fragments of code that make up the defining words act (are
6341: executed) at two completely separate times:
1.1 anton 6342:
1.29 crook 6343: @itemize @bullet
6344: @item
6345: At @i{define time}, the defining word executes @i{code1} to generate a
6346: child word
6347: @item
6348: At @i{child execution time}, when a child word is invoked, @i{code2}
6349: is executed, using parameters (data) that are private and specific to
6350: the child word.
6351: @end itemize
6352:
1.44 crook 6353: Another way of understanding the behaviour of @code{def-word} and
6354: @code{name} is to say that, if you make the following definitions:
1.33 anton 6355: @example
6356: : def-word1 ( "name" -- )
6357: CREATE @i{code1} ;
6358:
6359: : action1 ( ... -- ... )
6360: @i{code2} ;
6361:
6362: def-word1 name1
6363: @end example
6364:
1.44 crook 6365: @noindent
6366: Then using @code{name1 action1} is equivalent to using @code{name}.
1.1 anton 6367:
1.29 crook 6368: The classic example is that you can define @code{CONSTANT} in this way:
1.26 crook 6369:
1.1 anton 6370: @example
1.29 crook 6371: : CONSTANT ( w "name" -- )
6372: CREATE ,
1.26 crook 6373: DOES> ( -- w )
6374: @@ ;
1.1 anton 6375: @end example
6376:
1.29 crook 6377: @comment There is a beautiful description of how this works and what
6378: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6379: @comment commentary on the Counting Fruits problem.
6380:
6381: When you create a constant with @code{5 CONSTANT five}, a set of
6382: define-time actions take place; first a new word @code{five} is created,
6383: then the value 5 is laid down in the body of @code{five} with
1.44 crook 6384: @code{,}. When @code{five} is executed, the address of the body is put on
1.29 crook 6385: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6386: no code of its own; it simply contains a data field and a pointer to the
6387: code that follows @code{DOES>} in its defining word. That makes words
6388: created in this way very compact.
6389:
6390: The final example in this section is intended to remind you that space
6391: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6392: both read and written by a Standard program@footnote{Exercise: use this
6393: example as a starting point for your own implementation of @code{Value}
6394: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6395: @code{[']}.}:
6396:
6397: @example
6398: : foo ( "name" -- )
6399: CREATE -1 ,
6400: DOES> ( -- )
1.33 anton 6401: @@ . ;
1.29 crook 6402:
6403: foo first-word
6404: foo second-word
6405:
6406: 123 ' first-word >BODY !
6407: @end example
6408:
6409: If @code{first-word} had been a @code{CREATE}d word, we could simply
6410: have executed it to get the address of its data field. However, since it
6411: was defined to have @code{DOES>} actions, its execution semantics are to
6412: perform those @code{DOES>} actions. To get the address of its data field
6413: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6414: translate the xt into the address of the data field. When you execute
6415: @code{first-word}, it will display @code{123}. When you execute
6416: @code{second-word} it will display @code{-1}.
1.26 crook 6417:
6418: @cindex stack effect of @code{DOES>}-parts
6419: @cindex @code{DOES>}-parts, stack effect
1.29 crook 6420: In the examples above the stack comment after the @code{DOES>} specifies
1.26 crook 6421: the stack effect of the defined words, not the stack effect of the
6422: following code (the following code expects the address of the body on
6423: the top of stack, which is not reflected in the stack comment). This is
6424: the convention that I use and recommend (it clashes a bit with using
6425: locals declarations for stack effect specification, though).
1.1 anton 6426:
1.53 anton 6427: @menu
6428: * CREATE..DOES> applications::
6429: * CREATE..DOES> details::
1.63 anton 6430: * Advanced does> usage example::
1.155 anton 6431: * Const-does>::
1.53 anton 6432: @end menu
6433:
6434: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
1.26 crook 6435: @subsubsection Applications of @code{CREATE..DOES>}
6436: @cindex @code{CREATE} ... @code{DOES>}, applications
1.1 anton 6437:
1.26 crook 6438: You may wonder how to use this feature. Here are some usage patterns:
1.1 anton 6439:
1.26 crook 6440: @cindex factoring similar colon definitions
6441: When you see a sequence of code occurring several times, and you can
6442: identify a meaning, you will factor it out as a colon definition. When
6443: you see similar colon definitions, you can factor them using
6444: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6445: that look very similar:
1.1 anton 6446: @example
1.26 crook 6447: : ori, ( reg-target reg-source n -- )
6448: 0 asm-reg-reg-imm ;
6449: : andi, ( reg-target reg-source n -- )
6450: 1 asm-reg-reg-imm ;
1.1 anton 6451: @end example
6452:
1.26 crook 6453: @noindent
6454: This could be factored with:
6455: @example
6456: : reg-reg-imm ( op-code -- )
6457: CREATE ,
6458: DOES> ( reg-target reg-source n -- )
6459: @@ asm-reg-reg-imm ;
6460:
6461: 0 reg-reg-imm ori,
6462: 1 reg-reg-imm andi,
6463: @end example
1.1 anton 6464:
1.26 crook 6465: @cindex currying
6466: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6467: supply a part of the parameters for a word (known as @dfn{currying} in
6468: the functional language community). E.g., @code{+} needs two
6469: parameters. Creating versions of @code{+} with one parameter fixed can
6470: be done like this:
1.82 anton 6471:
1.1 anton 6472: @example
1.82 anton 6473: : curry+ ( n1 "name" -- )
1.26 crook 6474: CREATE ,
6475: DOES> ( n2 -- n1+n2 )
6476: @@ + ;
6477:
6478: 3 curry+ 3+
6479: -2 curry+ 2-
1.1 anton 6480: @end example
6481:
1.91 anton 6482:
1.63 anton 6483: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
1.26 crook 6484: @subsubsection The gory details of @code{CREATE..DOES>}
6485: @cindex @code{CREATE} ... @code{DOES>}, details
1.1 anton 6486:
1.26 crook 6487: doc-does>
1.1 anton 6488:
1.26 crook 6489: @cindex @code{DOES>} in a separate definition
6490: This means that you need not use @code{CREATE} and @code{DOES>} in the
6491: same definition; you can put the @code{DOES>}-part in a separate
1.29 crook 6492: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
1.26 crook 6493: @example
6494: : does1
6495: DOES> ( ... -- ... )
1.44 crook 6496: ... ;
6497:
6498: : does2
6499: DOES> ( ... -- ... )
6500: ... ;
6501:
6502: : def-word ( ... -- ... )
6503: create ...
6504: IF
6505: does1
6506: ELSE
6507: does2
6508: ENDIF ;
6509: @end example
6510:
6511: In this example, the selection of whether to use @code{does1} or
1.69 anton 6512: @code{does2} is made at definition-time; at the time that the child word is
1.44 crook 6513: @code{CREATE}d.
6514:
6515: @cindex @code{DOES>} in interpretation state
6516: In a standard program you can apply a @code{DOES>}-part only if the last
6517: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6518: will override the behaviour of the last word defined in any case. In a
6519: standard program, you can use @code{DOES>} only in a colon
6520: definition. In Gforth, you can also use it in interpretation state, in a
6521: kind of one-shot mode; for example:
6522: @example
6523: CREATE name ( ... -- ... )
6524: @i{initialization}
6525: DOES>
6526: @i{code} ;
6527: @end example
6528:
6529: @noindent
6530: is equivalent to the standard:
6531: @example
6532: :noname
6533: DOES>
6534: @i{code} ;
6535: CREATE name EXECUTE ( ... -- ... )
6536: @i{initialization}
6537: @end example
6538:
1.53 anton 6539: doc->body
6540:
1.152 pazsan 6541: @node Advanced does> usage example, Const-does>, CREATE..DOES> details, User-defined Defining Words
1.63 anton 6542: @subsubsection Advanced does> usage example
6543:
6544: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6545: for disassembling instructions, that follow a very repetetive scheme:
6546:
6547: @example
6548: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6549: @var{entry-num} cells @var{table} + !
6550: @end example
6551:
6552: Of course, this inspires the idea to factor out the commonalities to
6553: allow a definition like
6554:
6555: @example
6556: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6557: @end example
6558:
6559: The parameters @var{disasm-operands} and @var{table} are usually
1.69 anton 6560: correlated. Moreover, before I wrote the disassembler, there already
6561: existed code that defines instructions like this:
1.63 anton 6562:
6563: @example
6564: @var{entry-num} @var{inst-format} @var{inst-name}
6565: @end example
6566:
6567: This code comes from the assembler and resides in
6568: @file{arch/mips/insts.fs}.
6569:
6570: So I had to define the @var{inst-format} words that performed the scheme
6571: above when executed. At first I chose to use run-time code-generation:
6572:
6573: @example
6574: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6575: :noname Postpone @var{disasm-operands}
6576: name Postpone sliteral Postpone type Postpone ;
6577: swap cells @var{table} + ! ;
6578: @end example
6579:
6580: Note that this supplies the other two parameters of the scheme above.
1.44 crook 6581:
1.63 anton 6582: An alternative would have been to write this using
6583: @code{create}/@code{does>}:
6584:
6585: @example
6586: : @var{inst-format} ( entry-num "name" -- )
6587: here name string, ( entry-num c-addr ) \ parse and save "name"
6588: noname create , ( entry-num )
1.116 anton 6589: latestxt swap cells @var{table} + !
1.63 anton 6590: does> ( addr w -- )
6591: \ disassemble instruction w at addr
6592: @@ >r
6593: @var{disasm-operands}
6594: r> count type ;
6595: @end example
6596:
6597: Somehow the first solution is simpler, mainly because it's simpler to
6598: shift a string from definition-time to use-time with @code{sliteral}
6599: than with @code{string,} and friends.
6600:
6601: I wrote a lot of words following this scheme and soon thought about
6602: factoring out the commonalities among them. Note that this uses a
6603: two-level defining word, i.e., a word that defines ordinary defining
6604: words.
6605:
6606: This time a solution involving @code{postpone} and friends seemed more
6607: difficult (try it as an exercise), so I decided to use a
6608: @code{create}/@code{does>} word; since I was already at it, I also used
6609: @code{create}/@code{does>} for the lower level (try using
6610: @code{postpone} etc. as an exercise), resulting in the following
6611: definition:
6612:
6613: @example
6614: : define-format ( disasm-xt table-xt -- )
6615: \ define an instruction format that uses disasm-xt for
6616: \ disassembling and enters the defined instructions into table
6617: \ table-xt
6618: create 2,
6619: does> ( u "inst" -- )
6620: \ defines an anonymous word for disassembling instruction inst,
6621: \ and enters it as u-th entry into table-xt
6622: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6623: noname create 2, \ define anonymous word
1.116 anton 6624: execute latestxt swap ! \ enter xt of defined word into table-xt
1.63 anton 6625: does> ( addr w -- )
6626: \ disassemble instruction w at addr
6627: 2@@ >r ( addr w disasm-xt R: c-addr )
6628: execute ( R: c-addr ) \ disassemble operands
6629: r> count type ; \ print name
6630: @end example
6631:
6632: Note that the tables here (in contrast to above) do the @code{cells +}
6633: by themselves (that's why you have to pass an xt). This word is used in
6634: the following way:
6635:
6636: @example
6637: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6638: @end example
6639:
1.71 anton 6640: As shown above, the defined instruction format is then used like this:
6641:
6642: @example
6643: @var{entry-num} @var{inst-format} @var{inst-name}
6644: @end example
6645:
1.63 anton 6646: In terms of currying, this kind of two-level defining word provides the
6647: parameters in three stages: first @var{disasm-operands} and @var{table},
6648: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6649: the instruction to be disassembled.
6650:
6651: Of course this did not quite fit all the instruction format names used
6652: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6653: the parameters into the right form.
6654:
6655: If you have trouble following this section, don't worry. First, this is
6656: involved and takes time (and probably some playing around) to
6657: understand; second, this is the first two-level
6658: @code{create}/@code{does>} word I have written in seventeen years of
6659: Forth; and if I did not have @file{insts.fs} to start with, I may well
6660: have elected to use just a one-level defining word (with some repeating
6661: of parameters when using the defining word). So it is not necessary to
6662: understand this, but it may improve your understanding of Forth.
1.44 crook 6663:
6664:
1.152 pazsan 6665: @node Const-does>, , Advanced does> usage example, User-defined Defining Words
1.91 anton 6666: @subsubsection @code{Const-does>}
6667:
6668: A frequent use of @code{create}...@code{does>} is for transferring some
6669: values from definition-time to run-time. Gforth supports this use with
6670:
6671: doc-const-does>
6672:
6673: A typical use of this word is:
6674:
6675: @example
6676: : curry+ ( n1 "name" -- )
6677: 1 0 CONST-DOES> ( n2 -- n1+n2 )
6678: + ;
6679:
6680: 3 curry+ 3+
6681: @end example
6682:
6683: Here the @code{1 0} means that 1 cell and 0 floats are transferred from
6684: definition to run-time.
6685:
6686: The advantages of using @code{const-does>} are:
6687:
6688: @itemize
6689:
6690: @item
6691: You don't have to deal with storing and retrieving the values, i.e.,
6692: your program becomes more writable and readable.
6693:
6694: @item
6695: When using @code{does>}, you have to introduce a @code{@@} that cannot
6696: be optimized away (because you could change the data using
6697: @code{>body}...@code{!}); @code{const-does>} avoids this problem.
6698:
6699: @end itemize
6700:
6701: An ANS Forth implementation of @code{const-does>} is available in
6702: @file{compat/const-does.fs}.
6703:
6704:
1.170 pazsan 6705: @node Deferred Words, Aliases, User-defined Defining Words, Defining Words
6706: @subsection Deferred Words
1.44 crook 6707: @cindex deferred words
6708:
6709: The defining word @code{Defer} allows you to define a word by name
6710: without defining its behaviour; the definition of its behaviour is
6711: deferred. Here are two situation where this can be useful:
6712:
6713: @itemize @bullet
6714: @item
6715: Where you want to allow the behaviour of a word to be altered later, and
6716: for all precompiled references to the word to change when its behaviour
6717: is changed.
6718: @item
6719: For mutual recursion; @xref{Calls and returns}.
6720: @end itemize
6721:
6722: In the following example, @code{foo} always invokes the version of
6723: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6724: always invokes the version that prints ``@code{Hello}''. There is no way
6725: of getting @code{foo} to use the later version without re-ordering the
6726: source code and recompiling it.
6727:
6728: @example
6729: : greet ." Good morning" ;
6730: : foo ... greet ... ;
6731: : greet ." Hello" ;
6732: : bar ... greet ... ;
6733: @end example
6734:
6735: This problem can be solved by defining @code{greet} as a @code{Defer}red
6736: word. The behaviour of a @code{Defer}red word can be defined and
6737: redefined at any time by using @code{IS} to associate the xt of a
6738: previously-defined word with it. The previous example becomes:
6739:
6740: @example
1.69 anton 6741: Defer greet ( -- )
1.44 crook 6742: : foo ... greet ... ;
6743: : bar ... greet ... ;
1.69 anton 6744: : greet1 ( -- ) ." Good morning" ;
6745: : greet2 ( -- ) ." Hello" ;
1.132 anton 6746: ' greet2 IS greet \ make greet behave like greet2
1.44 crook 6747: @end example
6748:
1.69 anton 6749: @progstyle
6750: You should write a stack comment for every deferred word, and put only
6751: XTs into deferred words that conform to this stack effect. Otherwise
6752: it's too difficult to use the deferred word.
6753:
1.44 crook 6754: A deferred word can be used to improve the statistics-gathering example
6755: from @ref{User-defined Defining Words}; rather than edit the
6756: application's source code to change every @code{:} to a @code{my:}, do
6757: this:
6758:
6759: @example
6760: : real: : ; \ retain access to the original
6761: defer : \ redefine as a deferred word
1.132 anton 6762: ' my: IS : \ use special version of :
1.44 crook 6763: \
6764: \ load application here
6765: \
1.132 anton 6766: ' real: IS : \ go back to the original
1.44 crook 6767: @end example
6768:
6769:
1.132 anton 6770: One thing to note is that @code{IS} has special compilation semantics,
6771: such that it parses the name at compile time (like @code{TO}):
1.44 crook 6772:
6773: @example
6774: : set-greet ( xt -- )
1.132 anton 6775: IS greet ;
1.44 crook 6776:
6777: ' greet1 set-greet
6778: @end example
6779:
1.132 anton 6780: In situations where @code{IS} does not fit, use @code{defer!} instead.
6781:
1.69 anton 6782: A deferred word can only inherit execution semantics from the xt
6783: (because that is all that an xt can represent -- for more discussion of
6784: this @pxref{Tokens for Words}); by default it will have default
6785: interpretation and compilation semantics deriving from this execution
6786: semantics. However, you can change the interpretation and compilation
6787: semantics of the deferred word in the usual ways:
1.44 crook 6788:
6789: @example
1.132 anton 6790: : bar .... ; immediate
1.44 crook 6791: Defer fred immediate
6792: Defer jim
6793:
1.132 anton 6794: ' bar IS jim \ jim has default semantics
6795: ' bar IS fred \ fred is immediate
1.44 crook 6796: @end example
6797:
6798: doc-defer
1.132 anton 6799: doc-defer!
1.44 crook 6800: doc-is
1.132 anton 6801: doc-defer@
6802: doc-action-of
1.44 crook 6803: @comment TODO document these: what's defers [is]
6804: doc-defers
6805:
6806: @c Use @code{words-deferred} to see a list of deferred words.
6807:
1.132 anton 6808: Definitions of these words (except @code{defers}) in ANS Forth are
6809: provided in @file{compat/defer.fs}.
1.44 crook 6810:
6811:
1.170 pazsan 6812: @node Aliases, , Deferred Words, Defining Words
1.44 crook 6813: @subsection Aliases
6814: @cindex aliases
1.1 anton 6815:
1.44 crook 6816: The defining word @code{Alias} allows you to define a word by name that
6817: has the same behaviour as some other word. Here are two situation where
6818: this can be useful:
1.1 anton 6819:
1.44 crook 6820: @itemize @bullet
6821: @item
6822: When you want access to a word's definition from a different word list
6823: (for an example of this, see the definition of the @code{Root} word list
6824: in the Gforth source).
6825: @item
6826: When you want to create a synonym; a definition that can be known by
6827: either of two names (for example, @code{THEN} and @code{ENDIF} are
6828: aliases).
6829: @end itemize
1.1 anton 6830:
1.69 anton 6831: Like deferred words, an alias has default compilation and interpretation
6832: semantics at the beginning (not the modifications of the other word),
6833: but you can change them in the usual ways (@code{immediate},
6834: @code{compile-only}). For example:
1.1 anton 6835:
6836: @example
1.44 crook 6837: : foo ... ; immediate
6838:
6839: ' foo Alias bar \ bar is not an immediate word
6840: ' foo Alias fooby immediate \ fooby is an immediate word
1.1 anton 6841: @end example
6842:
1.44 crook 6843: Words that are aliases have the same xt, different headers in the
6844: dictionary, and consequently different name tokens (@pxref{Tokens for
6845: Words}) and possibly different immediate flags. An alias can only have
6846: default or immediate compilation semantics; you can define aliases for
6847: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
1.1 anton 6848:
1.44 crook 6849: doc-alias
1.1 anton 6850:
6851:
1.47 crook 6852: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6853: @section Interpretation and Compilation Semantics
1.26 crook 6854: @cindex semantics, interpretation and compilation
1.1 anton 6855:
1.71 anton 6856: @c !! state and ' are used without explanation
6857: @c example for immediate/compile-only? or is the tutorial enough
6858:
1.26 crook 6859: @cindex interpretation semantics
1.71 anton 6860: The @dfn{interpretation semantics} of a (named) word are what the text
1.26 crook 6861: interpreter does when it encounters the word in interpret state. It also
6862: appears in some other contexts, e.g., the execution token returned by
1.71 anton 6863: @code{' @i{word}} identifies the interpretation semantics of @i{word}
6864: (in other words, @code{' @i{word} execute} is equivalent to
1.29 crook 6865: interpret-state text interpretation of @code{@i{word}}).
1.1 anton 6866:
1.26 crook 6867: @cindex compilation semantics
1.71 anton 6868: The @dfn{compilation semantics} of a (named) word are what the text
6869: interpreter does when it encounters the word in compile state. It also
6870: appears in other contexts, e.g, @code{POSTPONE @i{word}}
6871: compiles@footnote{In standard terminology, ``appends to the current
6872: definition''.} the compilation semantics of @i{word}.
1.1 anton 6873:
1.26 crook 6874: @cindex execution semantics
6875: The standard also talks about @dfn{execution semantics}. They are used
6876: only for defining the interpretation and compilation semantics of many
6877: words. By default, the interpretation semantics of a word are to
6878: @code{execute} its execution semantics, and the compilation semantics of
6879: a word are to @code{compile,} its execution semantics.@footnote{In
6880: standard terminology: The default interpretation semantics are its
6881: execution semantics; the default compilation semantics are to append its
6882: execution semantics to the execution semantics of the current
6883: definition.}
6884:
1.71 anton 6885: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
6886: the text interpreter, ticked, or @code{postpone}d, so they have no
6887: interpretation or compilation semantics. Their behaviour is represented
6888: by their XT (@pxref{Tokens for Words}), and we call it execution
6889: semantics, too.
6890:
1.26 crook 6891: @comment TODO expand, make it co-operate with new sections on text interpreter.
6892:
6893: @cindex immediate words
6894: @cindex compile-only words
6895: You can change the semantics of the most-recently defined word:
6896:
1.44 crook 6897:
1.26 crook 6898: doc-immediate
6899: doc-compile-only
6900: doc-restrict
6901:
1.82 anton 6902: By convention, words with non-default compilation semantics (e.g.,
6903: immediate words) often have names surrounded with brackets (e.g.,
6904: @code{[']}, @pxref{Execution token}).
1.44 crook 6905:
1.26 crook 6906: Note that ticking (@code{'}) a compile-only word gives an error
6907: (``Interpreting a compile-only word'').
1.1 anton 6908:
1.47 crook 6909: @menu
1.67 anton 6910: * Combined words::
1.47 crook 6911: @end menu
1.44 crook 6912:
1.71 anton 6913:
1.48 anton 6914: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
1.44 crook 6915: @subsection Combined Words
6916: @cindex combined words
6917:
6918: Gforth allows you to define @dfn{combined words} -- words that have an
6919: arbitrary combination of interpretation and compilation semantics.
6920:
1.26 crook 6921: doc-interpret/compile:
1.1 anton 6922:
1.26 crook 6923: This feature was introduced for implementing @code{TO} and @code{S"}. I
6924: recommend that you do not define such words, as cute as they may be:
6925: they make it hard to get at both parts of the word in some contexts.
6926: E.g., assume you want to get an execution token for the compilation
6927: part. Instead, define two words, one that embodies the interpretation
6928: part, and one that embodies the compilation part. Once you have done
6929: that, you can define a combined word with @code{interpret/compile:} for
6930: the convenience of your users.
1.1 anton 6931:
1.26 crook 6932: You might try to use this feature to provide an optimizing
6933: implementation of the default compilation semantics of a word. For
6934: example, by defining:
1.1 anton 6935: @example
1.26 crook 6936: :noname
6937: foo bar ;
6938: :noname
6939: POSTPONE foo POSTPONE bar ;
1.29 crook 6940: interpret/compile: opti-foobar
1.1 anton 6941: @end example
1.26 crook 6942:
1.23 crook 6943: @noindent
1.26 crook 6944: as an optimizing version of:
6945:
1.1 anton 6946: @example
1.26 crook 6947: : foobar
6948: foo bar ;
1.1 anton 6949: @end example
6950:
1.26 crook 6951: Unfortunately, this does not work correctly with @code{[compile]},
6952: because @code{[compile]} assumes that the compilation semantics of all
6953: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
1.29 crook 6954: opti-foobar} would compile compilation semantics, whereas
6955: @code{[compile] foobar} would compile interpretation semantics.
1.1 anton 6956:
1.26 crook 6957: @cindex state-smart words (are a bad idea)
1.82 anton 6958: @anchor{state-smartness}
1.29 crook 6959: Some people try to use @dfn{state-smart} words to emulate the feature provided
1.26 crook 6960: by @code{interpret/compile:} (words are state-smart if they check
6961: @code{STATE} during execution). E.g., they would try to code
6962: @code{foobar} like this:
1.1 anton 6963:
1.26 crook 6964: @example
6965: : foobar
6966: STATE @@
6967: IF ( compilation state )
6968: POSTPONE foo POSTPONE bar
6969: ELSE
6970: foo bar
6971: ENDIF ; immediate
6972: @end example
1.1 anton 6973:
1.26 crook 6974: Although this works if @code{foobar} is only processed by the text
6975: interpreter, it does not work in other contexts (like @code{'} or
6976: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
6977: for a state-smart word, not for the interpretation semantics of the
6978: original @code{foobar}; when you execute this execution token (directly
6979: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
6980: state, the result will not be what you expected (i.e., it will not
6981: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
6982: write them@footnote{For a more detailed discussion of this topic, see
1.66 anton 6983: M. Anton Ertl,
6984: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
6985: it is Evil and How to Exorcise it}}, EuroForth '98.}!
1.1 anton 6986:
1.26 crook 6987: @cindex defining words with arbitrary semantics combinations
6988: It is also possible to write defining words that define words with
6989: arbitrary combinations of interpretation and compilation semantics. In
6990: general, they look like this:
1.1 anton 6991:
1.26 crook 6992: @example
6993: : def-word
6994: create-interpret/compile
1.29 crook 6995: @i{code1}
1.26 crook 6996: interpretation>
1.29 crook 6997: @i{code2}
1.26 crook 6998: <interpretation
6999: compilation>
1.29 crook 7000: @i{code3}
1.26 crook 7001: <compilation ;
7002: @end example
1.1 anton 7003:
1.29 crook 7004: For a @i{word} defined with @code{def-word}, the interpretation
7005: semantics are to push the address of the body of @i{word} and perform
7006: @i{code2}, and the compilation semantics are to push the address of
7007: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
1.26 crook 7008: can also be defined like this (except that the defined constants don't
7009: behave correctly when @code{[compile]}d):
1.1 anton 7010:
1.26 crook 7011: @example
7012: : constant ( n "name" -- )
7013: create-interpret/compile
7014: ,
7015: interpretation> ( -- n )
7016: @@
7017: <interpretation
7018: compilation> ( compilation. -- ; run-time. -- n )
7019: @@ postpone literal
7020: <compilation ;
7021: @end example
1.1 anton 7022:
1.44 crook 7023:
1.26 crook 7024: doc-create-interpret/compile
7025: doc-interpretation>
7026: doc-<interpretation
7027: doc-compilation>
7028: doc-<compilation
1.1 anton 7029:
1.44 crook 7030:
1.29 crook 7031: Words defined with @code{interpret/compile:} and
1.26 crook 7032: @code{create-interpret/compile} have an extended header structure that
7033: differs from other words; however, unless you try to access them with
7034: plain address arithmetic, you should not notice this. Words for
7035: accessing the header structure usually know how to deal with this; e.g.,
1.29 crook 7036: @code{'} @i{word} @code{>body} also gives you the body of a word created
7037: with @code{create-interpret/compile}.
1.1 anton 7038:
1.44 crook 7039:
1.47 crook 7040: @c -------------------------------------------------------------
1.81 anton 7041: @node Tokens for Words, Compiling words, Interpretation and Compilation Semantics, Words
1.47 crook 7042: @section Tokens for Words
7043: @cindex tokens for words
7044:
7045: This section describes the creation and use of tokens that represent
7046: words.
7047:
1.71 anton 7048: @menu
7049: * Execution token:: represents execution/interpretation semantics
7050: * Compilation token:: represents compilation semantics
7051: * Name token:: represents named words
7052: @end menu
1.47 crook 7053:
1.71 anton 7054: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
7055: @subsection Execution token
1.47 crook 7056:
7057: @cindex xt
7058: @cindex execution token
1.71 anton 7059: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
7060: You can use @code{execute} to invoke this behaviour.
1.47 crook 7061:
1.71 anton 7062: @cindex tick (')
7063: You can use @code{'} to get an execution token that represents the
7064: interpretation semantics of a named word:
1.47 crook 7065:
7066: @example
1.97 anton 7067: 5 ' . ( n xt )
7068: execute ( ) \ execute the xt (i.e., ".")
1.71 anton 7069: @end example
1.47 crook 7070:
1.71 anton 7071: doc-'
7072:
7073: @code{'} parses at run-time; there is also a word @code{[']} that parses
7074: when it is compiled, and compiles the resulting XT:
7075:
7076: @example
7077: : foo ['] . execute ;
7078: 5 foo
7079: : bar ' execute ; \ by contrast,
7080: 5 bar . \ ' parses "." when bar executes
7081: @end example
7082:
7083: doc-[']
7084:
7085: If you want the execution token of @i{word}, write @code{['] @i{word}}
7086: in compiled code and @code{' @i{word}} in interpreted code. Gforth's
7087: @code{'} and @code{[']} behave somewhat unusually by complaining about
7088: compile-only words (because these words have no interpretation
7089: semantics). You might get what you want by using @code{COMP' @i{word}
7090: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
7091: token}).
7092:
1.116 anton 7093: Another way to get an XT is @code{:noname} or @code{latestxt}
1.71 anton 7094: (@pxref{Anonymous Definitions}). For anonymous words this gives an xt
7095: for the only behaviour the word has (the execution semantics). For
1.116 anton 7096: named words, @code{latestxt} produces an XT for the same behaviour it
1.71 anton 7097: would produce if the word was defined anonymously.
7098:
7099: @example
7100: :noname ." hello" ;
7101: execute
1.47 crook 7102: @end example
7103:
1.71 anton 7104: An XT occupies one cell and can be manipulated like any other cell.
7105:
1.47 crook 7106: @cindex code field address
7107: @cindex CFA
1.71 anton 7108: In ANS Forth the XT is just an abstract data type (i.e., defined by the
7109: operations that produce or consume it). For old hands: In Gforth, the
7110: XT is implemented as a code field address (CFA).
7111:
7112: doc-execute
7113: doc-perform
7114:
7115: @node Compilation token, Name token, Execution token, Tokens for Words
7116: @subsection Compilation token
1.47 crook 7117:
7118: @cindex compilation token
1.71 anton 7119: @cindex CT (compilation token)
7120: Gforth represents the compilation semantics of a named word by a
1.47 crook 7121: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
7122: @i{xt} is an execution token. The compilation semantics represented by
7123: the compilation token can be performed with @code{execute}, which
7124: consumes the whole compilation token, with an additional stack effect
7125: determined by the represented compilation semantics.
7126:
7127: At present, the @i{w} part of a compilation token is an execution token,
7128: and the @i{xt} part represents either @code{execute} or
7129: @code{compile,}@footnote{Depending upon the compilation semantics of the
7130: word. If the word has default compilation semantics, the @i{xt} will
7131: represent @code{compile,}. Otherwise (e.g., for immediate words), the
7132: @i{xt} will represent @code{execute}.}. However, don't rely on that
7133: knowledge, unless necessary; future versions of Gforth may introduce
7134: unusual compilation tokens (e.g., a compilation token that represents
7135: the compilation semantics of a literal).
7136:
1.71 anton 7137: You can perform the compilation semantics represented by the compilation
7138: token with @code{execute}. You can compile the compilation semantics
7139: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
7140: equivalent to @code{postpone @i{word}}.
7141:
7142: doc-[comp']
7143: doc-comp'
7144: doc-postpone,
7145:
7146: @node Name token, , Compilation token, Tokens for Words
7147: @subsection Name token
1.47 crook 7148:
7149: @cindex name token
1.116 anton 7150: Gforth represents named words by the @dfn{name token}, (@i{nt}). Name
7151: token is an abstract data type that occurs as argument or result of the
7152: words below.
7153:
7154: @c !! put this elswhere?
1.47 crook 7155: @cindex name field address
7156: @cindex NFA
1.116 anton 7157: The closest thing to the nt in older Forth systems is the name field
7158: address (NFA), but there are significant differences: in older Forth
7159: systems each word had a unique NFA, LFA, CFA and PFA (in this order, or
7160: LFA, NFA, CFA, PFA) and there were words for getting from one to the
7161: next. In contrast, in Gforth 0@dots{}n nts correspond to one xt; there
7162: is a link field in the structure identified by the name token, but
7163: searching usually uses a hash table external to these structures; the
7164: name in Gforth has a cell-wide count-and-flags field, and the nt is not
7165: implemented as the address of that count field.
1.47 crook 7166:
7167: doc-find-name
1.116 anton 7168: doc-latest
7169: doc->name
1.47 crook 7170: doc-name>int
7171: doc-name?int
7172: doc-name>comp
7173: doc-name>string
1.109 anton 7174: doc-id.
7175: doc-.name
7176: doc-.id
1.47 crook 7177:
1.81 anton 7178: @c ----------------------------------------------------------
7179: @node Compiling words, The Text Interpreter, Tokens for Words, Words
7180: @section Compiling words
7181: @cindex compiling words
7182: @cindex macros
7183:
7184: In contrast to most other languages, Forth has no strict boundary
1.82 anton 7185: between compilation and run-time. E.g., you can run arbitrary code
7186: between defining words (or for computing data used by defining words
7187: like @code{constant}). Moreover, @code{Immediate} (@pxref{Interpretation
7188: and Compilation Semantics} and @code{[}...@code{]} (see below) allow
7189: running arbitrary code while compiling a colon definition (exception:
7190: you must not allot dictionary space).
7191:
7192: @menu
7193: * Literals:: Compiling data values
7194: * Macros:: Compiling words
7195: @end menu
7196:
7197: @node Literals, Macros, Compiling words, Compiling words
7198: @subsection Literals
7199: @cindex Literals
7200:
7201: The simplest and most frequent example is to compute a literal during
7202: compilation. E.g., the following definition prints an array of strings,
7203: one string per line:
7204:
7205: @example
7206: : .strings ( addr u -- ) \ gforth
7207: 2* cells bounds U+DO
7208: cr i 2@@ type
7209: 2 cells +LOOP ;
7210: @end example
1.81 anton 7211:
1.82 anton 7212: With a simple-minded compiler like Gforth's, this computes @code{2
7213: cells} on every loop iteration. You can compute this value once and for
7214: all at compile time and compile it into the definition like this:
7215:
7216: @example
7217: : .strings ( addr u -- ) \ gforth
7218: 2* cells bounds U+DO
7219: cr i 2@@ type
7220: [ 2 cells ] literal +LOOP ;
7221: @end example
7222:
7223: @code{[} switches the text interpreter to interpret state (you will get
7224: an @code{ok} prompt if you type this example interactively and insert a
7225: newline between @code{[} and @code{]}), so it performs the
7226: interpretation semantics of @code{2 cells}; this computes a number.
7227: @code{]} switches the text interpreter back into compile state. It then
7228: performs @code{Literal}'s compilation semantics, which are to compile
7229: this number into the current word. You can decompile the word with
7230: @code{see .strings} to see the effect on the compiled code.
1.81 anton 7231:
1.82 anton 7232: You can also optimize the @code{2* cells} into @code{[ 2 cells ] literal
7233: *} in this way.
1.81 anton 7234:
1.82 anton 7235: doc-[
7236: doc-]
1.81 anton 7237: doc-literal
7238: doc-]L
1.82 anton 7239:
7240: There are also words for compiling other data types than single cells as
7241: literals:
7242:
1.81 anton 7243: doc-2literal
7244: doc-fliteral
1.82 anton 7245: doc-sliteral
7246:
7247: @cindex colon-sys, passing data across @code{:}
7248: @cindex @code{:}, passing data across
7249: You might be tempted to pass data from outside a colon definition to the
7250: inside on the data stack. This does not work, because @code{:} puhes a
7251: colon-sys, making stuff below unaccessible. E.g., this does not work:
7252:
7253: @example
7254: 5 : foo literal ; \ error: "unstructured"
7255: @end example
7256:
7257: Instead, you have to pass the value in some other way, e.g., through a
7258: variable:
7259:
7260: @example
7261: variable temp
7262: 5 temp !
7263: : foo [ temp @@ ] literal ;
7264: @end example
7265:
7266:
7267: @node Macros, , Literals, Compiling words
7268: @subsection Macros
7269: @cindex Macros
7270: @cindex compiling compilation semantics
7271:
7272: @code{Literal} and friends compile data values into the current
7273: definition. You can also write words that compile other words into the
7274: current definition. E.g.,
7275:
7276: @example
7277: : compile-+ ( -- ) \ compiled code: ( n1 n2 -- n )
7278: POSTPONE + ;
7279:
7280: : foo ( n1 n2 -- n )
7281: [ compile-+ ] ;
7282: 1 2 foo .
7283: @end example
7284:
7285: This is equivalent to @code{: foo + ;} (@code{see foo} to check this).
7286: What happens in this example? @code{Postpone} compiles the compilation
7287: semantics of @code{+} into @code{compile-+}; later the text interpreter
7288: executes @code{compile-+} and thus the compilation semantics of +, which
7289: compile (the execution semantics of) @code{+} into
7290: @code{foo}.@footnote{A recent RFI answer requires that compiling words
7291: should only be executed in compile state, so this example is not
7292: guaranteed to work on all standard systems, but on any decent system it
7293: will work.}
7294:
7295: doc-postpone
7296: doc-[compile]
7297:
7298: Compiling words like @code{compile-+} are usually immediate (or similar)
7299: so you do not have to switch to interpret state to execute them;
7300: mopifying the last example accordingly produces:
7301:
7302: @example
7303: : [compile-+] ( compilation: --; interpretation: -- )
7304: \ compiled code: ( n1 n2 -- n )
7305: POSTPONE + ; immediate
7306:
7307: : foo ( n1 n2 -- n )
7308: [compile-+] ;
7309: 1 2 foo .
7310: @end example
7311:
7312: Immediate compiling words are similar to macros in other languages (in
7313: particular, Lisp). The important differences to macros in, e.g., C are:
7314:
7315: @itemize @bullet
7316:
7317: @item
7318: You use the same language for defining and processing macros, not a
7319: separate preprocessing language and processor.
7320:
7321: @item
7322: Consequently, the full power of Forth is available in macro definitions.
7323: E.g., you can perform arbitrarily complex computations, or generate
7324: different code conditionally or in a loop (e.g., @pxref{Advanced macros
7325: Tutorial}). This power is very useful when writing a parser generators
7326: or other code-generating software.
7327:
7328: @item
7329: Macros defined using @code{postpone} etc. deal with the language at a
7330: higher level than strings; name binding happens at macro definition
7331: time, so you can avoid the pitfalls of name collisions that can happen
7332: in C macros. Of course, Forth is a liberal language and also allows to
7333: shoot yourself in the foot with text-interpreted macros like
7334:
7335: @example
7336: : [compile-+] s" +" evaluate ; immediate
7337: @end example
7338:
7339: Apart from binding the name at macro use time, using @code{evaluate}
7340: also makes your definition @code{state}-smart (@pxref{state-smartness}).
7341: @end itemize
7342:
7343: You may want the macro to compile a number into a word. The word to do
7344: it is @code{literal}, but you have to @code{postpone} it, so its
7345: compilation semantics take effect when the macro is executed, not when
7346: it is compiled:
7347:
7348: @example
7349: : [compile-5] ( -- ) \ compiled code: ( -- n )
7350: 5 POSTPONE literal ; immediate
7351:
7352: : foo [compile-5] ;
7353: foo .
7354: @end example
7355:
7356: You may want to pass parameters to a macro, that the macro should
7357: compile into the current definition. If the parameter is a number, then
7358: you can use @code{postpone literal} (similar for other values).
7359:
7360: If you want to pass a word that is to be compiled, the usual way is to
7361: pass an execution token and @code{compile,} it:
7362:
7363: @example
7364: : twice1 ( xt -- ) \ compiled code: ... -- ...
7365: dup compile, compile, ;
7366:
7367: : 2+ ( n1 -- n2 )
7368: [ ' 1+ twice1 ] ;
7369: @end example
7370:
7371: doc-compile,
7372:
7373: An alternative available in Gforth, that allows you to pass compile-only
7374: words as parameters is to use the compilation token (@pxref{Compilation
7375: token}). The same example in this technique:
7376:
7377: @example
7378: : twice ( ... ct -- ... ) \ compiled code: ... -- ...
7379: 2dup 2>r execute 2r> execute ;
7380:
7381: : 2+ ( n1 -- n2 )
7382: [ comp' 1+ twice ] ;
7383: @end example
7384:
7385: In the example above @code{2>r} and @code{2r>} ensure that @code{twice}
7386: works even if the executed compilation semantics has an effect on the
7387: data stack.
7388:
7389: You can also define complete definitions with these words; this provides
7390: an alternative to using @code{does>} (@pxref{User-defined Defining
7391: Words}). E.g., instead of
7392:
7393: @example
7394: : curry+ ( n1 "name" -- )
7395: CREATE ,
7396: DOES> ( n2 -- n1+n2 )
7397: @@ + ;
7398: @end example
7399:
7400: you could define
7401:
7402: @example
7403: : curry+ ( n1 "name" -- )
7404: \ name execution: ( n2 -- n1+n2 )
7405: >r : r> POSTPONE literal POSTPONE + POSTPONE ; ;
1.81 anton 7406:
1.82 anton 7407: -3 curry+ 3-
7408: see 3-
7409: @end example
1.81 anton 7410:
1.82 anton 7411: The sequence @code{>r : r>} is necessary, because @code{:} puts a
7412: colon-sys on the data stack that makes everything below it unaccessible.
1.81 anton 7413:
1.82 anton 7414: This way of writing defining words is sometimes more, sometimes less
7415: convenient than using @code{does>} (@pxref{Advanced does> usage
7416: example}). One advantage of this method is that it can be optimized
7417: better, because the compiler knows that the value compiled with
7418: @code{literal} is fixed, whereas the data associated with a
7419: @code{create}d word can be changed.
1.47 crook 7420:
1.26 crook 7421: @c ----------------------------------------------------------
1.111 anton 7422: @node The Text Interpreter, The Input Stream, Compiling words, Words
1.26 crook 7423: @section The Text Interpreter
7424: @cindex interpreter - outer
7425: @cindex text interpreter
7426: @cindex outer interpreter
1.1 anton 7427:
1.34 anton 7428: @c Should we really describe all these ugly details? IMO the text
7429: @c interpreter should be much cleaner, but that may not be possible within
7430: @c ANS Forth. - anton
1.44 crook 7431: @c nac-> I wanted to explain how it works to show how you can exploit
7432: @c it in your own programs. When I was writing a cross-compiler, figuring out
7433: @c some of these gory details was very helpful to me. None of the textbooks
7434: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
7435: @c seems to positively avoid going into too much detail for some of
7436: @c the internals.
1.34 anton 7437:
1.71 anton 7438: @c anton: ok. I wonder, though, if this is the right place; for some stuff
7439: @c it is; for the ugly details, I would prefer another place. I wonder
7440: @c whether we should have a chapter before "Words" that describes some
7441: @c basic concepts referred to in words, and a chapter after "Words" that
7442: @c describes implementation details.
7443:
1.29 crook 7444: The text interpreter@footnote{This is an expanded version of the
7445: material in @ref{Introducing the Text Interpreter}.} is an endless loop
1.34 anton 7446: that processes input from the current input device. It is also called
7447: the outer interpreter, in contrast to the inner interpreter
7448: (@pxref{Engine}) which executes the compiled Forth code on interpretive
7449: implementations.
1.27 crook 7450:
1.29 crook 7451: @cindex interpret state
7452: @cindex compile state
7453: The text interpreter operates in one of two states: @dfn{interpret
7454: state} and @dfn{compile state}. The current state is defined by the
1.71 anton 7455: aptly-named variable @code{state}.
1.29 crook 7456:
7457: This section starts by describing how the text interpreter behaves when
7458: it is in interpret state, processing input from the user input device --
7459: the keyboard. This is the mode that a Forth system is in after it starts
7460: up.
7461:
7462: @cindex input buffer
7463: @cindex terminal input buffer
7464: The text interpreter works from an area of memory called the @dfn{input
7465: buffer}@footnote{When the text interpreter is processing input from the
7466: keyboard, this area of memory is called the @dfn{terminal input buffer}
7467: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7468: @code{#TIB}.}, which stores your keyboard input when you press the
1.30 anton 7469: @key{RET} key. Starting at the beginning of the input buffer, it skips
1.29 crook 7470: leading spaces (called @dfn{delimiters}) then parses a string (a
7471: sequence of non-space characters) until it reaches either a space
7472: character or the end of the buffer. Having parsed a string, it makes two
7473: attempts to process it:
1.27 crook 7474:
1.29 crook 7475: @cindex dictionary
1.27 crook 7476: @itemize @bullet
7477: @item
1.29 crook 7478: It looks for the string in a @dfn{dictionary} of definitions. If the
7479: string is found, the string names a @dfn{definition} (also known as a
7480: @dfn{word}) and the dictionary search returns information that allows
7481: the text interpreter to perform the word's @dfn{interpretation
7482: semantics}. In most cases, this simply means that the word will be
7483: executed.
1.27 crook 7484: @item
7485: If the string is not found in the dictionary, the text interpreter
1.29 crook 7486: attempts to treat it as a number, using the rules described in
7487: @ref{Number Conversion}. If the string represents a legal number in the
7488: current radix, the number is pushed onto a parameter stack (the data
7489: stack for integers, the floating-point stack for floating-point
7490: numbers).
7491: @end itemize
7492:
7493: If both attempts fail, or if the word is found in the dictionary but has
7494: no interpretation semantics@footnote{This happens if the word was
7495: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7496: remainder of the input buffer, issues an error message and waits for
7497: more input. If one of the attempts succeeds, the text interpreter
7498: repeats the parsing process until the whole of the input buffer has been
7499: processed, at which point it prints the status message ``@code{ ok}''
7500: and waits for more input.
7501:
1.71 anton 7502: @c anton: this should be in the input stream subsection (or below it)
7503:
1.29 crook 7504: @cindex parse area
7505: The text interpreter keeps track of its position in the input buffer by
7506: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7507: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7508: of the input buffer. The region from offset @code{>IN @@} to the end of
7509: the input buffer is called the @dfn{parse area}@footnote{In other words,
7510: the text interpreter processes the contents of the input buffer by
7511: parsing strings from the parse area until the parse area is empty.}.
7512: This example shows how @code{>IN} changes as the text interpreter parses
7513: the input buffer:
7514:
7515: @example
7516: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7517: CR ." ->" TYPE ." <-" ; IMMEDIATE
7518:
7519: 1 2 3 remaining + remaining .
7520:
7521: : foo 1 2 3 remaining SWAP remaining ;
7522: @end example
7523:
7524: @noindent
7525: The result is:
7526:
7527: @example
7528: ->+ remaining .<-
7529: ->.<-5 ok
7530:
7531: ->SWAP remaining ;-<
7532: ->;<- ok
7533: @end example
7534:
7535: @cindex parsing words
7536: The value of @code{>IN} can also be modified by a word in the input
7537: buffer that is executed by the text interpreter. This means that a word
7538: can ``trick'' the text interpreter into either skipping a section of the
7539: input buffer@footnote{This is how parsing words work.} or into parsing a
7540: section twice. For example:
1.27 crook 7541:
1.29 crook 7542: @example
1.71 anton 7543: : lat ." <<foo>>" ;
7544: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
1.29 crook 7545: @end example
7546:
7547: @noindent
7548: When @code{flat} is executed, this output is produced@footnote{Exercise
7549: for the reader: what would happen if the @code{3} were replaced with
7550: @code{4}?}:
7551:
7552: @example
1.71 anton 7553: <<bar>><<foo>>
1.29 crook 7554: @end example
7555:
1.71 anton 7556: This technique can be used to work around some of the interoperability
7557: problems of parsing words. Of course, it's better to avoid parsing
7558: words where possible.
7559:
1.29 crook 7560: @noindent
7561: Two important notes about the behaviour of the text interpreter:
1.27 crook 7562:
7563: @itemize @bullet
7564: @item
7565: It processes each input string to completion before parsing additional
1.29 crook 7566: characters from the input buffer.
7567: @item
7568: It treats the input buffer as a read-only region (and so must your code).
7569: @end itemize
7570:
7571: @noindent
7572: When the text interpreter is in compile state, its behaviour changes in
7573: these ways:
7574:
7575: @itemize @bullet
7576: @item
7577: If a parsed string is found in the dictionary, the text interpreter will
7578: perform the word's @dfn{compilation semantics}. In most cases, this
7579: simply means that the execution semantics of the word will be appended
7580: to the current definition.
1.27 crook 7581: @item
1.29 crook 7582: When a number is encountered, it is compiled into the current definition
7583: (as a literal) rather than being pushed onto a parameter stack.
7584: @item
7585: If an error occurs, @code{state} is modified to put the text interpreter
7586: back into interpret state.
7587: @item
7588: Each time a line is entered from the keyboard, Gforth prints
7589: ``@code{ compiled}'' rather than `` @code{ok}''.
7590: @end itemize
7591:
7592: @cindex text interpreter - input sources
7593: When the text interpreter is using an input device other than the
7594: keyboard, its behaviour changes in these ways:
7595:
7596: @itemize @bullet
7597: @item
7598: When the parse area is empty, the text interpreter attempts to refill
7599: the input buffer from the input source. When the input source is
1.71 anton 7600: exhausted, the input source is set back to the previous input source.
1.29 crook 7601: @item
7602: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7603: time the parse area is emptied.
7604: @item
7605: If an error occurs, the input source is set back to the user input
7606: device.
1.27 crook 7607: @end itemize
1.21 crook 7608:
1.49 anton 7609: You can read about this in more detail in @ref{Input Sources}.
1.44 crook 7610:
1.26 crook 7611: doc->in
1.27 crook 7612: doc-source
7613:
1.26 crook 7614: doc-tib
7615: doc-#tib
1.1 anton 7616:
1.44 crook 7617:
1.26 crook 7618: @menu
1.67 anton 7619: * Input Sources::
7620: * Number Conversion::
7621: * Interpret/Compile states::
7622: * Interpreter Directives::
1.26 crook 7623: @end menu
1.1 anton 7624:
1.29 crook 7625: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7626: @subsection Input Sources
7627: @cindex input sources
7628: @cindex text interpreter - input sources
7629:
1.44 crook 7630: By default, the text interpreter processes input from the user input
1.29 crook 7631: device (the keyboard) when Forth starts up. The text interpreter can
7632: process input from any of these sources:
7633:
7634: @itemize @bullet
7635: @item
7636: The user input device -- the keyboard.
7637: @item
7638: A file, using the words described in @ref{Forth source files}.
7639: @item
7640: A block, using the words described in @ref{Blocks}.
7641: @item
7642: A text string, using @code{evaluate}.
7643: @end itemize
7644:
7645: A program can identify the current input device from the values of
7646: @code{source-id} and @code{blk}.
7647:
1.44 crook 7648:
1.29 crook 7649: doc-source-id
7650: doc-blk
7651:
7652: doc-save-input
7653: doc-restore-input
7654:
7655: doc-evaluate
1.111 anton 7656: doc-query
1.1 anton 7657:
1.29 crook 7658:
1.44 crook 7659:
1.29 crook 7660: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
1.26 crook 7661: @subsection Number Conversion
7662: @cindex number conversion
7663: @cindex double-cell numbers, input format
7664: @cindex input format for double-cell numbers
7665: @cindex single-cell numbers, input format
7666: @cindex input format for single-cell numbers
7667: @cindex floating-point numbers, input format
7668: @cindex input format for floating-point numbers
1.1 anton 7669:
1.29 crook 7670: This section describes the rules that the text interpreter uses when it
7671: tries to convert a string into a number.
1.1 anton 7672:
1.26 crook 7673: Let <digit> represent any character that is a legal digit in the current
1.29 crook 7674: number base@footnote{For example, 0-9 when the number base is decimal or
7675: 0-9, A-F when the number base is hexadecimal.}.
1.1 anton 7676:
1.26 crook 7677: Let <decimal digit> represent any character in the range 0-9.
1.1 anton 7678:
1.29 crook 7679: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7680: in the braces (@i{a} or @i{b} or neither).
1.1 anton 7681:
1.26 crook 7682: Let * represent any number of instances of the previous character
7683: (including none).
1.1 anton 7684:
1.26 crook 7685: Let any other character represent itself.
1.1 anton 7686:
1.29 crook 7687: @noindent
1.26 crook 7688: Now, the conversion rules are:
1.21 crook 7689:
1.26 crook 7690: @itemize @bullet
7691: @item
7692: A string of the form <digit><digit>* is treated as a single-precision
1.29 crook 7693: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
1.26 crook 7694: @item
7695: A string of the form -<digit><digit>* is treated as a single-precision
1.29 crook 7696: (cell-sized) negative integer, and is represented using 2's-complement
1.26 crook 7697: arithmetic. Examples are -45 -5681 -0
7698: @item
7699: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
1.29 crook 7700: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7701: (all three of these represent the same number).
1.26 crook 7702: @item
7703: A string of the form -<digit><digit>*.<digit>* is treated as a
1.29 crook 7704: double-precision (double-cell-sized) negative integer, and is
1.26 crook 7705: represented using 2's-complement arithmetic. Examples are -3465. -3.465
1.29 crook 7706: -34.65 (all three of these represent the same number).
1.26 crook 7707: @item
1.29 crook 7708: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7709: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
1.35 anton 7710: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
1.29 crook 7711: number) +12.E-4
1.26 crook 7712: @end itemize
1.1 anton 7713:
1.174 anton 7714: By default, the number base used for integer number conversion is
7715: given by the contents of the variable @code{base}. Note that a lot of
1.35 anton 7716: confusion can result from unexpected values of @code{base}. If you
1.174 anton 7717: change @code{base} anywhere, make sure to save the old value and
7718: restore it afterwards; better yet, use @code{base-execute}, which does
7719: this for you. In general I recommend keeping @code{base} decimal, and
1.35 anton 7720: using the prefixes described below for the popular non-decimal bases.
1.1 anton 7721:
1.29 crook 7722: doc-dpl
1.174 anton 7723: doc-base-execute
1.26 crook 7724: doc-base
7725: doc-hex
7726: doc-decimal
1.1 anton 7727:
1.26 crook 7728: @cindex '-prefix for character strings
7729: @cindex &-prefix for decimal numbers
1.133 anton 7730: @cindex #-prefix for decimal numbers
1.26 crook 7731: @cindex %-prefix for binary numbers
7732: @cindex $-prefix for hexadecimal numbers
1.133 anton 7733: @cindex 0x-prefix for hexadecimal numbers
1.35 anton 7734: Gforth allows you to override the value of @code{base} by using a
1.29 crook 7735: prefix@footnote{Some Forth implementations provide a similar scheme by
7736: implementing @code{$} etc. as parsing words that process the subsequent
7737: number in the input stream and push it onto the stack. For example, see
7738: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7739: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7740: is required between the prefix and the number.} before the first digit
1.133 anton 7741: of an (integer) number. The following prefixes are supported:
1.1 anton 7742:
1.26 crook 7743: @itemize @bullet
7744: @item
1.35 anton 7745: @code{&} -- decimal
1.26 crook 7746: @item
1.133 anton 7747: @code{#} -- decimal
7748: @item
1.35 anton 7749: @code{%} -- binary
1.26 crook 7750: @item
1.35 anton 7751: @code{$} -- hexadecimal
1.26 crook 7752: @item
1.133 anton 7753: @code{0x} -- hexadecimal, if base<33.
7754: @item
7755: @code{'} -- numeric value (e.g., ASCII code) of next character; an
7756: optional @code{'} may be present after the character.
1.26 crook 7757: @end itemize
1.1 anton 7758:
1.26 crook 7759: Here are some examples, with the equivalent decimal number shown after
7760: in braces:
1.1 anton 7761:
1.26 crook 7762: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
1.133 anton 7763: 'A (65),
7764: -'a' (-97),
1.26 crook 7765: &905 (905), $abc (2478), $ABC (2478).
1.1 anton 7766:
1.26 crook 7767: @cindex number conversion - traps for the unwary
1.29 crook 7768: @noindent
1.26 crook 7769: Number conversion has a number of traps for the unwary:
1.1 anton 7770:
1.26 crook 7771: @itemize @bullet
7772: @item
7773: You cannot determine the current number base using the code sequence
1.35 anton 7774: @code{base @@ .} -- the number base is always 10 in the current number
7775: base. Instead, use something like @code{base @@ dec.}
1.26 crook 7776: @item
7777: If the number base is set to a value greater than 14 (for example,
7778: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7779: it to be intepreted as either a single-precision integer or a
7780: floating-point number (Gforth treats it as an integer). The ambiguity
7781: can be resolved by explicitly stating the sign of the mantissa and/or
7782: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7783: ambiguity arises; either representation will be treated as a
7784: floating-point number.
7785: @item
1.29 crook 7786: There is a word @code{bin} but it does @i{not} set the number base!
1.26 crook 7787: It is used to specify file types.
7788: @item
1.72 anton 7789: ANS Forth requires the @code{.} of a double-precision number to be the
7790: final character in the string. Gforth allows the @code{.} to be
7791: anywhere after the first digit.
1.26 crook 7792: @item
7793: The number conversion process does not check for overflow.
7794: @item
1.72 anton 7795: In an ANS Forth program @code{base} is required to be decimal when
7796: converting floating-point numbers. In Gforth, number conversion to
7797: floating-point numbers always uses base &10, irrespective of the value
7798: of @code{base}.
1.26 crook 7799: @end itemize
1.1 anton 7800:
1.49 anton 7801: You can read numbers into your programs with the words described in
1.181 ! anton 7802: @ref{Line input and conversion}.
1.1 anton 7803:
1.82 anton 7804: @node Interpret/Compile states, Interpreter Directives, Number Conversion, The Text Interpreter
1.26 crook 7805: @subsection Interpret/Compile states
7806: @cindex Interpret/Compile states
1.1 anton 7807:
1.29 crook 7808: A standard program is not permitted to change @code{state}
7809: explicitly. However, it can change @code{state} implicitly, using the
7810: words @code{[} and @code{]}. When @code{[} is executed it switches
7811: @code{state} to interpret state, and therefore the text interpreter
7812: starts interpreting. When @code{]} is executed it switches @code{state}
7813: to compile state and therefore the text interpreter starts
1.44 crook 7814: compiling. The most common usage for these words is for switching into
7815: interpret state and back from within a colon definition; this technique
1.49 anton 7816: can be used to compile a literal (for an example, @pxref{Literals}) or
7817: for conditional compilation (for an example, @pxref{Interpreter
7818: Directives}).
1.44 crook 7819:
1.35 anton 7820:
7821: @c This is a bad example: It's non-standard, and it's not necessary.
7822: @c However, I can't think of a good example for switching into compile
7823: @c state when there is no current word (@code{state}-smart words are not a
7824: @c good reason). So maybe we should use an example for switching into
7825: @c interpret @code{state} in a colon def. - anton
1.44 crook 7826: @c nac-> I agree. I started out by putting in the example, then realised
7827: @c that it was non-ANS, so wrote more words around it. I hope this
7828: @c re-written version is acceptable to you. I do want to keep the example
7829: @c as it is helpful for showing what is and what is not portable, particularly
7830: @c where it outlaws a style in common use.
7831:
1.72 anton 7832: @c anton: it's more important to show what's portable. After we have done
1.83 anton 7833: @c that, we can also show what's not. In any case, I have written a
7834: @c section Compiling Words which also deals with [ ].
1.35 anton 7835:
1.95 anton 7836: @c !! The following example does not work in Gforth 0.5.9 or later.
1.29 crook 7837:
1.95 anton 7838: @c @code{[} and @code{]} also give you the ability to switch into compile
7839: @c state and back, but we cannot think of any useful Standard application
7840: @c for this ability. Pre-ANS Forth textbooks have examples like this:
7841:
7842: @c @example
7843: @c : AA ." this is A" ;
7844: @c : BB ." this is B" ;
7845: @c : CC ." this is C" ;
7846:
7847: @c create table ] aa bb cc [
7848:
7849: @c : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7850: @c cells table + @@ execute ;
7851: @c @end example
7852:
7853: @c This example builds a jump table; @code{0 go} will display ``@code{this
7854: @c is A}''. Using @code{[} and @code{]} in this example is equivalent to
7855: @c defining @code{table} like this:
7856:
7857: @c @example
7858: @c create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
7859: @c @end example
7860:
7861: @c The problem with this code is that the definition of @code{table} is not
7862: @c portable -- it @i{compile}s execution tokens into code space. Whilst it
7863: @c @i{may} work on systems where code space and data space co-incide, the
7864: @c Standard only allows data space to be assigned for a @code{CREATE}d
7865: @c word. In addition, the Standard only allows @code{@@} to access data
7866: @c space, whilst this example is using it to access code space. The only
7867: @c portable, Standard way to build this table is to build it in data space,
7868: @c like this:
7869:
7870: @c @example
7871: @c create table ' aa , ' bb , ' cc ,
7872: @c @end example
1.29 crook 7873:
1.95 anton 7874: @c doc-state
1.44 crook 7875:
1.29 crook 7876:
1.82 anton 7877: @node Interpreter Directives, , Interpret/Compile states, The Text Interpreter
1.26 crook 7878: @subsection Interpreter Directives
7879: @cindex interpreter directives
1.72 anton 7880: @cindex conditional compilation
1.1 anton 7881:
1.29 crook 7882: These words are usually used in interpret state; typically to control
7883: which parts of a source file are processed by the text
1.26 crook 7884: interpreter. There are only a few ANS Forth Standard words, but Gforth
7885: supplements these with a rich set of immediate control structure words
7886: to compensate for the fact that the non-immediate versions can only be
1.29 crook 7887: used in compile state (@pxref{Control Structures}). Typical usages:
7888:
7889: @example
1.72 anton 7890: FALSE Constant HAVE-ASSEMBLER
1.29 crook 7891: .
7892: .
1.72 anton 7893: HAVE-ASSEMBLER [IF]
1.29 crook 7894: : ASSEMBLER-FEATURE
7895: ...
7896: ;
7897: [ENDIF]
7898: .
7899: .
7900: : SEE
7901: ... \ general-purpose SEE code
1.72 anton 7902: [ HAVE-ASSEMBLER [IF] ]
1.29 crook 7903: ... \ assembler-specific SEE code
7904: [ [ENDIF] ]
7905: ;
7906: @end example
1.1 anton 7907:
1.44 crook 7908:
1.26 crook 7909: doc-[IF]
7910: doc-[ELSE]
7911: doc-[THEN]
7912: doc-[ENDIF]
1.1 anton 7913:
1.26 crook 7914: doc-[IFDEF]
7915: doc-[IFUNDEF]
1.1 anton 7916:
1.26 crook 7917: doc-[?DO]
7918: doc-[DO]
7919: doc-[FOR]
7920: doc-[LOOP]
7921: doc-[+LOOP]
7922: doc-[NEXT]
1.1 anton 7923:
1.26 crook 7924: doc-[BEGIN]
7925: doc-[UNTIL]
7926: doc-[AGAIN]
7927: doc-[WHILE]
7928: doc-[REPEAT]
1.1 anton 7929:
1.27 crook 7930:
1.26 crook 7931: @c -------------------------------------------------------------
1.111 anton 7932: @node The Input Stream, Word Lists, The Text Interpreter, Words
7933: @section The Input Stream
7934: @cindex input stream
7935:
7936: @c !! integrate this better with the "Text Interpreter" section
7937: The text interpreter reads from the input stream, which can come from
7938: several sources (@pxref{Input Sources}). Some words, in particular
7939: defining words, but also words like @code{'}, read parameters from the
7940: input stream instead of from the stack.
7941:
7942: Such words are called parsing words, because they parse the input
7943: stream. Parsing words are hard to use in other words, because it is
7944: hard to pass program-generated parameters through the input stream.
7945: They also usually have an unintuitive combination of interpretation and
7946: compilation semantics when implemented naively, leading to various
7947: approaches that try to produce a more intuitive behaviour
7948: (@pxref{Combined words}).
7949:
7950: It should be obvious by now that parsing words are a bad idea. If you
7951: want to implement a parsing word for convenience, also provide a factor
7952: of the word that does not parse, but takes the parameters on the stack.
7953: To implement the parsing word on top if it, you can use the following
7954: words:
7955:
7956: @c anton: these belong in the input stream section
7957: doc-parse
1.138 anton 7958: doc-parse-name
1.111 anton 7959: doc-parse-word
7960: doc-name
7961: doc-word
7962: doc-\"-parse
7963: doc-refill
7964:
7965: Conversely, if you have the bad luck (or lack of foresight) to have to
7966: deal with parsing words without having such factors, how do you pass a
7967: string that is not in the input stream to it?
7968:
7969: doc-execute-parsing
7970:
1.146 anton 7971: A definition of this word in ANS Forth is provided in
7972: @file{compat/execute-parsing.fs}.
7973:
1.111 anton 7974: If you want to run a parsing word on a file, the following word should
7975: help:
7976:
7977: doc-execute-parsing-file
7978:
7979: @c -------------------------------------------------------------
7980: @node Word Lists, Environmental Queries, The Input Stream, Words
1.26 crook 7981: @section Word Lists
7982: @cindex word lists
1.32 anton 7983: @cindex header space
1.1 anton 7984:
1.36 anton 7985: A wordlist is a list of named words; you can add new words and look up
7986: words by name (and you can remove words in a restricted way with
7987: markers). Every named (and @code{reveal}ed) word is in one wordlist.
7988:
7989: @cindex search order stack
7990: The text interpreter searches the wordlists present in the search order
7991: (a stack of wordlists), from the top to the bottom. Within each
7992: wordlist, the search starts conceptually at the newest word; i.e., if
7993: two words in a wordlist have the same name, the newer word is found.
1.1 anton 7994:
1.26 crook 7995: @cindex compilation word list
1.36 anton 7996: New words are added to the @dfn{compilation wordlist} (aka current
7997: wordlist).
1.1 anton 7998:
1.36 anton 7999: @cindex wid
8000: A word list is identified by a cell-sized word list identifier (@i{wid})
8001: in much the same way as a file is identified by a file handle. The
8002: numerical value of the wid has no (portable) meaning, and might change
8003: from session to session.
1.1 anton 8004:
1.29 crook 8005: The ANS Forth ``Search order'' word set is intended to provide a set of
8006: low-level tools that allow various different schemes to be
1.74 anton 8007: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
1.26 crook 8008: word. @file{compat/vocabulary.fs} provides an implementation in ANS
1.45 crook 8009: Forth.
1.1 anton 8010:
1.27 crook 8011: @comment TODO: locals section refers to here, saying that every word list (aka
8012: @comment vocabulary) has its own methods for searching etc. Need to document that.
1.78 anton 8013: @c anton: but better in a separate subsection on wordlist internals
1.1 anton 8014:
1.45 crook 8015: @comment TODO: document markers, reveal, tables, mappedwordlist
8016:
8017: @comment the gforthman- prefix is used to pick out the true definition of a
1.27 crook 8018: @comment word from the source files, rather than some alias.
1.44 crook 8019:
1.26 crook 8020: doc-forth-wordlist
8021: doc-definitions
8022: doc-get-current
8023: doc-set-current
8024: doc-get-order
1.45 crook 8025: doc---gforthman-set-order
1.26 crook 8026: doc-wordlist
1.30 anton 8027: doc-table
1.79 anton 8028: doc->order
1.36 anton 8029: doc-previous
1.26 crook 8030: doc-also
1.45 crook 8031: doc---gforthman-forth
1.26 crook 8032: doc-only
1.45 crook 8033: doc---gforthman-order
1.15 anton 8034:
1.26 crook 8035: doc-find
8036: doc-search-wordlist
1.15 anton 8037:
1.26 crook 8038: doc-words
8039: doc-vlist
1.44 crook 8040: @c doc-words-deferred
1.1 anton 8041:
1.74 anton 8042: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
1.26 crook 8043: doc-root
8044: doc-vocabulary
8045: doc-seal
8046: doc-vocs
8047: doc-current
8048: doc-context
1.1 anton 8049:
1.44 crook 8050:
1.26 crook 8051: @menu
1.75 anton 8052: * Vocabularies::
1.67 anton 8053: * Why use word lists?::
1.75 anton 8054: * Word list example::
1.26 crook 8055: @end menu
8056:
1.75 anton 8057: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
8058: @subsection Vocabularies
8059: @cindex Vocabularies, detailed explanation
8060:
8061: Here is an example of creating and using a new wordlist using ANS
8062: Forth words:
8063:
8064: @example
8065: wordlist constant my-new-words-wordlist
8066: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
8067:
8068: \ add it to the search order
8069: also my-new-words
8070:
8071: \ alternatively, add it to the search order and make it
8072: \ the compilation word list
8073: also my-new-words definitions
8074: \ type "order" to see the problem
8075: @end example
8076:
8077: The problem with this example is that @code{order} has no way to
8078: associate the name @code{my-new-words} with the wid of the word list (in
8079: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
8080: that has no associated name). There is no Standard way of associating a
8081: name with a wid.
8082:
8083: In Gforth, this example can be re-coded using @code{vocabulary}, which
8084: associates a name with a wid:
8085:
8086: @example
8087: vocabulary my-new-words
8088:
8089: \ add it to the search order
8090: also my-new-words
8091:
8092: \ alternatively, add it to the search order and make it
8093: \ the compilation word list
8094: my-new-words definitions
8095: \ type "order" to see that the problem is solved
8096: @end example
8097:
8098:
8099: @node Why use word lists?, Word list example, Vocabularies, Word Lists
1.26 crook 8100: @subsection Why use word lists?
8101: @cindex word lists - why use them?
8102:
1.74 anton 8103: Here are some reasons why people use wordlists:
1.26 crook 8104:
8105: @itemize @bullet
1.74 anton 8106:
8107: @c anton: Gforth's hashing implementation makes the search speed
8108: @c independent from the number of words. But it is linear with the number
8109: @c of wordlists that have to be searched, so in effect using more wordlists
8110: @c actually slows down compilation.
8111:
8112: @c @item
8113: @c To improve compilation speed by reducing the number of header space
8114: @c entries that must be searched. This is achieved by creating a new
8115: @c word list that contains all of the definitions that are used in the
8116: @c definition of a Forth system but which would not usually be used by
8117: @c programs running on that system. That word list would be on the search
8118: @c list when the Forth system was compiled but would be removed from the
8119: @c search list for normal operation. This can be a useful technique for
8120: @c low-performance systems (for example, 8-bit processors in embedded
8121: @c systems) but is unlikely to be necessary in high-performance desktop
8122: @c systems.
8123:
1.26 crook 8124: @item
8125: To prevent a set of words from being used outside the context in which
8126: they are valid. Two classic examples of this are an integrated editor
8127: (all of the edit commands are defined in a separate word list; the
8128: search order is set to the editor word list when the editor is invoked;
8129: the old search order is restored when the editor is terminated) and an
8130: integrated assembler (the op-codes for the machine are defined in a
8131: separate word list which is used when a @code{CODE} word is defined).
1.74 anton 8132:
8133: @item
8134: To organize the words of an application or library into a user-visible
8135: set (in @code{forth-wordlist} or some other common wordlist) and a set
8136: of helper words used just for the implementation (hidden in a separate
1.75 anton 8137: wordlist). This keeps @code{words}' output smaller, separates
8138: implementation and interface, and reduces the chance of name conflicts
8139: within the common wordlist.
1.74 anton 8140:
1.26 crook 8141: @item
8142: To prevent a name-space clash between multiple definitions with the same
8143: name. For example, when building a cross-compiler you might have a word
8144: @code{IF} that generates conditional code for your target system. By
8145: placing this definition in a different word list you can control whether
8146: the host system's @code{IF} or the target system's @code{IF} get used in
8147: any particular context by controlling the order of the word lists on the
8148: search order stack.
1.74 anton 8149:
1.26 crook 8150: @end itemize
1.1 anton 8151:
1.74 anton 8152: The downsides of using wordlists are:
8153:
8154: @itemize
8155:
8156: @item
8157: Debugging becomes more cumbersome.
8158:
8159: @item
8160: Name conflicts worked around with wordlists are still there, and you
8161: have to arrange the search order carefully to get the desired results;
8162: if you forget to do that, you get hard-to-find errors (as in any case
8163: where you read the code differently from the compiler; @code{see} can
1.75 anton 8164: help seeing which of several possible words the name resolves to in such
8165: cases). @code{See} displays just the name of the words, not what
8166: wordlist they belong to, so it might be misleading. Using unique names
8167: is a better approach to avoid name conflicts.
1.74 anton 8168:
8169: @item
8170: You have to explicitly undo any changes to the search order. In many
8171: cases it would be more convenient if this happened implicitly. Gforth
8172: currently does not provide such a feature, but it may do so in the
8173: future.
8174: @end itemize
8175:
8176:
1.75 anton 8177: @node Word list example, , Why use word lists?, Word Lists
8178: @subsection Word list example
8179: @cindex word lists - example
1.1 anton 8180:
1.74 anton 8181: The following example is from the
8182: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
8183: garbage collector} and uses wordlists to separate public words from
8184: helper words:
8185:
8186: @example
8187: get-current ( wid )
8188: vocabulary garbage-collector also garbage-collector definitions
8189: ... \ define helper words
8190: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
8191: ... \ define the public (i.e., API) words
8192: \ they can refer to the helper words
8193: previous \ restore original search order (helper words become invisible)
8194: @end example
8195:
1.26 crook 8196: @c -------------------------------------------------------------
8197: @node Environmental Queries, Files, Word Lists, Words
8198: @section Environmental Queries
8199: @cindex environmental queries
1.21 crook 8200:
1.26 crook 8201: ANS Forth introduced the idea of ``environmental queries'' as a way
8202: for a program running on a system to determine certain characteristics of the system.
8203: The Standard specifies a number of strings that might be recognised by a system.
1.21 crook 8204:
1.32 anton 8205: The Standard requires that the header space used for environmental queries
8206: be distinct from the header space used for definitions.
1.21 crook 8207:
1.26 crook 8208: Typically, environmental queries are supported by creating a set of
1.29 crook 8209: definitions in a word list that is @i{only} used during environmental
1.26 crook 8210: queries; that is what Gforth does. There is no Standard way of adding
8211: definitions to the set of recognised environmental queries, but any
8212: implementation that supports the loading of optional word sets must have
8213: some mechanism for doing this (after loading the word set, the
8214: associated environmental query string must return @code{true}). In
8215: Gforth, the word list used to honour environmental queries can be
8216: manipulated just like any other word list.
1.21 crook 8217:
1.44 crook 8218:
1.26 crook 8219: doc-environment?
8220: doc-environment-wordlist
1.21 crook 8221:
1.26 crook 8222: doc-gforth
8223: doc-os-class
1.21 crook 8224:
1.44 crook 8225:
1.26 crook 8226: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
8227: returning two items on the stack, querying it using @code{environment?}
8228: will return an additional item; the @code{true} flag that shows that the
8229: string was recognised.
1.21 crook 8230:
1.26 crook 8231: @comment TODO Document the standard strings or note where they are documented herein
1.21 crook 8232:
1.26 crook 8233: Here are some examples of using environmental queries:
1.21 crook 8234:
1.26 crook 8235: @example
8236: s" address-unit-bits" environment? 0=
8237: [IF]
8238: cr .( environmental attribute address-units-bits unknown... ) cr
1.75 anton 8239: [ELSE]
8240: drop \ ensure balanced stack effect
1.26 crook 8241: [THEN]
1.21 crook 8242:
1.75 anton 8243: \ this might occur in the prelude of a standard program that uses THROW
8244: s" exception" environment? [IF]
8245: 0= [IF]
8246: : throw abort" exception thrown" ;
8247: [THEN]
8248: [ELSE] \ we don't know, so make sure
8249: : throw abort" exception thrown" ;
8250: [THEN]
1.21 crook 8251:
1.26 crook 8252: s" gforth" environment? [IF] .( Gforth version ) TYPE
8253: [ELSE] .( Not Gforth..) [THEN]
1.75 anton 8254:
8255: \ a program using v*
8256: s" gforth" environment? [IF]
8257: s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
8258: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8259: >r swap 2swap swap 0e r> 0 ?DO
8260: dup f@ over + 2swap dup f@ f* f+ over + 2swap
8261: LOOP
8262: 2drop 2drop ;
8263: [THEN]
8264: [ELSE] \
8265: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8266: ...
8267: [THEN]
1.26 crook 8268: @end example
1.21 crook 8269:
1.26 crook 8270: Here is an example of adding a definition to the environment word list:
1.21 crook 8271:
1.26 crook 8272: @example
8273: get-current environment-wordlist set-current
8274: true constant block
8275: true constant block-ext
8276: set-current
8277: @end example
1.21 crook 8278:
1.26 crook 8279: You can see what definitions are in the environment word list like this:
1.21 crook 8280:
1.26 crook 8281: @example
1.79 anton 8282: environment-wordlist >order words previous
1.26 crook 8283: @end example
1.21 crook 8284:
8285:
1.26 crook 8286: @c -------------------------------------------------------------
8287: @node Files, Blocks, Environmental Queries, Words
8288: @section Files
1.28 crook 8289: @cindex files
8290: @cindex I/O - file-handling
1.21 crook 8291:
1.26 crook 8292: Gforth provides facilities for accessing files that are stored in the
8293: host operating system's file-system. Files that are processed by Gforth
8294: can be divided into two categories:
1.21 crook 8295:
1.23 crook 8296: @itemize @bullet
8297: @item
1.29 crook 8298: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
1.23 crook 8299: @item
1.29 crook 8300: Files that are processed by some other program (@dfn{general files}).
1.26 crook 8301: @end itemize
8302:
8303: @menu
1.48 anton 8304: * Forth source files::
8305: * General files::
1.167 anton 8306: * Redirection::
1.48 anton 8307: * Search Paths::
1.26 crook 8308: @end menu
8309:
8310: @c -------------------------------------------------------------
8311: @node Forth source files, General files, Files, Files
8312: @subsection Forth source files
8313: @cindex including files
8314: @cindex Forth source files
1.21 crook 8315:
1.26 crook 8316: The simplest way to interpret the contents of a file is to use one of
8317: these two formats:
1.21 crook 8318:
1.26 crook 8319: @example
8320: include mysource.fs
8321: s" mysource.fs" included
8322: @end example
1.21 crook 8323:
1.75 anton 8324: You usually want to include a file only if it is not included already
1.26 crook 8325: (by, say, another source file). In that case, you can use one of these
1.45 crook 8326: three formats:
1.21 crook 8327:
1.26 crook 8328: @example
8329: require mysource.fs
8330: needs mysource.fs
8331: s" mysource.fs" required
8332: @end example
1.21 crook 8333:
1.26 crook 8334: @cindex stack effect of included files
8335: @cindex including files, stack effect
1.45 crook 8336: It is good practice to write your source files such that interpreting them
8337: does not change the stack. Source files designed in this way can be used with
1.26 crook 8338: @code{required} and friends without complications. For example:
1.21 crook 8339:
1.26 crook 8340: @example
1.75 anton 8341: 1024 require foo.fs drop
1.26 crook 8342: @end example
1.21 crook 8343:
1.75 anton 8344: Here you want to pass the argument 1024 (e.g., a buffer size) to
8345: @file{foo.fs}. Interpreting @file{foo.fs} has the stack effect ( n -- n
8346: ), which allows its use with @code{require}. Of course with such
8347: parameters to required files, you have to ensure that the first
8348: @code{require} fits for all uses (i.e., @code{require} it early in the
8349: master load file).
1.44 crook 8350:
1.26 crook 8351: doc-include-file
8352: doc-included
1.28 crook 8353: doc-included?
1.26 crook 8354: doc-include
8355: doc-required
8356: doc-require
8357: doc-needs
1.75 anton 8358: @c doc-init-included-files @c internal
8359: doc-sourcefilename
8360: doc-sourceline#
1.44 crook 8361:
1.26 crook 8362: A definition in ANS Forth for @code{required} is provided in
8363: @file{compat/required.fs}.
1.21 crook 8364:
1.26 crook 8365: @c -------------------------------------------------------------
1.167 anton 8366: @node General files, Redirection, Forth source files, Files
1.26 crook 8367: @subsection General files
8368: @cindex general files
8369: @cindex file-handling
1.21 crook 8370:
1.75 anton 8371: Files are opened/created by name and type. The following file access
8372: methods (FAMs) are recognised:
1.44 crook 8373:
1.75 anton 8374: @cindex fam (file access method)
1.26 crook 8375: doc-r/o
8376: doc-r/w
8377: doc-w/o
8378: doc-bin
1.1 anton 8379:
1.44 crook 8380:
1.26 crook 8381: When a file is opened/created, it returns a file identifier,
1.29 crook 8382: @i{wfileid} that is used for all other file commands. All file
8383: commands also return a status value, @i{wior}, that is 0 for a
1.26 crook 8384: successful operation and an implementation-defined non-zero value in the
8385: case of an error.
1.21 crook 8386:
1.44 crook 8387:
1.26 crook 8388: doc-open-file
8389: doc-create-file
1.21 crook 8390:
1.26 crook 8391: doc-close-file
8392: doc-delete-file
8393: doc-rename-file
8394: doc-read-file
8395: doc-read-line
1.154 anton 8396: doc-key-file
8397: doc-key?-file
1.26 crook 8398: doc-write-file
8399: doc-write-line
8400: doc-emit-file
8401: doc-flush-file
1.21 crook 8402:
1.26 crook 8403: doc-file-status
8404: doc-file-position
8405: doc-reposition-file
8406: doc-file-size
8407: doc-resize-file
1.21 crook 8408:
1.93 anton 8409: doc-slurp-file
8410: doc-slurp-fid
1.112 anton 8411: doc-stdin
8412: doc-stdout
8413: doc-stderr
1.44 crook 8414:
1.26 crook 8415: @c ---------------------------------------------------------
1.167 anton 8416: @node Redirection, Search Paths, General files, Files
8417: @subsection Redirection
8418: @cindex Redirection
8419: @cindex Input Redirection
8420: @cindex Output Redirection
8421:
8422: You can redirect the output of @code{type} and @code{emit} and all the
8423: words that use them (all output words that don't have an explicit
1.174 anton 8424: target file) to an arbitrary file with the @code{outfile-execute},
8425: used like this:
1.167 anton 8426:
8427: @example
1.174 anton 8428: : some-warning ( n -- )
8429: cr ." warning# " . ;
8430:
1.167 anton 8431: : print-some-warning ( n -- )
1.174 anton 8432: ['] some-warning stderr outfile-execute ;
1.167 anton 8433: @end example
8434:
1.174 anton 8435: After @code{some-warning} is executed, the original output direction
8436: is restored; this construct is safe against exceptions. Similarly,
8437: there is @code{infile-execute} for redirecting the input of @code{key}
8438: and its users (any input word that does not take a file explicitly).
8439:
8440: doc-outfile-execute
8441: doc-infile-execute
1.167 anton 8442:
8443: If you do not want to redirect the input or output to a file, you can
8444: also make use of the fact that @code{key}, @code{emit} and @code{type}
8445: are deferred words (@pxref{Deferred Words}). However, in that case
8446: you have to worry about the restoration and the protection against
8447: exceptions yourself; also, note that for redirecting the output in
8448: this way, you have to redirect both @code{emit} and @code{type}.
8449:
8450: @c ---------------------------------------------------------
8451: @node Search Paths, , Redirection, Files
1.26 crook 8452: @subsection Search Paths
8453: @cindex path for @code{included}
8454: @cindex file search path
8455: @cindex @code{include} search path
8456: @cindex search path for files
1.21 crook 8457:
1.26 crook 8458: If you specify an absolute filename (i.e., a filename starting with
8459: @file{/} or @file{~}, or with @file{:} in the second position (as in
8460: @samp{C:...})) for @code{included} and friends, that file is included
8461: just as you would expect.
1.21 crook 8462:
1.75 anton 8463: If the filename starts with @file{./}, this refers to the directory that
8464: the present file was @code{included} from. This allows files to include
8465: other files relative to their own position (irrespective of the current
8466: working directory or the absolute position). This feature is essential
8467: for libraries consisting of several files, where a file may include
8468: other files from the library. It corresponds to @code{#include "..."}
8469: in C. If the current input source is not a file, @file{.} refers to the
8470: directory of the innermost file being included, or, if there is no file
8471: being included, to the current working directory.
8472:
8473: For relative filenames (not starting with @file{./}), Gforth uses a
8474: search path similar to Forth's search order (@pxref{Word Lists}). It
8475: tries to find the given filename in the directories present in the path,
8476: and includes the first one it finds. There are separate search paths for
8477: Forth source files and general files. If the search path contains the
8478: directory @file{.}, this refers to the directory of the current file, or
8479: the working directory, as if the file had been specified with @file{./}.
1.21 crook 8480:
1.26 crook 8481: Use @file{~+} to refer to the current working directory (as in the
8482: @code{bash}).
1.1 anton 8483:
1.75 anton 8484: @c anton: fold the following subsubsections into this subsection?
1.1 anton 8485:
1.48 anton 8486: @menu
1.75 anton 8487: * Source Search Paths::
1.48 anton 8488: * General Search Paths::
8489: @end menu
8490:
1.26 crook 8491: @c ---------------------------------------------------------
1.75 anton 8492: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
8493: @subsubsection Source Search Paths
8494: @cindex search path control, source files
1.5 anton 8495:
1.26 crook 8496: The search path is initialized when you start Gforth (@pxref{Invoking
1.75 anton 8497: Gforth}). You can display it and change it using @code{fpath} in
8498: combination with the general path handling words.
1.5 anton 8499:
1.75 anton 8500: doc-fpath
8501: @c the functionality of the following words is easily available through
8502: @c fpath and the general path words. The may go away.
8503: @c doc-.fpath
8504: @c doc-fpath+
8505: @c doc-fpath=
8506: @c doc-open-fpath-file
1.44 crook 8507:
8508: @noindent
1.26 crook 8509: Here is an example of using @code{fpath} and @code{require}:
1.5 anton 8510:
1.26 crook 8511: @example
1.75 anton 8512: fpath path= /usr/lib/forth/|./
1.26 crook 8513: require timer.fs
8514: @end example
1.5 anton 8515:
1.75 anton 8516:
1.26 crook 8517: @c ---------------------------------------------------------
1.75 anton 8518: @node General Search Paths, , Source Search Paths, Search Paths
1.26 crook 8519: @subsubsection General Search Paths
1.75 anton 8520: @cindex search path control, source files
1.5 anton 8521:
1.26 crook 8522: Your application may need to search files in several directories, like
8523: @code{included} does. To facilitate this, Gforth allows you to define
8524: and use your own search paths, by providing generic equivalents of the
8525: Forth search path words:
1.5 anton 8526:
1.75 anton 8527: doc-open-path-file
8528: doc-path-allot
8529: doc-clear-path
8530: doc-also-path
1.26 crook 8531: doc-.path
8532: doc-path+
8533: doc-path=
1.5 anton 8534:
1.75 anton 8535: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
1.44 crook 8536:
1.75 anton 8537: Here's an example of creating an empty search path:
8538: @c
1.26 crook 8539: @example
1.75 anton 8540: create mypath 500 path-allot \ maximum length 500 chars (is checked)
1.26 crook 8541: @end example
1.5 anton 8542:
1.26 crook 8543: @c -------------------------------------------------------------
8544: @node Blocks, Other I/O, Files, Words
8545: @section Blocks
1.28 crook 8546: @cindex I/O - blocks
8547: @cindex blocks
8548:
8549: When you run Gforth on a modern desk-top computer, it runs under the
8550: control of an operating system which provides certain services. One of
8551: these services is @var{file services}, which allows Forth source code
8552: and data to be stored in files and read into Gforth (@pxref{Files}).
8553:
8554: Traditionally, Forth has been an important programming language on
8555: systems where it has interfaced directly to the underlying hardware with
8556: no intervening operating system. Forth provides a mechanism, called
1.29 crook 8557: @dfn{blocks}, for accessing mass storage on such systems.
1.28 crook 8558:
8559: A block is a 1024-byte data area, which can be used to hold data or
8560: Forth source code. No structure is imposed on the contents of the
8561: block. A block is identified by its number; blocks are numbered
8562: contiguously from 1 to an implementation-defined maximum.
8563:
8564: A typical system that used blocks but no operating system might use a
8565: single floppy-disk drive for mass storage, with the disks formatted to
8566: provide 256-byte sectors. Blocks would be implemented by assigning the
8567: first four sectors of the disk to block 1, the second four sectors to
8568: block 2 and so on, up to the limit of the capacity of the disk. The disk
8569: would not contain any file system information, just the set of blocks.
8570:
1.29 crook 8571: @cindex blocks file
1.28 crook 8572: On systems that do provide file services, blocks are typically
1.29 crook 8573: implemented by storing a sequence of blocks within a single @dfn{blocks
1.28 crook 8574: file}. The size of the blocks file will be an exact multiple of 1024
8575: bytes, corresponding to the number of blocks it contains. This is the
8576: mechanism that Gforth uses.
8577:
1.29 crook 8578: @cindex @file{blocks.fb}
1.75 anton 8579: Only one blocks file can be open at a time. If you use block words without
1.28 crook 8580: having specified a blocks file, Gforth defaults to the blocks file
8581: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
1.75 anton 8582: locate a blocks file (@pxref{Source Search Paths}).
1.28 crook 8583:
1.29 crook 8584: @cindex block buffers
1.28 crook 8585: When you read and write blocks under program control, Gforth uses a
1.29 crook 8586: number of @dfn{block buffers} as intermediate storage. These buffers are
1.28 crook 8587: not used when you use @code{load} to interpret the contents of a block.
8588:
1.75 anton 8589: The behaviour of the block buffers is analagous to that of a cache.
8590: Each block buffer has three states:
1.28 crook 8591:
8592: @itemize @bullet
8593: @item
8594: Unassigned
8595: @item
8596: Assigned-clean
8597: @item
8598: Assigned-dirty
8599: @end itemize
8600:
1.29 crook 8601: Initially, all block buffers are @i{unassigned}. In order to access a
1.28 crook 8602: block, the block (specified by its block number) must be assigned to a
8603: block buffer.
8604:
8605: The assignment of a block to a block buffer is performed by @code{block}
8606: or @code{buffer}. Use @code{block} when you wish to modify the existing
8607: contents of a block. Use @code{buffer} when you don't care about the
8608: existing contents of the block@footnote{The ANS Forth definition of
1.35 anton 8609: @code{buffer} is intended not to cause disk I/O; if the data associated
1.28 crook 8610: with the particular block is already stored in a block buffer due to an
8611: earlier @code{block} command, @code{buffer} will return that block
8612: buffer and the existing contents of the block will be
8613: available. Otherwise, @code{buffer} will simply assign a new, empty
1.29 crook 8614: block buffer for the block.}.
1.28 crook 8615:
1.47 crook 8616: Once a block has been assigned to a block buffer using @code{block} or
1.75 anton 8617: @code{buffer}, that block buffer becomes the @i{current block
8618: buffer}. Data may only be manipulated (read or written) within the
8619: current block buffer.
1.47 crook 8620:
8621: When the contents of the current block buffer has been modified it is
1.48 anton 8622: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
1.75 anton 8623: either abandon the changes (by doing nothing) or mark the block as
8624: changed (assigned-dirty), using @code{update}. Using @code{update} does
8625: not change the blocks file; it simply changes a block buffer's state to
8626: @i{assigned-dirty}. The block will be written implicitly when it's
8627: buffer is needed for another block, or explicitly by @code{flush} or
8628: @code{save-buffers}.
8629:
8630: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
8631: blocks file on disk. Leaving Gforth with @code{bye} also performs a
8632: @code{flush}.
1.28 crook 8633:
1.29 crook 8634: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
1.28 crook 8635: algorithm to assign a block buffer to a block. That means that any
8636: particular block can only be assigned to one specific block buffer,
1.29 crook 8637: called (for the particular operation) the @i{victim buffer}. If the
1.47 crook 8638: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8639: the new block immediately. If it is @i{assigned-dirty} its current
8640: contents are written back to the blocks file on disk before it is
1.28 crook 8641: allocated to the new block.
8642:
8643: Although no structure is imposed on the contents of a block, it is
8644: traditional to display the contents as 16 lines each of 64 characters. A
8645: block provides a single, continuous stream of input (for example, it
8646: acts as a single parse area) -- there are no end-of-line characters
8647: within a block, and no end-of-file character at the end of a
8648: block. There are two consequences of this:
1.26 crook 8649:
1.28 crook 8650: @itemize @bullet
8651: @item
8652: The last character of one line wraps straight into the first character
8653: of the following line
8654: @item
8655: The word @code{\} -- comment to end of line -- requires special
8656: treatment; in the context of a block it causes all characters until the
8657: end of the current 64-character ``line'' to be ignored.
8658: @end itemize
8659:
8660: In Gforth, when you use @code{block} with a non-existent block number,
1.45 crook 8661: the current blocks file will be extended to the appropriate size and the
1.28 crook 8662: block buffer will be initialised with spaces.
8663:
1.47 crook 8664: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8665: for details) but doesn't encourage the use of blocks; the mechanism is
8666: only provided for backward compatibility -- ANS Forth requires blocks to
8667: be available when files are.
1.28 crook 8668:
8669: Common techniques that are used when working with blocks include:
8670:
8671: @itemize @bullet
8672: @item
8673: A screen editor that allows you to edit blocks without leaving the Forth
8674: environment.
8675: @item
8676: Shadow screens; where every code block has an associated block
8677: containing comments (for example: code in odd block numbers, comments in
8678: even block numbers). Typically, the block editor provides a convenient
8679: mechanism to toggle between code and comments.
8680: @item
8681: Load blocks; a single block (typically block 1) contains a number of
8682: @code{thru} commands which @code{load} the whole of the application.
8683: @end itemize
1.26 crook 8684:
1.29 crook 8685: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8686: integrated into a Forth programming environment.
1.26 crook 8687:
8688: @comment TODO what about errors on open-blocks?
1.44 crook 8689:
1.26 crook 8690: doc-open-blocks
8691: doc-use
1.75 anton 8692: doc-block-offset
1.26 crook 8693: doc-get-block-fid
8694: doc-block-position
1.28 crook 8695:
1.75 anton 8696: doc-list
1.28 crook 8697: doc-scr
8698:
1.45 crook 8699: doc---gforthman-block
1.28 crook 8700: doc-buffer
8701:
1.75 anton 8702: doc-empty-buffers
8703: doc-empty-buffer
1.26 crook 8704: doc-update
1.28 crook 8705: doc-updated?
1.26 crook 8706: doc-save-buffers
1.75 anton 8707: doc-save-buffer
1.26 crook 8708: doc-flush
1.28 crook 8709:
1.26 crook 8710: doc-load
8711: doc-thru
8712: doc-+load
8713: doc-+thru
1.45 crook 8714: doc---gforthman--->
1.26 crook 8715: doc-block-included
8716:
1.44 crook 8717:
1.26 crook 8718: @c -------------------------------------------------------------
1.126 pazsan 8719: @node Other I/O, OS command line arguments, Blocks, Words
1.26 crook 8720: @section Other I/O
1.28 crook 8721: @cindex I/O - keyboard and display
1.26 crook 8722:
8723: @menu
8724: * Simple numeric output:: Predefined formats
8725: * Formatted numeric output:: Formatted (pictured) output
8726: * String Formats:: How Forth stores strings in memory
1.67 anton 8727: * Displaying characters and strings:: Other stuff
1.175 anton 8728: * Terminal output:: Cursor positioning etc.
1.181 ! anton 8729: * Single-key input::
! 8730: * Line input and conversion::
1.112 anton 8731: * Pipes:: How to create your own pipes
1.149 pazsan 8732: * Xchars and Unicode:: Non-ASCII characters
1.26 crook 8733: @end menu
8734:
8735: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8736: @subsection Simple numeric output
1.28 crook 8737: @cindex numeric output - simple/free-format
1.5 anton 8738:
1.26 crook 8739: The simplest output functions are those that display numbers from the
8740: data or floating-point stacks. Floating-point output is always displayed
8741: using base 10. Numbers displayed from the data stack use the value stored
8742: in @code{base}.
1.5 anton 8743:
1.44 crook 8744:
1.26 crook 8745: doc-.
8746: doc-dec.
8747: doc-hex.
8748: doc-u.
8749: doc-.r
8750: doc-u.r
8751: doc-d.
8752: doc-ud.
8753: doc-d.r
8754: doc-ud.r
8755: doc-f.
8756: doc-fe.
8757: doc-fs.
1.111 anton 8758: doc-f.rdp
1.44 crook 8759:
1.26 crook 8760: Examples of printing the number 1234.5678E23 in the different floating-point output
8761: formats are shown below:
1.5 anton 8762:
8763: @example
1.26 crook 8764: f. 123456779999999000000000000.
8765: fe. 123.456779999999E24
8766: fs. 1.23456779999999E26
1.5 anton 8767: @end example
8768:
8769:
1.26 crook 8770: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8771: @subsection Formatted numeric output
1.28 crook 8772: @cindex formatted numeric output
1.26 crook 8773: @cindex pictured numeric output
1.28 crook 8774: @cindex numeric output - formatted
1.26 crook 8775:
1.29 crook 8776: Forth traditionally uses a technique called @dfn{pictured numeric
1.26 crook 8777: output} for formatted printing of integers. In this technique, digits
8778: are extracted from the number (using the current output radix defined by
8779: @code{base}), converted to ASCII codes and appended to a string that is
8780: built in a scratch-pad area of memory (@pxref{core-idef,
8781: Implementation-defined options, Implementation-defined
8782: options}). Arbitrary characters can be appended to the string during the
8783: extraction process. The completed string is specified by an address
8784: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8785: under program control.
1.5 anton 8786:
1.75 anton 8787: All of the integer output words described in the previous section
8788: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
8789: numeric output.
1.5 anton 8790:
1.47 crook 8791: Three important things to remember about pictured numeric output:
1.5 anton 8792:
1.26 crook 8793: @itemize @bullet
8794: @item
1.28 crook 8795: It always operates on double-precision numbers; to display a
1.49 anton 8796: single-precision number, convert it first (for ways of doing this
8797: @pxref{Double precision}).
1.26 crook 8798: @item
1.28 crook 8799: It always treats the double-precision number as though it were
8800: unsigned. The examples below show ways of printing signed numbers.
1.26 crook 8801: @item
8802: The string is built up from right to left; least significant digit first.
8803: @end itemize
1.5 anton 8804:
1.44 crook 8805:
1.26 crook 8806: doc-<#
1.47 crook 8807: doc-<<#
1.26 crook 8808: doc-#
8809: doc-#s
8810: doc-hold
8811: doc-sign
8812: doc-#>
1.47 crook 8813: doc-#>>
1.5 anton 8814:
1.26 crook 8815: doc-represent
1.111 anton 8816: doc-f>str-rdp
8817: doc-f>buf-rdp
1.5 anton 8818:
1.44 crook 8819:
8820: @noindent
1.26 crook 8821: Here are some examples of using pictured numeric output:
1.5 anton 8822:
8823: @example
1.26 crook 8824: : my-u. ( u -- )
8825: \ Simplest use of pns.. behaves like Standard u.
8826: 0 \ convert to unsigned double
1.75 anton 8827: <<# \ start conversion
1.26 crook 8828: #s \ convert all digits
8829: #> \ complete conversion
1.75 anton 8830: TYPE SPACE \ display, with trailing space
8831: #>> ; \ release hold area
1.5 anton 8832:
1.26 crook 8833: : cents-only ( u -- )
8834: 0 \ convert to unsigned double
1.75 anton 8835: <<# \ start conversion
1.26 crook 8836: # # \ convert two least-significant digits
8837: #> \ complete conversion, discard other digits
1.75 anton 8838: TYPE SPACE \ display, with trailing space
8839: #>> ; \ release hold area
1.5 anton 8840:
1.26 crook 8841: : dollars-and-cents ( u -- )
8842: 0 \ convert to unsigned double
1.75 anton 8843: <<# \ start conversion
1.26 crook 8844: # # \ convert two least-significant digits
8845: [char] . hold \ insert decimal point
8846: #s \ convert remaining digits
8847: [char] $ hold \ append currency symbol
8848: #> \ complete conversion
1.75 anton 8849: TYPE SPACE \ display, with trailing space
8850: #>> ; \ release hold area
1.5 anton 8851:
1.26 crook 8852: : my-. ( n -- )
8853: \ handling negatives.. behaves like Standard .
8854: s>d \ convert to signed double
8855: swap over dabs \ leave sign byte followed by unsigned double
1.75 anton 8856: <<# \ start conversion
1.26 crook 8857: #s \ convert all digits
8858: rot sign \ get at sign byte, append "-" if needed
8859: #> \ complete conversion
1.75 anton 8860: TYPE SPACE \ display, with trailing space
8861: #>> ; \ release hold area
1.5 anton 8862:
1.26 crook 8863: : account. ( n -- )
1.75 anton 8864: \ accountants don't like minus signs, they use parentheses
1.26 crook 8865: \ for negative numbers
8866: s>d \ convert to signed double
8867: swap over dabs \ leave sign byte followed by unsigned double
1.75 anton 8868: <<# \ start conversion
1.26 crook 8869: 2 pick \ get copy of sign byte
8870: 0< IF [char] ) hold THEN \ right-most character of output
8871: #s \ convert all digits
8872: rot \ get at sign byte
8873: 0< IF [char] ( hold THEN
8874: #> \ complete conversion
1.75 anton 8875: TYPE SPACE \ display, with trailing space
8876: #>> ; \ release hold area
8877:
1.5 anton 8878: @end example
8879:
1.26 crook 8880: Here are some examples of using these words:
1.5 anton 8881:
8882: @example
1.26 crook 8883: 1 my-u. 1
8884: hex -1 my-u. decimal FFFFFFFF
8885: 1 cents-only 01
8886: 1234 cents-only 34
8887: 2 dollars-and-cents $0.02
8888: 1234 dollars-and-cents $12.34
8889: 123 my-. 123
8890: -123 my. -123
8891: 123 account. 123
8892: -456 account. (456)
1.5 anton 8893: @end example
8894:
8895:
1.26 crook 8896: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8897: @subsection String Formats
1.27 crook 8898: @cindex strings - see character strings
8899: @cindex character strings - formats
1.28 crook 8900: @cindex I/O - see character strings
1.75 anton 8901: @cindex counted strings
8902:
8903: @c anton: this does not really belong here; maybe the memory section,
8904: @c or the principles chapter
1.26 crook 8905:
1.27 crook 8906: Forth commonly uses two different methods for representing character
8907: strings:
1.26 crook 8908:
8909: @itemize @bullet
8910: @item
8911: @cindex address of counted string
1.45 crook 8912: @cindex counted string
1.29 crook 8913: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8914: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8915: string and the string occupies the subsequent @i{n} char addresses in
1.26 crook 8916: memory.
8917: @item
1.29 crook 8918: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8919: of the string in characters, and @i{c-addr} is the address of the
1.26 crook 8920: first byte of the string.
8921: @end itemize
8922:
8923: ANS Forth encourages the use of the second format when representing
1.75 anton 8924: strings.
1.26 crook 8925:
1.44 crook 8926:
1.26 crook 8927: doc-count
8928:
1.44 crook 8929:
1.49 anton 8930: For words that move, copy and search for strings see @ref{Memory
8931: Blocks}. For words that display characters and strings see
8932: @ref{Displaying characters and strings}.
1.26 crook 8933:
1.175 anton 8934: @node Displaying characters and strings, Terminal output, String Formats, Other I/O
1.26 crook 8935: @subsection Displaying characters and strings
1.27 crook 8936: @cindex characters - compiling and displaying
8937: @cindex character strings - compiling and displaying
1.26 crook 8938:
8939: This section starts with a glossary of Forth words and ends with a set
8940: of examples.
8941:
8942: doc-bl
8943: doc-space
8944: doc-spaces
8945: doc-emit
8946: doc-toupper
8947: doc-."
8948: doc-.(
1.98 anton 8949: doc-.\"
1.26 crook 8950: doc-type
1.44 crook 8951: doc-typewhite
1.26 crook 8952: doc-cr
1.27 crook 8953: @cindex cursor control
1.26 crook 8954: doc-s"
1.98 anton 8955: doc-s\"
1.26 crook 8956: doc-c"
8957: doc-char
8958: doc-[char]
8959:
1.44 crook 8960:
8961: @noindent
1.26 crook 8962: As an example, consider the following text, stored in a file @file{test.fs}:
1.5 anton 8963:
8964: @example
1.26 crook 8965: .( text-1)
8966: : my-word
8967: ." text-2" cr
8968: .( text-3)
8969: ;
8970:
8971: ." text-4"
8972:
8973: : my-char
8974: [char] ALPHABET emit
8975: char emit
8976: ;
1.5 anton 8977: @end example
8978:
1.26 crook 8979: When you load this code into Gforth, the following output is generated:
1.5 anton 8980:
1.26 crook 8981: @example
1.30 anton 8982: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
1.26 crook 8983: @end example
1.5 anton 8984:
1.26 crook 8985: @itemize @bullet
8986: @item
8987: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
8988: is an immediate word; it behaves in the same way whether it is used inside
8989: or outside a colon definition.
8990: @item
8991: Message @code{text-4} is displayed because of Gforth's added interpretation
8992: semantics for @code{."}.
8993: @item
1.29 crook 8994: Message @code{text-2} is @i{not} displayed, because the text interpreter
1.26 crook 8995: performs the compilation semantics for @code{."} within the definition of
8996: @code{my-word}.
8997: @end itemize
1.5 anton 8998:
1.26 crook 8999: Here are some examples of executing @code{my-word} and @code{my-char}:
1.5 anton 9000:
1.26 crook 9001: @example
1.30 anton 9002: @kbd{my-word @key{RET}} text-2
1.26 crook 9003: ok
1.30 anton 9004: @kbd{my-char fred @key{RET}} Af ok
9005: @kbd{my-char jim @key{RET}} Aj ok
1.26 crook 9006: @end example
1.5 anton 9007:
9008: @itemize @bullet
9009: @item
1.26 crook 9010: Message @code{text-2} is displayed because of the run-time behaviour of
9011: @code{."}.
9012: @item
9013: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
9014: on the stack at run-time. @code{emit} always displays the character
9015: when @code{my-char} is executed.
9016: @item
9017: @code{char} parses a string at run-time and the second @code{emit} displays
9018: the first character of the string.
1.5 anton 9019: @item
1.26 crook 9020: If you type @code{see my-char} you can see that @code{[char]} discarded
9021: the text ``LPHABET'' and only compiled the display code for ``A'' into the
9022: definition of @code{my-char}.
1.5 anton 9023: @end itemize
9024:
9025:
1.181 ! anton 9026: @node Terminal output, Single-key input, Displaying characters and strings, Other I/O
1.175 anton 9027: @subsection Terminal output
9028: @cindex output to terminal
9029: @cindex terminal output
9030:
9031: If you are outputting to a terminal, you may want to control the
9032: positioning of the cursor:
9033: @cindex cursor positioning
9034:
9035: doc-at-xy
9036:
9037: In order to know where to position the cursor, it is often helpful to
9038: know the size of the screen:
9039: @cindex terminal size
9040:
9041: doc-form
9042:
9043: And sometimes you want to use:
9044: @cindex clear screen
9045:
9046: doc-page
9047:
9048: Note that on non-terminals you should use @code{12 emit}, not
9049: @code{page}, to get a form feed.
9050:
1.5 anton 9051:
1.181 ! anton 9052: @node Single-key input, Line input and conversion, Terminal output, Other I/O
! 9053: @subsection Single-key input
! 9054: @cindex single-key input
! 9055: @cindex input, single-key
! 9056:
! 9057: If you want to get a single printable character, you can use
! 9058: @code{key}; to check whether a character is available for @code{key},
! 9059: you can use @code{key?}.
1.5 anton 9060:
1.181 ! anton 9061: doc-key
! 9062: doc-key?
1.27 crook 9063:
1.181 ! anton 9064: If you want to process a mix of printable and non-printable
! 9065: characters, you can do that with @code{ekey} and friends. @code{Ekey}
! 9066: produces a keyboard event that you have to convert into a character
! 9067: with @code{ekey>char} or into a key identifier with @code{ekey>fkey}.
! 9068:
! 9069: Typical code for using EKEY looks like this:
! 9070:
! 9071: @example
! 9072: ekey ekey>char if ( c )
! 9073: ... \ do something with the character
! 9074: else ekey>fkey if ( key-id )
! 9075: case
! 9076: k-up of ... endof
! 9077: k-f1 of ... endof
! 9078: k-left k-shift-mask or k-ctrl-mask or of ... endof
! 9079: ...
! 9080: endcase
! 9081: else ( keyboard-event )
! 9082: drop \ just ignore an unknown keyboard event type
! 9083: then then
! 9084: @end example
1.44 crook 9085:
1.45 crook 9086: doc-ekey
1.141 anton 9087: doc-ekey>char
1.181 ! anton 9088: doc-ekey>fkey
1.45 crook 9089: doc-ekey?
1.141 anton 9090:
1.181 ! anton 9091: The key identifiers for cursor keys are:
1.141 anton 9092:
9093: doc-k-left
9094: doc-k-right
9095: doc-k-up
9096: doc-k-down
9097: doc-k-home
9098: doc-k-end
9099: doc-k-prior
9100: doc-k-next
9101: doc-k-insert
9102: doc-k-delete
9103:
1.181 ! anton 9104: The key identifiers for function keys (aka keypad keys) are:
1.141 anton 9105:
1.181 ! anton 9106: doc-k-f1
! 9107: doc-k-f2
! 9108: doc-k-f3
! 9109: doc-k-f4
! 9110: doc-k-f5
! 9111: doc-k-f6
! 9112: doc-k-f7
! 9113: doc-k-f8
! 9114: doc-k-f9
! 9115: doc-k-f10
! 9116: doc-k-f11
! 9117: doc-k-f12
! 9118:
! 9119: Note that @code{k-f11} and @code{k-f12} are not as widely available.
! 9120:
! 9121: You can combine these key identifiers with masks for various shift keys:
! 9122:
! 9123: doc-k-shift-mask
! 9124: doc-k-ctrl-mask
! 9125: doc-k-alt-mask
! 9126:
! 9127: Note that, even if a Forth system has @code{ekey>fkey} and the key
! 9128: identifier words, the keys are not necessarily available or it may not
! 9129: necessarily be able to report all the keys and all the possible
! 9130: combinations with shift masks. Therefore, write your programs in such
! 9131: a way that they are still useful even if the keys and key combinations
! 9132: cannot be pressed or are not recognized.
! 9133:
! 9134: Examples: Older keyboards often do not have an F11 and F12 key. If
! 9135: you run Gforth in an xterm, the xterm catches a number of combinations
! 9136: (e.g., @key{Shift-Up}), and never passes it to Gforth. Finally,
! 9137: Gforth currently does not recognize and report combinations with
! 9138: multiple shift keys (so the @key{shift-ctrl-left} case in the example
! 9139: above would never be entered).
! 9140:
! 9141: Gforth recognizes various keys available on ANSI terminals (in MS-DOS
! 9142: you need the ANSI.SYS driver to get that behaviour); it works by
! 9143: recognizing the escape sequences that ANSI terminals send when such a
! 9144: key is pressed. If you have a terminal that sends other escape
! 9145: sequences, you will not get useful results on Gforth. Other Forth
! 9146: systems may work in a different way.
! 9147:
! 9148:
! 9149: @node Line input and conversion, Pipes, Single-key input, Other I/O
! 9150: @subsection Line input and conversion
! 9151: @cindex line input from terminal
! 9152: @cindex input, linewise from terminal
! 9153: @cindex convertin strings to numbers
! 9154: @cindex I/O - see input
! 9155:
! 9156: For ways of storing character strings in memory see @ref{String Formats}.
! 9157:
! 9158: @comment TODO examples for >number >float accept key key? pad parse word refill
! 9159: @comment then index them
1.141 anton 9160:
9161: Words for inputting one line from the keyboard:
9162:
9163: doc-accept
9164: doc-edit-line
9165:
9166: Conversion words:
9167:
1.143 anton 9168: doc-s>number?
9169: doc-s>unumber?
1.26 crook 9170: doc->number
9171: doc->float
1.143 anton 9172:
1.141 anton 9173:
1.27 crook 9174: @comment obsolescent words..
1.141 anton 9175: Obsolescent input and conversion words:
9176:
1.27 crook 9177: doc-convert
1.26 crook 9178: doc-expect
1.27 crook 9179: doc-span
1.5 anton 9180:
9181:
1.181 ! anton 9182: @node Pipes, Xchars and Unicode, Line input and conversion, Other I/O
1.112 anton 9183: @subsection Pipes
9184: @cindex pipes, creating your own
9185:
9186: In addition to using Gforth in pipes created by other processes
9187: (@pxref{Gforth in pipes}), you can create your own pipe with
9188: @code{open-pipe}, and read from or write to it.
9189:
9190: doc-open-pipe
9191: doc-close-pipe
9192:
9193: If you write to a pipe, Gforth can throw a @code{broken-pipe-error}; if
9194: you don't catch this exception, Gforth will catch it and exit, usually
9195: silently (@pxref{Gforth in pipes}). Since you probably do not want
9196: this, you should wrap a @code{catch} or @code{try} block around the code
9197: from @code{open-pipe} to @code{close-pipe}, so you can deal with the
9198: problem yourself, and then return to regular processing.
9199:
9200: doc-broken-pipe-error
9201:
1.155 anton 9202: @node Xchars and Unicode, , Pipes, Other I/O
9203: @subsection Xchars and Unicode
1.149 pazsan 9204:
9205: This chapter needs completion
1.112 anton 9206:
1.121 anton 9207: @node OS command line arguments, Locals, Other I/O, Words
9208: @section OS command line arguments
9209: @cindex OS command line arguments
9210: @cindex command line arguments, OS
9211: @cindex arguments, OS command line
9212:
9213: The usual way to pass arguments to Gforth programs on the command line
9214: is via the @option{-e} option, e.g.
9215:
9216: @example
9217: gforth -e "123 456" foo.fs -e bye
9218: @end example
9219:
9220: However, you may want to interpret the command-line arguments directly.
9221: In that case, you can access the (image-specific) command-line arguments
1.123 anton 9222: through @code{next-arg}:
1.121 anton 9223:
1.123 anton 9224: doc-next-arg
1.121 anton 9225:
1.123 anton 9226: Here's an example program @file{echo.fs} for @code{next-arg}:
1.121 anton 9227:
9228: @example
9229: : echo ( -- )
1.122 anton 9230: begin
1.123 anton 9231: next-arg 2dup 0 0 d<> while
9232: type space
9233: repeat
9234: 2drop ;
1.121 anton 9235:
9236: echo cr bye
9237: @end example
9238:
9239: This can be invoked with
9240:
9241: @example
9242: gforth echo.fs hello world
9243: @end example
1.123 anton 9244:
9245: and it will print
9246:
9247: @example
9248: hello world
9249: @end example
9250:
9251: The next lower level of dealing with the OS command line are the
9252: following words:
9253:
9254: doc-arg
9255: doc-shift-args
9256:
9257: Finally, at the lowest level Gforth provides the following words:
9258:
9259: doc-argc
9260: doc-argv
1.121 anton 9261:
1.78 anton 9262: @c -------------------------------------------------------------
1.126 pazsan 9263: @node Locals, Structures, OS command line arguments, Words
1.78 anton 9264: @section Locals
9265: @cindex locals
9266:
9267: Local variables can make Forth programming more enjoyable and Forth
9268: programs easier to read. Unfortunately, the locals of ANS Forth are
9269: laden with restrictions. Therefore, we provide not only the ANS Forth
9270: locals wordset, but also our own, more powerful locals wordset (we
9271: implemented the ANS Forth locals wordset through our locals wordset).
1.44 crook 9272:
1.78 anton 9273: The ideas in this section have also been published in M. Anton Ertl,
9274: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
9275: Automatic Scoping of Local Variables}}, EuroForth '94.
1.12 anton 9276:
9277: @menu
1.78 anton 9278: * Gforth locals::
9279: * ANS Forth locals::
1.5 anton 9280: @end menu
9281:
1.78 anton 9282: @node Gforth locals, ANS Forth locals, Locals, Locals
9283: @subsection Gforth locals
9284: @cindex Gforth locals
9285: @cindex locals, Gforth style
1.5 anton 9286:
1.78 anton 9287: Locals can be defined with
1.44 crook 9288:
1.78 anton 9289: @example
9290: @{ local1 local2 ... -- comment @}
9291: @end example
9292: or
9293: @example
9294: @{ local1 local2 ... @}
9295: @end example
1.5 anton 9296:
1.78 anton 9297: E.g.,
9298: @example
9299: : max @{ n1 n2 -- n3 @}
9300: n1 n2 > if
9301: n1
9302: else
9303: n2
9304: endif ;
9305: @end example
1.44 crook 9306:
1.78 anton 9307: The similarity of locals definitions with stack comments is intended. A
9308: locals definition often replaces the stack comment of a word. The order
9309: of the locals corresponds to the order in a stack comment and everything
9310: after the @code{--} is really a comment.
1.77 anton 9311:
1.78 anton 9312: This similarity has one disadvantage: It is too easy to confuse locals
9313: declarations with stack comments, causing bugs and making them hard to
9314: find. However, this problem can be avoided by appropriate coding
9315: conventions: Do not use both notations in the same program. If you do,
9316: they should be distinguished using additional means, e.g. by position.
1.77 anton 9317:
1.78 anton 9318: @cindex types of locals
9319: @cindex locals types
9320: The name of the local may be preceded by a type specifier, e.g.,
9321: @code{F:} for a floating point value:
1.5 anton 9322:
1.78 anton 9323: @example
9324: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
9325: \ complex multiplication
9326: Ar Br f* Ai Bi f* f-
9327: Ar Bi f* Ai Br f* f+ ;
9328: @end example
1.44 crook 9329:
1.78 anton 9330: @cindex flavours of locals
9331: @cindex locals flavours
9332: @cindex value-flavoured locals
9333: @cindex variable-flavoured locals
9334: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9335: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9336: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9337: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9338: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9339: produces its address (which becomes invalid when the variable's scope is
9340: left). E.g., the standard word @code{emit} can be defined in terms of
9341: @code{type} like this:
1.5 anton 9342:
1.78 anton 9343: @example
9344: : emit @{ C^ char* -- @}
9345: char* 1 type ;
9346: @end example
1.5 anton 9347:
1.78 anton 9348: @cindex default type of locals
9349: @cindex locals, default type
9350: A local without type specifier is a @code{W:} local. Both flavours of
9351: locals are initialized with values from the data or FP stack.
1.44 crook 9352:
1.78 anton 9353: Currently there is no way to define locals with user-defined data
9354: structures, but we are working on it.
1.5 anton 9355:
1.78 anton 9356: Gforth allows defining locals everywhere in a colon definition. This
9357: poses the following questions:
1.5 anton 9358:
1.78 anton 9359: @menu
9360: * Where are locals visible by name?::
9361: * How long do locals live?::
9362: * Locals programming style::
9363: * Locals implementation::
9364: @end menu
1.44 crook 9365:
1.78 anton 9366: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9367: @subsubsection Where are locals visible by name?
9368: @cindex locals visibility
9369: @cindex visibility of locals
9370: @cindex scope of locals
1.5 anton 9371:
1.78 anton 9372: Basically, the answer is that locals are visible where you would expect
9373: it in block-structured languages, and sometimes a little longer. If you
9374: want to restrict the scope of a local, enclose its definition in
9375: @code{SCOPE}...@code{ENDSCOPE}.
1.5 anton 9376:
9377:
1.78 anton 9378: doc-scope
9379: doc-endscope
1.5 anton 9380:
9381:
1.78 anton 9382: These words behave like control structure words, so you can use them
9383: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9384: arbitrary ways.
1.77 anton 9385:
1.78 anton 9386: If you want a more exact answer to the visibility question, here's the
9387: basic principle: A local is visible in all places that can only be
9388: reached through the definition of the local@footnote{In compiler
9389: construction terminology, all places dominated by the definition of the
9390: local.}. In other words, it is not visible in places that can be reached
9391: without going through the definition of the local. E.g., locals defined
9392: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9393: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9394: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
1.77 anton 9395:
1.78 anton 9396: The reasoning behind this solution is: We want to have the locals
9397: visible as long as it is meaningful. The user can always make the
9398: visibility shorter by using explicit scoping. In a place that can
9399: only be reached through the definition of a local, the meaning of a
9400: local name is clear. In other places it is not: How is the local
9401: initialized at the control flow path that does not contain the
9402: definition? Which local is meant, if the same name is defined twice in
9403: two independent control flow paths?
1.77 anton 9404:
1.78 anton 9405: This should be enough detail for nearly all users, so you can skip the
9406: rest of this section. If you really must know all the gory details and
9407: options, read on.
1.77 anton 9408:
1.78 anton 9409: In order to implement this rule, the compiler has to know which places
9410: are unreachable. It knows this automatically after @code{AHEAD},
9411: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9412: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9413: compiler that the control flow never reaches that place. If
9414: @code{UNREACHABLE} is not used where it could, the only consequence is
9415: that the visibility of some locals is more limited than the rule above
9416: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9417: lie to the compiler), buggy code will be produced.
1.77 anton 9418:
1.5 anton 9419:
1.78 anton 9420: doc-unreachable
1.5 anton 9421:
1.23 crook 9422:
1.78 anton 9423: Another problem with this rule is that at @code{BEGIN}, the compiler
9424: does not know which locals will be visible on the incoming
9425: back-edge. All problems discussed in the following are due to this
9426: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9427: loops as examples; the discussion also applies to @code{?DO} and other
9428: loops). Perhaps the most insidious example is:
1.26 crook 9429: @example
1.78 anton 9430: AHEAD
9431: BEGIN
9432: x
9433: [ 1 CS-ROLL ] THEN
9434: @{ x @}
9435: ...
9436: UNTIL
1.26 crook 9437: @end example
1.23 crook 9438:
1.78 anton 9439: This should be legal according to the visibility rule. The use of
9440: @code{x} can only be reached through the definition; but that appears
9441: textually below the use.
9442:
9443: From this example it is clear that the visibility rules cannot be fully
9444: implemented without major headaches. Our implementation treats common
9445: cases as advertised and the exceptions are treated in a safe way: The
9446: compiler makes a reasonable guess about the locals visible after a
9447: @code{BEGIN}; if it is too pessimistic, the
9448: user will get a spurious error about the local not being defined; if the
9449: compiler is too optimistic, it will notice this later and issue a
9450: warning. In the case above the compiler would complain about @code{x}
9451: being undefined at its use. You can see from the obscure examples in
9452: this section that it takes quite unusual control structures to get the
9453: compiler into trouble, and even then it will often do fine.
1.23 crook 9454:
1.78 anton 9455: If the @code{BEGIN} is reachable from above, the most optimistic guess
9456: is that all locals visible before the @code{BEGIN} will also be
9457: visible after the @code{BEGIN}. This guess is valid for all loops that
9458: are entered only through the @code{BEGIN}, in particular, for normal
9459: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9460: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9461: compiler. When the branch to the @code{BEGIN} is finally generated by
9462: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9463: warns the user if it was too optimistic:
1.26 crook 9464: @example
1.78 anton 9465: IF
9466: @{ x @}
9467: BEGIN
9468: \ x ?
9469: [ 1 cs-roll ] THEN
9470: ...
9471: UNTIL
1.26 crook 9472: @end example
1.23 crook 9473:
1.78 anton 9474: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9475: optimistically assumes that it lives until the @code{THEN}. It notices
9476: this difference when it compiles the @code{UNTIL} and issues a
9477: warning. The user can avoid the warning, and make sure that @code{x}
9478: is not used in the wrong area by using explicit scoping:
9479: @example
9480: IF
9481: SCOPE
9482: @{ x @}
9483: ENDSCOPE
9484: BEGIN
9485: [ 1 cs-roll ] THEN
9486: ...
9487: UNTIL
9488: @end example
1.23 crook 9489:
1.78 anton 9490: Since the guess is optimistic, there will be no spurious error messages
9491: about undefined locals.
1.44 crook 9492:
1.78 anton 9493: If the @code{BEGIN} is not reachable from above (e.g., after
9494: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9495: optimistic guess, as the locals visible after the @code{BEGIN} may be
9496: defined later. Therefore, the compiler assumes that no locals are
9497: visible after the @code{BEGIN}. However, the user can use
9498: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9499: visible at the BEGIN as at the point where the top control-flow stack
9500: item was created.
1.23 crook 9501:
1.44 crook 9502:
1.78 anton 9503: doc-assume-live
1.26 crook 9504:
1.23 crook 9505:
1.78 anton 9506: @noindent
9507: E.g.,
9508: @example
9509: @{ x @}
9510: AHEAD
9511: ASSUME-LIVE
9512: BEGIN
9513: x
9514: [ 1 CS-ROLL ] THEN
9515: ...
9516: UNTIL
9517: @end example
1.44 crook 9518:
1.78 anton 9519: Other cases where the locals are defined before the @code{BEGIN} can be
9520: handled by inserting an appropriate @code{CS-ROLL} before the
9521: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9522: behind the @code{ASSUME-LIVE}).
1.23 crook 9523:
1.78 anton 9524: Cases where locals are defined after the @code{BEGIN} (but should be
9525: visible immediately after the @code{BEGIN}) can only be handled by
9526: rearranging the loop. E.g., the ``most insidious'' example above can be
9527: arranged into:
9528: @example
9529: BEGIN
9530: @{ x @}
9531: ... 0=
9532: WHILE
9533: x
9534: REPEAT
9535: @end example
1.44 crook 9536:
1.78 anton 9537: @node How long do locals live?, Locals programming style, Where are locals visible by name?, Gforth locals
9538: @subsubsection How long do locals live?
9539: @cindex locals lifetime
9540: @cindex lifetime of locals
1.23 crook 9541:
1.78 anton 9542: The right answer for the lifetime question would be: A local lives at
9543: least as long as it can be accessed. For a value-flavoured local this
9544: means: until the end of its visibility. However, a variable-flavoured
9545: local could be accessed through its address far beyond its visibility
9546: scope. Ultimately, this would mean that such locals would have to be
9547: garbage collected. Since this entails un-Forth-like implementation
9548: complexities, I adopted the same cowardly solution as some other
9549: languages (e.g., C): The local lives only as long as it is visible;
9550: afterwards its address is invalid (and programs that access it
9551: afterwards are erroneous).
1.23 crook 9552:
1.78 anton 9553: @node Locals programming style, Locals implementation, How long do locals live?, Gforth locals
9554: @subsubsection Locals programming style
9555: @cindex locals programming style
9556: @cindex programming style, locals
1.23 crook 9557:
1.78 anton 9558: The freedom to define locals anywhere has the potential to change
9559: programming styles dramatically. In particular, the need to use the
9560: return stack for intermediate storage vanishes. Moreover, all stack
9561: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9562: determined arguments) can be eliminated: If the stack items are in the
9563: wrong order, just write a locals definition for all of them; then
9564: write the items in the order you want.
1.23 crook 9565:
1.78 anton 9566: This seems a little far-fetched and eliminating stack manipulations is
9567: unlikely to become a conscious programming objective. Still, the number
9568: of stack manipulations will be reduced dramatically if local variables
9569: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
9570: a traditional implementation of @code{max}).
1.23 crook 9571:
1.78 anton 9572: This shows one potential benefit of locals: making Forth programs more
9573: readable. Of course, this benefit will only be realized if the
9574: programmers continue to honour the principle of factoring instead of
9575: using the added latitude to make the words longer.
1.23 crook 9576:
1.78 anton 9577: @cindex single-assignment style for locals
9578: Using @code{TO} can and should be avoided. Without @code{TO},
9579: every value-flavoured local has only a single assignment and many
9580: advantages of functional languages apply to Forth. I.e., programs are
9581: easier to analyse, to optimize and to read: It is clear from the
9582: definition what the local stands for, it does not turn into something
9583: different later.
1.23 crook 9584:
1.78 anton 9585: E.g., a definition using @code{TO} might look like this:
9586: @example
9587: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9588: u1 u2 min 0
9589: ?do
9590: addr1 c@@ addr2 c@@ -
9591: ?dup-if
9592: unloop exit
9593: then
9594: addr1 char+ TO addr1
9595: addr2 char+ TO addr2
9596: loop
9597: u1 u2 - ;
1.26 crook 9598: @end example
1.78 anton 9599: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9600: every loop iteration. @code{strcmp} is a typical example of the
9601: readability problems of using @code{TO}. When you start reading
9602: @code{strcmp}, you think that @code{addr1} refers to the start of the
9603: string. Only near the end of the loop you realize that it is something
9604: else.
1.23 crook 9605:
1.78 anton 9606: This can be avoided by defining two locals at the start of the loop that
9607: are initialized with the right value for the current iteration.
9608: @example
9609: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9610: addr1 addr2
9611: u1 u2 min 0
9612: ?do @{ s1 s2 @}
9613: s1 c@@ s2 c@@ -
9614: ?dup-if
9615: unloop exit
9616: then
9617: s1 char+ s2 char+
9618: loop
9619: 2drop
9620: u1 u2 - ;
9621: @end example
9622: Here it is clear from the start that @code{s1} has a different value
9623: in every loop iteration.
1.23 crook 9624:
1.78 anton 9625: @node Locals implementation, , Locals programming style, Gforth locals
9626: @subsubsection Locals implementation
9627: @cindex locals implementation
9628: @cindex implementation of locals
1.23 crook 9629:
1.78 anton 9630: @cindex locals stack
9631: Gforth uses an extra locals stack. The most compelling reason for
9632: this is that the return stack is not float-aligned; using an extra stack
9633: also eliminates the problems and restrictions of using the return stack
9634: as locals stack. Like the other stacks, the locals stack grows toward
9635: lower addresses. A few primitives allow an efficient implementation:
9636:
9637:
9638: doc-@local#
9639: doc-f@local#
9640: doc-laddr#
9641: doc-lp+!#
9642: doc-lp!
9643: doc->l
9644: doc-f>l
9645:
9646:
9647: In addition to these primitives, some specializations of these
9648: primitives for commonly occurring inline arguments are provided for
9649: efficiency reasons, e.g., @code{@@local0} as specialization of
9650: @code{@@local#} for the inline argument 0. The following compiling words
9651: compile the right specialized version, or the general version, as
9652: appropriate:
1.23 crook 9653:
1.5 anton 9654:
1.107 dvdkhlng 9655: @c doc-compile-@local
9656: @c doc-compile-f@local
1.78 anton 9657: doc-compile-lp+!
1.5 anton 9658:
9659:
1.78 anton 9660: Combinations of conditional branches and @code{lp+!#} like
9661: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9662: is taken) are provided for efficiency and correctness in loops.
1.5 anton 9663:
1.78 anton 9664: A special area in the dictionary space is reserved for keeping the
9665: local variable names. @code{@{} switches the dictionary pointer to this
9666: area and @code{@}} switches it back and generates the locals
9667: initializing code. @code{W:} etc.@ are normal defining words. This
9668: special area is cleared at the start of every colon definition.
1.5 anton 9669:
1.78 anton 9670: @cindex word list for defining locals
9671: A special feature of Gforth's dictionary is used to implement the
9672: definition of locals without type specifiers: every word list (aka
9673: vocabulary) has its own methods for searching
9674: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9675: with a special search method: When it is searched for a word, it
9676: actually creates that word using @code{W:}. @code{@{} changes the search
9677: order to first search the word list containing @code{@}}, @code{W:} etc.,
9678: and then the word list for defining locals without type specifiers.
1.5 anton 9679:
1.78 anton 9680: The lifetime rules support a stack discipline within a colon
9681: definition: The lifetime of a local is either nested with other locals
9682: lifetimes or it does not overlap them.
1.23 crook 9683:
1.78 anton 9684: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9685: pointer manipulation is generated. Between control structure words
9686: locals definitions can push locals onto the locals stack. @code{AGAIN}
9687: is the simplest of the other three control flow words. It has to
9688: restore the locals stack depth of the corresponding @code{BEGIN}
9689: before branching. The code looks like this:
9690: @format
9691: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9692: @code{branch} <begin>
9693: @end format
1.26 crook 9694:
1.78 anton 9695: @code{UNTIL} is a little more complicated: If it branches back, it
9696: must adjust the stack just like @code{AGAIN}. But if it falls through,
9697: the locals stack must not be changed. The compiler generates the
9698: following code:
9699: @format
9700: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9701: @end format
9702: The locals stack pointer is only adjusted if the branch is taken.
1.44 crook 9703:
1.78 anton 9704: @code{THEN} can produce somewhat inefficient code:
9705: @format
9706: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9707: <orig target>:
9708: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9709: @end format
9710: The second @code{lp+!#} adjusts the locals stack pointer from the
9711: level at the @i{orig} point to the level after the @code{THEN}. The
9712: first @code{lp+!#} adjusts the locals stack pointer from the current
9713: level to the level at the orig point, so the complete effect is an
9714: adjustment from the current level to the right level after the
9715: @code{THEN}.
1.26 crook 9716:
1.78 anton 9717: @cindex locals information on the control-flow stack
9718: @cindex control-flow stack items, locals information
9719: In a conventional Forth implementation a dest control-flow stack entry
9720: is just the target address and an orig entry is just the address to be
9721: patched. Our locals implementation adds a word list to every orig or dest
9722: item. It is the list of locals visible (or assumed visible) at the point
9723: described by the entry. Our implementation also adds a tag to identify
9724: the kind of entry, in particular to differentiate between live and dead
9725: (reachable and unreachable) orig entries.
1.26 crook 9726:
1.78 anton 9727: A few unusual operations have to be performed on locals word lists:
1.44 crook 9728:
1.5 anton 9729:
1.78 anton 9730: doc-common-list
9731: doc-sub-list?
9732: doc-list-size
1.52 anton 9733:
9734:
1.78 anton 9735: Several features of our locals word list implementation make these
9736: operations easy to implement: The locals word lists are organised as
9737: linked lists; the tails of these lists are shared, if the lists
9738: contain some of the same locals; and the address of a name is greater
9739: than the address of the names behind it in the list.
1.5 anton 9740:
1.78 anton 9741: Another important implementation detail is the variable
9742: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9743: determine if they can be reached directly or only through the branch
9744: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9745: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9746: definition, by @code{BEGIN} and usually by @code{THEN}.
1.5 anton 9747:
1.78 anton 9748: Counted loops are similar to other loops in most respects, but
9749: @code{LEAVE} requires special attention: It performs basically the same
9750: service as @code{AHEAD}, but it does not create a control-flow stack
9751: entry. Therefore the information has to be stored elsewhere;
9752: traditionally, the information was stored in the target fields of the
9753: branches created by the @code{LEAVE}s, by organizing these fields into a
9754: linked list. Unfortunately, this clever trick does not provide enough
9755: space for storing our extended control flow information. Therefore, we
9756: introduce another stack, the leave stack. It contains the control-flow
9757: stack entries for all unresolved @code{LEAVE}s.
1.44 crook 9758:
1.78 anton 9759: Local names are kept until the end of the colon definition, even if
9760: they are no longer visible in any control-flow path. In a few cases
9761: this may lead to increased space needs for the locals name area, but
9762: usually less than reclaiming this space would cost in code size.
1.5 anton 9763:
1.44 crook 9764:
1.78 anton 9765: @node ANS Forth locals, , Gforth locals, Locals
9766: @subsection ANS Forth locals
9767: @cindex locals, ANS Forth style
1.5 anton 9768:
1.78 anton 9769: The ANS Forth locals wordset does not define a syntax for locals, but
9770: words that make it possible to define various syntaxes. One of the
9771: possible syntaxes is a subset of the syntax we used in the Gforth locals
9772: wordset, i.e.:
1.29 crook 9773:
9774: @example
1.78 anton 9775: @{ local1 local2 ... -- comment @}
9776: @end example
9777: @noindent
9778: or
9779: @example
9780: @{ local1 local2 ... @}
1.29 crook 9781: @end example
9782:
1.78 anton 9783: The order of the locals corresponds to the order in a stack comment. The
9784: restrictions are:
1.5 anton 9785:
1.78 anton 9786: @itemize @bullet
9787: @item
9788: Locals can only be cell-sized values (no type specifiers are allowed).
9789: @item
9790: Locals can be defined only outside control structures.
9791: @item
9792: Locals can interfere with explicit usage of the return stack. For the
9793: exact (and long) rules, see the standard. If you don't use return stack
9794: accessing words in a definition using locals, you will be all right. The
9795: purpose of this rule is to make locals implementation on the return
9796: stack easier.
9797: @item
9798: The whole definition must be in one line.
9799: @end itemize
1.5 anton 9800:
1.78 anton 9801: Locals defined in ANS Forth behave like @code{VALUE}s
9802: (@pxref{Values}). I.e., they are initialized from the stack. Using their
9803: name produces their value. Their value can be changed using @code{TO}.
1.77 anton 9804:
1.78 anton 9805: Since the syntax above is supported by Gforth directly, you need not do
9806: anything to use it. If you want to port a program using this syntax to
9807: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9808: syntax on the other system.
1.5 anton 9809:
1.78 anton 9810: Note that a syntax shown in the standard, section A.13 looks
9811: similar, but is quite different in having the order of locals
9812: reversed. Beware!
1.5 anton 9813:
1.78 anton 9814: The ANS Forth locals wordset itself consists of one word:
1.5 anton 9815:
1.78 anton 9816: doc-(local)
1.5 anton 9817:
1.78 anton 9818: The ANS Forth locals extension wordset defines a syntax using
9819: @code{locals|}, but it is so awful that we strongly recommend not to use
9820: it. We have implemented this syntax to make porting to Gforth easy, but
9821: do not document it here. The problem with this syntax is that the locals
9822: are defined in an order reversed with respect to the standard stack
9823: comment notation, making programs harder to read, and easier to misread
9824: and miswrite. The only merit of this syntax is that it is easy to
9825: implement using the ANS Forth locals wordset.
1.53 anton 9826:
9827:
1.78 anton 9828: @c ----------------------------------------------------------
9829: @node Structures, Object-oriented Forth, Locals, Words
9830: @section Structures
9831: @cindex structures
9832: @cindex records
1.53 anton 9833:
1.78 anton 9834: This section presents the structure package that comes with Gforth. A
9835: version of the package implemented in ANS Forth is available in
9836: @file{compat/struct.fs}. This package was inspired by a posting on
9837: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9838: possibly John Hayes). A version of this section has been published in
9839: M. Anton Ertl,
9840: @uref{http://www.complang.tuwien.ac.at/forth/objects/structs.html, Yet
9841: Another Forth Structures Package}, Forth Dimensions 19(3), pages
9842: 13--16. Marcel Hendrix provided helpful comments.
1.53 anton 9843:
1.78 anton 9844: @menu
9845: * Why explicit structure support?::
9846: * Structure Usage::
9847: * Structure Naming Convention::
9848: * Structure Implementation::
9849: * Structure Glossary::
9850: @end menu
1.55 anton 9851:
1.78 anton 9852: @node Why explicit structure support?, Structure Usage, Structures, Structures
9853: @subsection Why explicit structure support?
1.53 anton 9854:
1.78 anton 9855: @cindex address arithmetic for structures
9856: @cindex structures using address arithmetic
9857: If we want to use a structure containing several fields, we could simply
9858: reserve memory for it, and access the fields using address arithmetic
9859: (@pxref{Address arithmetic}). As an example, consider a structure with
9860: the following fields
1.57 anton 9861:
1.78 anton 9862: @table @code
9863: @item a
9864: is a float
9865: @item b
9866: is a cell
9867: @item c
9868: is a float
9869: @end table
1.57 anton 9870:
1.78 anton 9871: Given the (float-aligned) base address of the structure we get the
9872: address of the field
1.52 anton 9873:
1.78 anton 9874: @table @code
9875: @item a
9876: without doing anything further.
9877: @item b
9878: with @code{float+}
9879: @item c
9880: with @code{float+ cell+ faligned}
9881: @end table
1.52 anton 9882:
1.78 anton 9883: It is easy to see that this can become quite tiring.
1.52 anton 9884:
1.78 anton 9885: Moreover, it is not very readable, because seeing a
9886: @code{cell+} tells us neither which kind of structure is
9887: accessed nor what field is accessed; we have to somehow infer the kind
9888: of structure, and then look up in the documentation, which field of
9889: that structure corresponds to that offset.
1.53 anton 9890:
1.78 anton 9891: Finally, this kind of address arithmetic also causes maintenance
9892: troubles: If you add or delete a field somewhere in the middle of the
9893: structure, you have to find and change all computations for the fields
9894: afterwards.
1.52 anton 9895:
1.78 anton 9896: So, instead of using @code{cell+} and friends directly, how
9897: about storing the offsets in constants:
1.52 anton 9898:
1.78 anton 9899: @example
9900: 0 constant a-offset
9901: 0 float+ constant b-offset
9902: 0 float+ cell+ faligned c-offset
9903: @end example
1.64 pazsan 9904:
1.78 anton 9905: Now we can get the address of field @code{x} with @code{x-offset
9906: +}. This is much better in all respects. Of course, you still
9907: have to change all later offset definitions if you add a field. You can
9908: fix this by declaring the offsets in the following way:
1.57 anton 9909:
1.78 anton 9910: @example
9911: 0 constant a-offset
9912: a-offset float+ constant b-offset
9913: b-offset cell+ faligned constant c-offset
9914: @end example
1.57 anton 9915:
1.78 anton 9916: Since we always use the offsets with @code{+}, we could use a defining
9917: word @code{cfield} that includes the @code{+} in the action of the
9918: defined word:
1.64 pazsan 9919:
1.78 anton 9920: @example
9921: : cfield ( n "name" -- )
9922: create ,
9923: does> ( name execution: addr1 -- addr2 )
9924: @@ + ;
1.64 pazsan 9925:
1.78 anton 9926: 0 cfield a
9927: 0 a float+ cfield b
9928: 0 b cell+ faligned cfield c
9929: @end example
1.64 pazsan 9930:
1.78 anton 9931: Instead of @code{x-offset +}, we now simply write @code{x}.
1.64 pazsan 9932:
1.78 anton 9933: The structure field words now can be used quite nicely. However,
9934: their definition is still a bit cumbersome: We have to repeat the
9935: name, the information about size and alignment is distributed before
9936: and after the field definitions etc. The structure package presented
9937: here addresses these problems.
1.64 pazsan 9938:
1.78 anton 9939: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9940: @subsection Structure Usage
9941: @cindex structure usage
1.57 anton 9942:
1.78 anton 9943: @cindex @code{field} usage
9944: @cindex @code{struct} usage
9945: @cindex @code{end-struct} usage
9946: You can define a structure for a (data-less) linked list with:
1.57 anton 9947: @example
1.78 anton 9948: struct
9949: cell% field list-next
9950: end-struct list%
1.57 anton 9951: @end example
9952:
1.78 anton 9953: With the address of the list node on the stack, you can compute the
9954: address of the field that contains the address of the next node with
9955: @code{list-next}. E.g., you can determine the length of a list
9956: with:
1.57 anton 9957:
9958: @example
1.78 anton 9959: : list-length ( list -- n )
9960: \ "list" is a pointer to the first element of a linked list
9961: \ "n" is the length of the list
9962: 0 BEGIN ( list1 n1 )
9963: over
9964: WHILE ( list1 n1 )
9965: 1+ swap list-next @@ swap
9966: REPEAT
9967: nip ;
1.57 anton 9968: @end example
9969:
1.78 anton 9970: You can reserve memory for a list node in the dictionary with
9971: @code{list% %allot}, which leaves the address of the list node on the
9972: stack. For the equivalent allocation on the heap you can use @code{list%
9973: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9974: use @code{list% %allocate}). You can get the the size of a list
9975: node with @code{list% %size} and its alignment with @code{list%
9976: %alignment}.
9977:
9978: Note that in ANS Forth the body of a @code{create}d word is
9979: @code{aligned} but not necessarily @code{faligned};
9980: therefore, if you do a:
1.57 anton 9981:
9982: @example
1.78 anton 9983: create @emph{name} foo% %allot drop
1.57 anton 9984: @end example
9985:
1.78 anton 9986: @noindent
9987: then the memory alloted for @code{foo%} is guaranteed to start at the
9988: body of @code{@emph{name}} only if @code{foo%} contains only character,
9989: cell and double fields. Therefore, if your structure contains floats,
9990: better use
1.57 anton 9991:
9992: @example
1.78 anton 9993: foo% %allot constant @emph{name}
1.57 anton 9994: @end example
9995:
1.78 anton 9996: @cindex structures containing structures
9997: You can include a structure @code{foo%} as a field of
9998: another structure, like this:
1.65 anton 9999: @example
1.78 anton 10000: struct
10001: ...
10002: foo% field ...
10003: ...
10004: end-struct ...
1.65 anton 10005: @end example
1.52 anton 10006:
1.78 anton 10007: @cindex structure extension
10008: @cindex extended records
10009: Instead of starting with an empty structure, you can extend an
10010: existing structure. E.g., a plain linked list without data, as defined
10011: above, is hardly useful; You can extend it to a linked list of integers,
10012: like this:@footnote{This feature is also known as @emph{extended
10013: records}. It is the main innovation in the Oberon language; in other
10014: words, adding this feature to Modula-2 led Wirth to create a new
10015: language, write a new compiler etc. Adding this feature to Forth just
10016: required a few lines of code.}
1.52 anton 10017:
1.78 anton 10018: @example
10019: list%
10020: cell% field intlist-int
10021: end-struct intlist%
10022: @end example
1.55 anton 10023:
1.78 anton 10024: @code{intlist%} is a structure with two fields:
10025: @code{list-next} and @code{intlist-int}.
1.55 anton 10026:
1.78 anton 10027: @cindex structures containing arrays
10028: You can specify an array type containing @emph{n} elements of
10029: type @code{foo%} like this:
1.55 anton 10030:
10031: @example
1.78 anton 10032: foo% @emph{n} *
1.56 anton 10033: @end example
1.55 anton 10034:
1.78 anton 10035: You can use this array type in any place where you can use a normal
10036: type, e.g., when defining a @code{field}, or with
10037: @code{%allot}.
10038:
10039: @cindex first field optimization
10040: The first field is at the base address of a structure and the word for
10041: this field (e.g., @code{list-next}) actually does not change the address
10042: on the stack. You may be tempted to leave it away in the interest of
10043: run-time and space efficiency. This is not necessary, because the
10044: structure package optimizes this case: If you compile a first-field
10045: words, no code is generated. So, in the interest of readability and
10046: maintainability you should include the word for the field when accessing
10047: the field.
1.52 anton 10048:
10049:
1.78 anton 10050: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
10051: @subsection Structure Naming Convention
10052: @cindex structure naming convention
1.52 anton 10053:
1.78 anton 10054: The field names that come to (my) mind are often quite generic, and,
10055: if used, would cause frequent name clashes. E.g., many structures
10056: probably contain a @code{counter} field. The structure names
10057: that come to (my) mind are often also the logical choice for the names
10058: of words that create such a structure.
1.52 anton 10059:
1.78 anton 10060: Therefore, I have adopted the following naming conventions:
1.52 anton 10061:
1.78 anton 10062: @itemize @bullet
10063: @cindex field naming convention
10064: @item
10065: The names of fields are of the form
10066: @code{@emph{struct}-@emph{field}}, where
10067: @code{@emph{struct}} is the basic name of the structure, and
10068: @code{@emph{field}} is the basic name of the field. You can
10069: think of field words as converting the (address of the)
10070: structure into the (address of the) field.
1.52 anton 10071:
1.78 anton 10072: @cindex structure naming convention
10073: @item
10074: The names of structures are of the form
10075: @code{@emph{struct}%}, where
10076: @code{@emph{struct}} is the basic name of the structure.
10077: @end itemize
1.52 anton 10078:
1.78 anton 10079: This naming convention does not work that well for fields of extended
10080: structures; e.g., the integer list structure has a field
10081: @code{intlist-int}, but has @code{list-next}, not
10082: @code{intlist-next}.
1.53 anton 10083:
1.78 anton 10084: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
10085: @subsection Structure Implementation
10086: @cindex structure implementation
10087: @cindex implementation of structures
1.52 anton 10088:
1.78 anton 10089: The central idea in the implementation is to pass the data about the
10090: structure being built on the stack, not in some global
10091: variable. Everything else falls into place naturally once this design
10092: decision is made.
1.53 anton 10093:
1.78 anton 10094: The type description on the stack is of the form @emph{align
10095: size}. Keeping the size on the top-of-stack makes dealing with arrays
10096: very simple.
1.53 anton 10097:
1.78 anton 10098: @code{field} is a defining word that uses @code{Create}
10099: and @code{DOES>}. The body of the field contains the offset
10100: of the field, and the normal @code{DOES>} action is simply:
1.53 anton 10101:
10102: @example
1.78 anton 10103: @@ +
1.53 anton 10104: @end example
10105:
1.78 anton 10106: @noindent
10107: i.e., add the offset to the address, giving the stack effect
10108: @i{addr1 -- addr2} for a field.
10109:
10110: @cindex first field optimization, implementation
10111: This simple structure is slightly complicated by the optimization
10112: for fields with offset 0, which requires a different
10113: @code{DOES>}-part (because we cannot rely on there being
10114: something on the stack if such a field is invoked during
10115: compilation). Therefore, we put the different @code{DOES>}-parts
10116: in separate words, and decide which one to invoke based on the
10117: offset. For a zero offset, the field is basically a noop; it is
10118: immediate, and therefore no code is generated when it is compiled.
1.53 anton 10119:
1.78 anton 10120: @node Structure Glossary, , Structure Implementation, Structures
10121: @subsection Structure Glossary
10122: @cindex structure glossary
1.53 anton 10123:
1.5 anton 10124:
1.78 anton 10125: doc-%align
10126: doc-%alignment
10127: doc-%alloc
10128: doc-%allocate
10129: doc-%allot
10130: doc-cell%
10131: doc-char%
10132: doc-dfloat%
10133: doc-double%
10134: doc-end-struct
10135: doc-field
10136: doc-float%
10137: doc-naligned
10138: doc-sfloat%
10139: doc-%size
10140: doc-struct
1.54 anton 10141:
10142:
1.26 crook 10143: @c -------------------------------------------------------------
1.78 anton 10144: @node Object-oriented Forth, Programming Tools, Structures, Words
10145: @section Object-oriented Forth
10146:
10147: Gforth comes with three packages for object-oriented programming:
10148: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
10149: is preloaded, so you have to @code{include} them before use. The most
10150: important differences between these packages (and others) are discussed
10151: in @ref{Comparison with other object models}. All packages are written
10152: in ANS Forth and can be used with any other ANS Forth.
1.5 anton 10153:
1.78 anton 10154: @menu
10155: * Why object-oriented programming?::
10156: * Object-Oriented Terminology::
10157: * Objects::
10158: * OOF::
10159: * Mini-OOF::
10160: * Comparison with other object models::
10161: @end menu
1.5 anton 10162:
1.78 anton 10163: @c ----------------------------------------------------------------
10164: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
10165: @subsection Why object-oriented programming?
10166: @cindex object-oriented programming motivation
10167: @cindex motivation for object-oriented programming
1.44 crook 10168:
1.78 anton 10169: Often we have to deal with several data structures (@emph{objects}),
10170: that have to be treated similarly in some respects, but differently in
10171: others. Graphical objects are the textbook example: circles, triangles,
10172: dinosaurs, icons, and others, and we may want to add more during program
10173: development. We want to apply some operations to any graphical object,
10174: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
10175: has to do something different for every kind of object.
10176: @comment TODO add some other operations eg perimeter, area
10177: @comment and tie in to concrete examples later..
1.5 anton 10178:
1.78 anton 10179: We could implement @code{draw} as a big @code{CASE}
10180: control structure that executes the appropriate code depending on the
10181: kind of object to be drawn. This would be not be very elegant, and,
10182: moreover, we would have to change @code{draw} every time we add
10183: a new kind of graphical object (say, a spaceship).
1.44 crook 10184:
1.78 anton 10185: What we would rather do is: When defining spaceships, we would tell
10186: the system: ``Here's how you @code{draw} a spaceship; you figure
10187: out the rest''.
1.5 anton 10188:
1.78 anton 10189: This is the problem that all systems solve that (rightfully) call
10190: themselves object-oriented; the object-oriented packages presented here
10191: solve this problem (and not much else).
10192: @comment TODO ?list properties of oo systems.. oo vs o-based?
1.44 crook 10193:
1.78 anton 10194: @c ------------------------------------------------------------------------
10195: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
10196: @subsection Object-Oriented Terminology
10197: @cindex object-oriented terminology
10198: @cindex terminology for object-oriented programming
1.5 anton 10199:
1.78 anton 10200: This section is mainly for reference, so you don't have to understand
10201: all of it right away. The terminology is mainly Smalltalk-inspired. In
10202: short:
1.44 crook 10203:
1.78 anton 10204: @table @emph
10205: @cindex class
10206: @item class
10207: a data structure definition with some extras.
1.5 anton 10208:
1.78 anton 10209: @cindex object
10210: @item object
10211: an instance of the data structure described by the class definition.
1.5 anton 10212:
1.78 anton 10213: @cindex instance variables
10214: @item instance variables
10215: fields of the data structure.
1.5 anton 10216:
1.78 anton 10217: @cindex selector
10218: @cindex method selector
10219: @cindex virtual function
10220: @item selector
10221: (or @emph{method selector}) a word (e.g.,
10222: @code{draw}) that performs an operation on a variety of data
10223: structures (classes). A selector describes @emph{what} operation to
10224: perform. In C++ terminology: a (pure) virtual function.
1.5 anton 10225:
1.78 anton 10226: @cindex method
10227: @item method
10228: the concrete definition that performs the operation
10229: described by the selector for a specific class. A method specifies
10230: @emph{how} the operation is performed for a specific class.
1.5 anton 10231:
1.78 anton 10232: @cindex selector invocation
10233: @cindex message send
10234: @cindex invoking a selector
10235: @item selector invocation
10236: a call of a selector. One argument of the call (the TOS (top-of-stack))
10237: is used for determining which method is used. In Smalltalk terminology:
10238: a message (consisting of the selector and the other arguments) is sent
10239: to the object.
1.5 anton 10240:
1.78 anton 10241: @cindex receiving object
10242: @item receiving object
10243: the object used for determining the method executed by a selector
10244: invocation. In the @file{objects.fs} model, it is the object that is on
10245: the TOS when the selector is invoked. (@emph{Receiving} comes from
10246: the Smalltalk @emph{message} terminology.)
1.5 anton 10247:
1.78 anton 10248: @cindex child class
10249: @cindex parent class
10250: @cindex inheritance
10251: @item child class
10252: a class that has (@emph{inherits}) all properties (instance variables,
10253: selectors, methods) from a @emph{parent class}. In Smalltalk
10254: terminology: The subclass inherits from the superclass. In C++
10255: terminology: The derived class inherits from the base class.
1.5 anton 10256:
1.78 anton 10257: @end table
1.5 anton 10258:
1.78 anton 10259: @c If you wonder about the message sending terminology, it comes from
10260: @c a time when each object had it's own task and objects communicated via
10261: @c message passing; eventually the Smalltalk developers realized that
10262: @c they can do most things through simple (indirect) calls. They kept the
10263: @c terminology.
1.5 anton 10264:
1.78 anton 10265: @c --------------------------------------------------------------
10266: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
10267: @subsection The @file{objects.fs} model
10268: @cindex objects
10269: @cindex object-oriented programming
1.26 crook 10270:
1.78 anton 10271: @cindex @file{objects.fs}
10272: @cindex @file{oof.fs}
1.26 crook 10273:
1.78 anton 10274: This section describes the @file{objects.fs} package. This material also
10275: has been published in M. Anton Ertl,
10276: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
10277: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
10278: 37--43.
10279: @c McKewan's and Zsoter's packages
1.26 crook 10280:
1.78 anton 10281: This section assumes that you have read @ref{Structures}.
1.5 anton 10282:
1.78 anton 10283: The techniques on which this model is based have been used to implement
10284: the parser generator, Gray, and have also been used in Gforth for
10285: implementing the various flavours of word lists (hashed or not,
10286: case-sensitive or not, special-purpose word lists for locals etc.).
1.5 anton 10287:
10288:
1.26 crook 10289: @menu
1.78 anton 10290: * Properties of the Objects model::
10291: * Basic Objects Usage::
10292: * The Objects base class::
10293: * Creating objects::
10294: * Object-Oriented Programming Style::
10295: * Class Binding::
10296: * Method conveniences::
10297: * Classes and Scoping::
10298: * Dividing classes::
10299: * Object Interfaces::
10300: * Objects Implementation::
10301: * Objects Glossary::
1.26 crook 10302: @end menu
1.5 anton 10303:
1.78 anton 10304: Marcel Hendrix provided helpful comments on this section.
1.5 anton 10305:
1.78 anton 10306: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
10307: @subsubsection Properties of the @file{objects.fs} model
10308: @cindex @file{objects.fs} properties
1.5 anton 10309:
1.78 anton 10310: @itemize @bullet
10311: @item
10312: It is straightforward to pass objects on the stack. Passing
10313: selectors on the stack is a little less convenient, but possible.
1.44 crook 10314:
1.78 anton 10315: @item
10316: Objects are just data structures in memory, and are referenced by their
10317: address. You can create words for objects with normal defining words
10318: like @code{constant}. Likewise, there is no difference between instance
10319: variables that contain objects and those that contain other data.
1.5 anton 10320:
1.78 anton 10321: @item
10322: Late binding is efficient and easy to use.
1.44 crook 10323:
1.78 anton 10324: @item
10325: It avoids parsing, and thus avoids problems with state-smartness
10326: and reduced extensibility; for convenience there are a few parsing
10327: words, but they have non-parsing counterparts. There are also a few
10328: defining words that parse. This is hard to avoid, because all standard
10329: defining words parse (except @code{:noname}); however, such
10330: words are not as bad as many other parsing words, because they are not
10331: state-smart.
1.5 anton 10332:
1.78 anton 10333: @item
10334: It does not try to incorporate everything. It does a few things and does
10335: them well (IMO). In particular, this model was not designed to support
10336: information hiding (although it has features that may help); you can use
10337: a separate package for achieving this.
1.5 anton 10338:
1.78 anton 10339: @item
10340: It is layered; you don't have to learn and use all features to use this
10341: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10342: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
10343: are optional and independent of each other.
1.5 anton 10344:
1.78 anton 10345: @item
10346: An implementation in ANS Forth is available.
1.5 anton 10347:
1.78 anton 10348: @end itemize
1.5 anton 10349:
1.44 crook 10350:
1.78 anton 10351: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10352: @subsubsection Basic @file{objects.fs} Usage
10353: @cindex basic objects usage
10354: @cindex objects, basic usage
1.5 anton 10355:
1.78 anton 10356: You can define a class for graphical objects like this:
1.44 crook 10357:
1.78 anton 10358: @cindex @code{class} usage
10359: @cindex @code{end-class} usage
10360: @cindex @code{selector} usage
1.5 anton 10361: @example
1.78 anton 10362: object class \ "object" is the parent class
10363: selector draw ( x y graphical -- )
10364: end-class graphical
10365: @end example
10366:
10367: This code defines a class @code{graphical} with an
10368: operation @code{draw}. We can perform the operation
10369: @code{draw} on any @code{graphical} object, e.g.:
10370:
10371: @example
10372: 100 100 t-rex draw
1.26 crook 10373: @end example
1.5 anton 10374:
1.78 anton 10375: @noindent
10376: where @code{t-rex} is a word (say, a constant) that produces a
10377: graphical object.
10378:
10379: @comment TODO add a 2nd operation eg perimeter.. and use for
10380: @comment a concrete example
1.5 anton 10381:
1.78 anton 10382: @cindex abstract class
10383: How do we create a graphical object? With the present definitions,
10384: we cannot create a useful graphical object. The class
10385: @code{graphical} describes graphical objects in general, but not
10386: any concrete graphical object type (C++ users would call it an
10387: @emph{abstract class}); e.g., there is no method for the selector
10388: @code{draw} in the class @code{graphical}.
1.5 anton 10389:
1.78 anton 10390: For concrete graphical objects, we define child classes of the
10391: class @code{graphical}, e.g.:
1.5 anton 10392:
1.78 anton 10393: @cindex @code{overrides} usage
10394: @cindex @code{field} usage in class definition
1.26 crook 10395: @example
1.78 anton 10396: graphical class \ "graphical" is the parent class
10397: cell% field circle-radius
1.5 anton 10398:
1.78 anton 10399: :noname ( x y circle -- )
10400: circle-radius @@ draw-circle ;
10401: overrides draw
1.5 anton 10402:
1.78 anton 10403: :noname ( n-radius circle -- )
10404: circle-radius ! ;
10405: overrides construct
1.5 anton 10406:
1.78 anton 10407: end-class circle
10408: @end example
1.44 crook 10409:
1.78 anton 10410: Here we define a class @code{circle} as a child of @code{graphical},
10411: with field @code{circle-radius} (which behaves just like a field
10412: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10413: for the selectors @code{draw} and @code{construct} (@code{construct} is
10414: defined in @code{object}, the parent class of @code{graphical}).
1.5 anton 10415:
1.78 anton 10416: Now we can create a circle on the heap (i.e.,
10417: @code{allocate}d memory) with:
1.44 crook 10418:
1.78 anton 10419: @cindex @code{heap-new} usage
1.5 anton 10420: @example
1.78 anton 10421: 50 circle heap-new constant my-circle
1.5 anton 10422: @end example
10423:
1.78 anton 10424: @noindent
10425: @code{heap-new} invokes @code{construct}, thus
10426: initializing the field @code{circle-radius} with 50. We can draw
10427: this new circle at (100,100) with:
1.5 anton 10428:
10429: @example
1.78 anton 10430: 100 100 my-circle draw
1.5 anton 10431: @end example
10432:
1.78 anton 10433: @cindex selector invocation, restrictions
10434: @cindex class definition, restrictions
10435: Note: You can only invoke a selector if the object on the TOS
10436: (the receiving object) belongs to the class where the selector was
10437: defined or one of its descendents; e.g., you can invoke
10438: @code{draw} only for objects belonging to @code{graphical}
10439: or its descendents (e.g., @code{circle}). Immediately before
10440: @code{end-class}, the search order has to be the same as
10441: immediately after @code{class}.
10442:
10443: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10444: @subsubsection The @file{object.fs} base class
10445: @cindex @code{object} class
10446:
10447: When you define a class, you have to specify a parent class. So how do
10448: you start defining classes? There is one class available from the start:
10449: @code{object}. It is ancestor for all classes and so is the
10450: only class that has no parent. It has two selectors: @code{construct}
10451: and @code{print}.
10452:
10453: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10454: @subsubsection Creating objects
10455: @cindex creating objects
10456: @cindex object creation
10457: @cindex object allocation options
10458:
10459: @cindex @code{heap-new} discussion
10460: @cindex @code{dict-new} discussion
10461: @cindex @code{construct} discussion
10462: You can create and initialize an object of a class on the heap with
10463: @code{heap-new} ( ... class -- object ) and in the dictionary
10464: (allocation with @code{allot}) with @code{dict-new} (
10465: ... class -- object ). Both words invoke @code{construct}, which
10466: consumes the stack items indicated by "..." above.
10467:
10468: @cindex @code{init-object} discussion
10469: @cindex @code{class-inst-size} discussion
10470: If you want to allocate memory for an object yourself, you can get its
10471: alignment and size with @code{class-inst-size 2@@} ( class --
10472: align size ). Once you have memory for an object, you can initialize
10473: it with @code{init-object} ( ... class object -- );
10474: @code{construct} does only a part of the necessary work.
10475:
10476: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10477: @subsubsection Object-Oriented Programming Style
10478: @cindex object-oriented programming style
10479: @cindex programming style, object-oriented
1.5 anton 10480:
1.78 anton 10481: This section is not exhaustive.
1.5 anton 10482:
1.78 anton 10483: @cindex stack effects of selectors
10484: @cindex selectors and stack effects
10485: In general, it is a good idea to ensure that all methods for the
10486: same selector have the same stack effect: when you invoke a selector,
10487: you often have no idea which method will be invoked, so, unless all
10488: methods have the same stack effect, you will not know the stack effect
10489: of the selector invocation.
1.5 anton 10490:
1.78 anton 10491: One exception to this rule is methods for the selector
10492: @code{construct}. We know which method is invoked, because we
10493: specify the class to be constructed at the same place. Actually, I
10494: defined @code{construct} as a selector only to give the users a
10495: convenient way to specify initialization. The way it is used, a
10496: mechanism different from selector invocation would be more natural
10497: (but probably would take more code and more space to explain).
1.5 anton 10498:
1.78 anton 10499: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10500: @subsubsection Class Binding
10501: @cindex class binding
10502: @cindex early binding
1.5 anton 10503:
1.78 anton 10504: @cindex late binding
10505: Normal selector invocations determine the method at run-time depending
10506: on the class of the receiving object. This run-time selection is called
10507: @i{late binding}.
1.5 anton 10508:
1.78 anton 10509: Sometimes it's preferable to invoke a different method. For example,
10510: you might want to use the simple method for @code{print}ing
10511: @code{object}s instead of the possibly long-winded @code{print} method
10512: of the receiver class. You can achieve this by replacing the invocation
10513: of @code{print} with:
1.5 anton 10514:
1.78 anton 10515: @cindex @code{[bind]} usage
1.5 anton 10516: @example
1.78 anton 10517: [bind] object print
1.5 anton 10518: @end example
10519:
1.78 anton 10520: @noindent
10521: in compiled code or:
10522:
10523: @cindex @code{bind} usage
1.5 anton 10524: @example
1.78 anton 10525: bind object print
1.5 anton 10526: @end example
10527:
1.78 anton 10528: @cindex class binding, alternative to
10529: @noindent
10530: in interpreted code. Alternatively, you can define the method with a
10531: name (e.g., @code{print-object}), and then invoke it through the
10532: name. Class binding is just a (often more convenient) way to achieve
10533: the same effect; it avoids name clutter and allows you to invoke
10534: methods directly without naming them first.
1.5 anton 10535:
1.78 anton 10536: @cindex superclass binding
10537: @cindex parent class binding
10538: A frequent use of class binding is this: When we define a method
10539: for a selector, we often want the method to do what the selector does
10540: in the parent class, and a little more. There is a special word for
10541: this purpose: @code{[parent]}; @code{[parent]
10542: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10543: selector}}, where @code{@emph{parent}} is the parent
10544: class of the current class. E.g., a method definition might look like:
1.44 crook 10545:
1.78 anton 10546: @cindex @code{[parent]} usage
10547: @example
10548: :noname
10549: dup [parent] foo \ do parent's foo on the receiving object
10550: ... \ do some more
10551: ; overrides foo
10552: @end example
1.6 pazsan 10553:
1.78 anton 10554: @cindex class binding as optimization
10555: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10556: March 1997), Andrew McKewan presents class binding as an optimization
10557: technique. I recommend not using it for this purpose unless you are in
10558: an emergency. Late binding is pretty fast with this model anyway, so the
10559: benefit of using class binding is small; the cost of using class binding
10560: where it is not appropriate is reduced maintainability.
1.44 crook 10561:
1.78 anton 10562: While we are at programming style questions: You should bind
10563: selectors only to ancestor classes of the receiving object. E.g., say,
10564: you know that the receiving object is of class @code{foo} or its
10565: descendents; then you should bind only to @code{foo} and its
10566: ancestors.
1.12 anton 10567:
1.78 anton 10568: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10569: @subsubsection Method conveniences
10570: @cindex method conveniences
1.44 crook 10571:
1.78 anton 10572: In a method you usually access the receiving object pretty often. If
10573: you define the method as a plain colon definition (e.g., with
10574: @code{:noname}), you may have to do a lot of stack
10575: gymnastics. To avoid this, you can define the method with @code{m:
10576: ... ;m}. E.g., you could define the method for
10577: @code{draw}ing a @code{circle} with
1.6 pazsan 10578:
1.78 anton 10579: @cindex @code{this} usage
10580: @cindex @code{m:} usage
10581: @cindex @code{;m} usage
10582: @example
10583: m: ( x y circle -- )
10584: ( x y ) this circle-radius @@ draw-circle ;m
10585: @end example
1.6 pazsan 10586:
1.78 anton 10587: @cindex @code{exit} in @code{m: ... ;m}
10588: @cindex @code{exitm} discussion
10589: @cindex @code{catch} in @code{m: ... ;m}
10590: When this method is executed, the receiver object is removed from the
10591: stack; you can access it with @code{this} (admittedly, in this
10592: example the use of @code{m: ... ;m} offers no advantage). Note
10593: that I specify the stack effect for the whole method (i.e. including
10594: the receiver object), not just for the code between @code{m:}
10595: and @code{;m}. You cannot use @code{exit} in
10596: @code{m:...;m}; instead, use
10597: @code{exitm}.@footnote{Moreover, for any word that calls
10598: @code{catch} and was defined before loading
10599: @code{objects.fs}, you have to redefine it like I redefined
10600: @code{catch}: @code{: catch this >r catch r> to-this ;}}
1.12 anton 10601:
1.78 anton 10602: @cindex @code{inst-var} usage
10603: You will frequently use sequences of the form @code{this
10604: @emph{field}} (in the example above: @code{this
10605: circle-radius}). If you use the field only in this way, you can
10606: define it with @code{inst-var} and eliminate the
10607: @code{this} before the field name. E.g., the @code{circle}
10608: class above could also be defined with:
1.6 pazsan 10609:
1.78 anton 10610: @example
10611: graphical class
10612: cell% inst-var radius
1.6 pazsan 10613:
1.78 anton 10614: m: ( x y circle -- )
10615: radius @@ draw-circle ;m
10616: overrides draw
1.6 pazsan 10617:
1.78 anton 10618: m: ( n-radius circle -- )
10619: radius ! ;m
10620: overrides construct
1.6 pazsan 10621:
1.78 anton 10622: end-class circle
10623: @end example
1.6 pazsan 10624:
1.78 anton 10625: @code{radius} can only be used in @code{circle} and its
10626: descendent classes and inside @code{m:...;m}.
1.6 pazsan 10627:
1.78 anton 10628: @cindex @code{inst-value} usage
10629: You can also define fields with @code{inst-value}, which is
10630: to @code{inst-var} what @code{value} is to
10631: @code{variable}. You can change the value of such a field with
10632: @code{[to-inst]}. E.g., we could also define the class
10633: @code{circle} like this:
1.44 crook 10634:
1.78 anton 10635: @example
10636: graphical class
10637: inst-value radius
1.6 pazsan 10638:
1.78 anton 10639: m: ( x y circle -- )
10640: radius draw-circle ;m
10641: overrides draw
1.44 crook 10642:
1.78 anton 10643: m: ( n-radius circle -- )
10644: [to-inst] radius ;m
10645: overrides construct
1.6 pazsan 10646:
1.78 anton 10647: end-class circle
10648: @end example
1.6 pazsan 10649:
1.78 anton 10650: @c !! :m is easy to confuse with m:. Another name would be better.
1.6 pazsan 10651:
1.78 anton 10652: @c Finally, you can define named methods with @code{:m}. One use of this
10653: @c feature is the definition of words that occur only in one class and are
10654: @c not intended to be overridden, but which still need method context
10655: @c (e.g., for accessing @code{inst-var}s). Another use is for methods that
10656: @c would be bound frequently, if defined anonymously.
1.6 pazsan 10657:
10658:
1.78 anton 10659: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
10660: @subsubsection Classes and Scoping
10661: @cindex classes and scoping
10662: @cindex scoping and classes
1.6 pazsan 10663:
1.78 anton 10664: Inheritance is frequent, unlike structure extension. This exacerbates
10665: the problem with the field name convention (@pxref{Structure Naming
10666: Convention}): One always has to remember in which class the field was
10667: originally defined; changing a part of the class structure would require
10668: changes for renaming in otherwise unaffected code.
1.6 pazsan 10669:
1.78 anton 10670: @cindex @code{inst-var} visibility
10671: @cindex @code{inst-value} visibility
10672: To solve this problem, I added a scoping mechanism (which was not in my
10673: original charter): A field defined with @code{inst-var} (or
10674: @code{inst-value}) is visible only in the class where it is defined and in
10675: the descendent classes of this class. Using such fields only makes
10676: sense in @code{m:}-defined methods in these classes anyway.
1.6 pazsan 10677:
1.78 anton 10678: This scoping mechanism allows us to use the unadorned field name,
10679: because name clashes with unrelated words become much less likely.
1.6 pazsan 10680:
1.78 anton 10681: @cindex @code{protected} discussion
10682: @cindex @code{private} discussion
10683: Once we have this mechanism, we can also use it for controlling the
10684: visibility of other words: All words defined after
10685: @code{protected} are visible only in the current class and its
10686: descendents. @code{public} restores the compilation
10687: (i.e. @code{current}) word list that was in effect before. If you
10688: have several @code{protected}s without an intervening
10689: @code{public} or @code{set-current}, @code{public}
10690: will restore the compilation word list in effect before the first of
10691: these @code{protected}s.
1.6 pazsan 10692:
1.78 anton 10693: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10694: @subsubsection Dividing classes
10695: @cindex Dividing classes
10696: @cindex @code{methods}...@code{end-methods}
1.6 pazsan 10697:
1.78 anton 10698: You may want to do the definition of methods separate from the
10699: definition of the class, its selectors, fields, and instance variables,
10700: i.e., separate the implementation from the definition. You can do this
10701: in the following way:
1.6 pazsan 10702:
1.78 anton 10703: @example
10704: graphical class
10705: inst-value radius
10706: end-class circle
1.6 pazsan 10707:
1.78 anton 10708: ... \ do some other stuff
1.6 pazsan 10709:
1.78 anton 10710: circle methods \ now we are ready
1.44 crook 10711:
1.78 anton 10712: m: ( x y circle -- )
10713: radius draw-circle ;m
10714: overrides draw
1.6 pazsan 10715:
1.78 anton 10716: m: ( n-radius circle -- )
10717: [to-inst] radius ;m
10718: overrides construct
1.44 crook 10719:
1.78 anton 10720: end-methods
10721: @end example
1.7 pazsan 10722:
1.78 anton 10723: You can use several @code{methods}...@code{end-methods} sections. The
10724: only things you can do to the class in these sections are: defining
10725: methods, and overriding the class's selectors. You must not define new
10726: selectors or fields.
1.7 pazsan 10727:
1.78 anton 10728: Note that you often have to override a selector before using it. In
10729: particular, you usually have to override @code{construct} with a new
10730: method before you can invoke @code{heap-new} and friends. E.g., you
10731: must not create a circle before the @code{overrides construct} sequence
10732: in the example above.
1.7 pazsan 10733:
1.78 anton 10734: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
10735: @subsubsection Object Interfaces
10736: @cindex object interfaces
10737: @cindex interfaces for objects
1.7 pazsan 10738:
1.78 anton 10739: In this model you can only call selectors defined in the class of the
10740: receiving objects or in one of its ancestors. If you call a selector
10741: with a receiving object that is not in one of these classes, the
10742: result is undefined; if you are lucky, the program crashes
10743: immediately.
1.7 pazsan 10744:
1.78 anton 10745: @cindex selectors common to hardly-related classes
10746: Now consider the case when you want to have a selector (or several)
10747: available in two classes: You would have to add the selector to a
10748: common ancestor class, in the worst case to @code{object}. You
10749: may not want to do this, e.g., because someone else is responsible for
10750: this ancestor class.
1.7 pazsan 10751:
1.78 anton 10752: The solution for this problem is interfaces. An interface is a
10753: collection of selectors. If a class implements an interface, the
10754: selectors become available to the class and its descendents. A class
10755: can implement an unlimited number of interfaces. For the problem
10756: discussed above, we would define an interface for the selector(s), and
10757: both classes would implement the interface.
1.7 pazsan 10758:
1.78 anton 10759: As an example, consider an interface @code{storage} for
10760: writing objects to disk and getting them back, and a class
10761: @code{foo} that implements it. The code would look like this:
1.7 pazsan 10762:
1.78 anton 10763: @cindex @code{interface} usage
10764: @cindex @code{end-interface} usage
10765: @cindex @code{implementation} usage
10766: @example
10767: interface
10768: selector write ( file object -- )
10769: selector read1 ( file object -- )
10770: end-interface storage
1.13 pazsan 10771:
1.78 anton 10772: bar class
10773: storage implementation
1.13 pazsan 10774:
1.78 anton 10775: ... overrides write
10776: ... overrides read1
10777: ...
10778: end-class foo
10779: @end example
1.13 pazsan 10780:
1.78 anton 10781: @noindent
10782: (I would add a word @code{read} @i{( file -- object )} that uses
10783: @code{read1} internally, but that's beyond the point illustrated
10784: here.)
1.13 pazsan 10785:
1.78 anton 10786: Note that you cannot use @code{protected} in an interface; and
10787: of course you cannot define fields.
1.13 pazsan 10788:
1.78 anton 10789: In the Neon model, all selectors are available for all classes;
10790: therefore it does not need interfaces. The price you pay in this model
10791: is slower late binding, and therefore, added complexity to avoid late
10792: binding.
1.13 pazsan 10793:
1.78 anton 10794: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10795: @subsubsection @file{objects.fs} Implementation
10796: @cindex @file{objects.fs} implementation
1.13 pazsan 10797:
1.78 anton 10798: @cindex @code{object-map} discussion
10799: An object is a piece of memory, like one of the data structures
10800: described with @code{struct...end-struct}. It has a field
10801: @code{object-map} that points to the method map for the object's
10802: class.
1.13 pazsan 10803:
1.78 anton 10804: @cindex method map
10805: @cindex virtual function table
10806: The @emph{method map}@footnote{This is Self terminology; in C++
10807: terminology: virtual function table.} is an array that contains the
10808: execution tokens (@i{xt}s) of the methods for the object's class. Each
10809: selector contains an offset into a method map.
1.13 pazsan 10810:
1.78 anton 10811: @cindex @code{selector} implementation, class
10812: @code{selector} is a defining word that uses
10813: @code{CREATE} and @code{DOES>}. The body of the
10814: selector contains the offset; the @code{DOES>} action for a
10815: class selector is, basically:
1.8 pazsan 10816:
10817: @example
1.78 anton 10818: ( object addr ) @@ over object-map @@ + @@ execute
1.13 pazsan 10819: @end example
10820:
1.78 anton 10821: Since @code{object-map} is the first field of the object, it
10822: does not generate any code. As you can see, calling a selector has a
10823: small, constant cost.
1.26 crook 10824:
1.78 anton 10825: @cindex @code{current-interface} discussion
10826: @cindex class implementation and representation
10827: A class is basically a @code{struct} combined with a method
10828: map. During the class definition the alignment and size of the class
10829: are passed on the stack, just as with @code{struct}s, so
10830: @code{field} can also be used for defining class
10831: fields. However, passing more items on the stack would be
10832: inconvenient, so @code{class} builds a data structure in memory,
10833: which is accessed through the variable
10834: @code{current-interface}. After its definition is complete, the
10835: class is represented on the stack by a pointer (e.g., as parameter for
10836: a child class definition).
1.26 crook 10837:
1.78 anton 10838: A new class starts off with the alignment and size of its parent,
10839: and a copy of the parent's method map. Defining new fields extends the
10840: size and alignment; likewise, defining new selectors extends the
10841: method map. @code{overrides} just stores a new @i{xt} in the method
10842: map at the offset given by the selector.
1.13 pazsan 10843:
1.78 anton 10844: @cindex class binding, implementation
10845: Class binding just gets the @i{xt} at the offset given by the selector
10846: from the class's method map and @code{compile,}s (in the case of
10847: @code{[bind]}) it.
1.13 pazsan 10848:
1.78 anton 10849: @cindex @code{this} implementation
10850: @cindex @code{catch} and @code{this}
10851: @cindex @code{this} and @code{catch}
10852: I implemented @code{this} as a @code{value}. At the
10853: start of an @code{m:...;m} method the old @code{this} is
10854: stored to the return stack and restored at the end; and the object on
10855: the TOS is stored @code{TO this}. This technique has one
10856: disadvantage: If the user does not leave the method via
10857: @code{;m}, but via @code{throw} or @code{exit},
10858: @code{this} is not restored (and @code{exit} may
10859: crash). To deal with the @code{throw} problem, I have redefined
10860: @code{catch} to save and restore @code{this}; the same
10861: should be done with any word that can catch an exception. As for
10862: @code{exit}, I simply forbid it (as a replacement, there is
10863: @code{exitm}).
1.13 pazsan 10864:
1.78 anton 10865: @cindex @code{inst-var} implementation
10866: @code{inst-var} is just the same as @code{field}, with
10867: a different @code{DOES>} action:
1.13 pazsan 10868: @example
1.78 anton 10869: @@ this +
1.8 pazsan 10870: @end example
1.78 anton 10871: Similar for @code{inst-value}.
1.8 pazsan 10872:
1.78 anton 10873: @cindex class scoping implementation
10874: Each class also has a word list that contains the words defined with
10875: @code{inst-var} and @code{inst-value}, and its protected
10876: words. It also has a pointer to its parent. @code{class} pushes
10877: the word lists of the class and all its ancestors onto the search order stack,
10878: and @code{end-class} drops them.
1.20 pazsan 10879:
1.78 anton 10880: @cindex interface implementation
10881: An interface is like a class without fields, parent and protected
10882: words; i.e., it just has a method map. If a class implements an
10883: interface, its method map contains a pointer to the method map of the
10884: interface. The positive offsets in the map are reserved for class
10885: methods, therefore interface map pointers have negative
10886: offsets. Interfaces have offsets that are unique throughout the
10887: system, unlike class selectors, whose offsets are only unique for the
10888: classes where the selector is available (invokable).
1.20 pazsan 10889:
1.78 anton 10890: This structure means that interface selectors have to perform one
10891: indirection more than class selectors to find their method. Their body
10892: contains the interface map pointer offset in the class method map, and
10893: the method offset in the interface method map. The
10894: @code{does>} action for an interface selector is, basically:
1.20 pazsan 10895:
10896: @example
1.78 anton 10897: ( object selector-body )
10898: 2dup selector-interface @@ ( object selector-body object interface-offset )
10899: swap object-map @@ + @@ ( object selector-body map )
10900: swap selector-offset @@ + @@ execute
1.20 pazsan 10901: @end example
10902:
1.78 anton 10903: where @code{object-map} and @code{selector-offset} are
10904: first fields and generate no code.
1.20 pazsan 10905:
1.78 anton 10906: As a concrete example, consider the following code:
1.20 pazsan 10907:
10908: @example
1.78 anton 10909: interface
10910: selector if1sel1
10911: selector if1sel2
10912: end-interface if1
1.20 pazsan 10913:
1.78 anton 10914: object class
10915: if1 implementation
10916: selector cl1sel1
10917: cell% inst-var cl1iv1
1.20 pazsan 10918:
1.78 anton 10919: ' m1 overrides construct
10920: ' m2 overrides if1sel1
10921: ' m3 overrides if1sel2
10922: ' m4 overrides cl1sel2
10923: end-class cl1
1.20 pazsan 10924:
1.78 anton 10925: create obj1 object dict-new drop
10926: create obj2 cl1 dict-new drop
10927: @end example
1.20 pazsan 10928:
1.78 anton 10929: The data structure created by this code (including the data structure
10930: for @code{object}) is shown in the
10931: @uref{objects-implementation.eps,figure}, assuming a cell size of 4.
10932: @comment TODO add this diagram..
1.20 pazsan 10933:
1.78 anton 10934: @node Objects Glossary, , Objects Implementation, Objects
10935: @subsubsection @file{objects.fs} Glossary
10936: @cindex @file{objects.fs} Glossary
1.20 pazsan 10937:
10938:
1.78 anton 10939: doc---objects-bind
10940: doc---objects-<bind>
10941: doc---objects-bind'
10942: doc---objects-[bind]
10943: doc---objects-class
10944: doc---objects-class->map
10945: doc---objects-class-inst-size
10946: doc---objects-class-override!
1.79 anton 10947: doc---objects-class-previous
10948: doc---objects-class>order
1.78 anton 10949: doc---objects-construct
10950: doc---objects-current'
10951: doc---objects-[current]
10952: doc---objects-current-interface
10953: doc---objects-dict-new
10954: doc---objects-end-class
10955: doc---objects-end-class-noname
10956: doc---objects-end-interface
10957: doc---objects-end-interface-noname
10958: doc---objects-end-methods
10959: doc---objects-exitm
10960: doc---objects-heap-new
10961: doc---objects-implementation
10962: doc---objects-init-object
10963: doc---objects-inst-value
10964: doc---objects-inst-var
10965: doc---objects-interface
10966: doc---objects-m:
10967: doc---objects-:m
10968: doc---objects-;m
10969: doc---objects-method
10970: doc---objects-methods
10971: doc---objects-object
10972: doc---objects-overrides
10973: doc---objects-[parent]
10974: doc---objects-print
10975: doc---objects-protected
10976: doc---objects-public
10977: doc---objects-selector
10978: doc---objects-this
10979: doc---objects-<to-inst>
10980: doc---objects-[to-inst]
10981: doc---objects-to-this
10982: doc---objects-xt-new
1.20 pazsan 10983:
10984:
1.78 anton 10985: @c -------------------------------------------------------------
10986: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10987: @subsection The @file{oof.fs} model
10988: @cindex oof
10989: @cindex object-oriented programming
1.20 pazsan 10990:
1.78 anton 10991: @cindex @file{objects.fs}
10992: @cindex @file{oof.fs}
1.20 pazsan 10993:
1.78 anton 10994: This section describes the @file{oof.fs} package.
1.20 pazsan 10995:
1.78 anton 10996: The package described in this section has been used in bigFORTH since 1991, and
10997: used for two large applications: a chromatographic system used to
10998: create new medicaments, and a graphic user interface library (MINOS).
1.20 pazsan 10999:
1.78 anton 11000: You can find a description (in German) of @file{oof.fs} in @cite{Object
11001: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
11002: 10(2), 1994.
1.20 pazsan 11003:
1.78 anton 11004: @menu
11005: * Properties of the OOF model::
11006: * Basic OOF Usage::
11007: * The OOF base class::
11008: * Class Declaration::
11009: * Class Implementation::
11010: @end menu
1.20 pazsan 11011:
1.78 anton 11012: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
11013: @subsubsection Properties of the @file{oof.fs} model
11014: @cindex @file{oof.fs} properties
1.20 pazsan 11015:
1.78 anton 11016: @itemize @bullet
11017: @item
11018: This model combines object oriented programming with information
11019: hiding. It helps you writing large application, where scoping is
11020: necessary, because it provides class-oriented scoping.
1.20 pazsan 11021:
1.78 anton 11022: @item
11023: Named objects, object pointers, and object arrays can be created,
11024: selector invocation uses the ``object selector'' syntax. Selector invocation
11025: to objects and/or selectors on the stack is a bit less convenient, but
11026: possible.
1.44 crook 11027:
1.78 anton 11028: @item
11029: Selector invocation and instance variable usage of the active object is
11030: straightforward, since both make use of the active object.
1.44 crook 11031:
1.78 anton 11032: @item
11033: Late binding is efficient and easy to use.
1.20 pazsan 11034:
1.78 anton 11035: @item
11036: State-smart objects parse selectors. However, extensibility is provided
11037: using a (parsing) selector @code{postpone} and a selector @code{'}.
1.20 pazsan 11038:
1.78 anton 11039: @item
11040: An implementation in ANS Forth is available.
1.20 pazsan 11041:
1.78 anton 11042: @end itemize
1.23 crook 11043:
11044:
1.78 anton 11045: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
11046: @subsubsection Basic @file{oof.fs} Usage
11047: @cindex @file{oof.fs} usage
1.23 crook 11048:
1.78 anton 11049: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
1.23 crook 11050:
1.78 anton 11051: You can define a class for graphical objects like this:
1.23 crook 11052:
1.78 anton 11053: @cindex @code{class} usage
11054: @cindex @code{class;} usage
11055: @cindex @code{method} usage
11056: @example
11057: object class graphical \ "object" is the parent class
1.139 pazsan 11058: method draw ( x y -- )
1.78 anton 11059: class;
11060: @end example
1.23 crook 11061:
1.78 anton 11062: This code defines a class @code{graphical} with an
11063: operation @code{draw}. We can perform the operation
11064: @code{draw} on any @code{graphical} object, e.g.:
1.23 crook 11065:
1.78 anton 11066: @example
11067: 100 100 t-rex draw
11068: @end example
1.23 crook 11069:
1.78 anton 11070: @noindent
11071: where @code{t-rex} is an object or object pointer, created with e.g.
11072: @code{graphical : t-rex}.
1.23 crook 11073:
1.78 anton 11074: @cindex abstract class
11075: How do we create a graphical object? With the present definitions,
11076: we cannot create a useful graphical object. The class
11077: @code{graphical} describes graphical objects in general, but not
11078: any concrete graphical object type (C++ users would call it an
11079: @emph{abstract class}); e.g., there is no method for the selector
11080: @code{draw} in the class @code{graphical}.
1.23 crook 11081:
1.78 anton 11082: For concrete graphical objects, we define child classes of the
11083: class @code{graphical}, e.g.:
1.23 crook 11084:
1.78 anton 11085: @example
11086: graphical class circle \ "graphical" is the parent class
11087: cell var circle-radius
11088: how:
11089: : draw ( x y -- )
11090: circle-radius @@ draw-circle ;
1.23 crook 11091:
1.139 pazsan 11092: : init ( n-radius -- )
1.78 anton 11093: circle-radius ! ;
11094: class;
11095: @end example
1.1 anton 11096:
1.78 anton 11097: Here we define a class @code{circle} as a child of @code{graphical},
11098: with a field @code{circle-radius}; it defines new methods for the
11099: selectors @code{draw} and @code{init} (@code{init} is defined in
11100: @code{object}, the parent class of @code{graphical}).
1.1 anton 11101:
1.78 anton 11102: Now we can create a circle in the dictionary with:
1.1 anton 11103:
1.78 anton 11104: @example
11105: 50 circle : my-circle
11106: @end example
1.21 crook 11107:
1.78 anton 11108: @noindent
11109: @code{:} invokes @code{init}, thus initializing the field
11110: @code{circle-radius} with 50. We can draw this new circle at (100,100)
11111: with:
1.1 anton 11112:
1.78 anton 11113: @example
11114: 100 100 my-circle draw
11115: @end example
1.1 anton 11116:
1.78 anton 11117: @cindex selector invocation, restrictions
11118: @cindex class definition, restrictions
11119: Note: You can only invoke a selector if the receiving object belongs to
11120: the class where the selector was defined or one of its descendents;
11121: e.g., you can invoke @code{draw} only for objects belonging to
11122: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
11123: mechanism will check if you try to invoke a selector that is not
11124: defined in this class hierarchy, so you'll get an error at compilation
11125: time.
1.1 anton 11126:
11127:
1.78 anton 11128: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
11129: @subsubsection The @file{oof.fs} base class
11130: @cindex @file{oof.fs} base class
1.1 anton 11131:
1.78 anton 11132: When you define a class, you have to specify a parent class. So how do
11133: you start defining classes? There is one class available from the start:
11134: @code{object}. You have to use it as ancestor for all classes. It is the
11135: only class that has no parent. Classes are also objects, except that
11136: they don't have instance variables; class manipulation such as
11137: inheritance or changing definitions of a class is handled through
11138: selectors of the class @code{object}.
1.1 anton 11139:
1.78 anton 11140: @code{object} provides a number of selectors:
1.1 anton 11141:
1.78 anton 11142: @itemize @bullet
11143: @item
11144: @code{class} for subclassing, @code{definitions} to add definitions
11145: later on, and @code{class?} to get type informations (is the class a
11146: subclass of the class passed on the stack?).
1.1 anton 11147:
1.78 anton 11148: doc---object-class
11149: doc---object-definitions
11150: doc---object-class?
1.1 anton 11151:
11152:
1.26 crook 11153: @item
1.78 anton 11154: @code{init} and @code{dispose} as constructor and destructor of the
11155: object. @code{init} is invocated after the object's memory is allocated,
11156: while @code{dispose} also handles deallocation. Thus if you redefine
11157: @code{dispose}, you have to call the parent's dispose with @code{super
11158: dispose}, too.
11159:
11160: doc---object-init
11161: doc---object-dispose
11162:
1.1 anton 11163:
1.26 crook 11164: @item
1.78 anton 11165: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
11166: @code{[]} to create named and unnamed objects and object arrays or
11167: object pointers.
11168:
11169: doc---object-new
11170: doc---object-new[]
11171: doc---object-:
11172: doc---object-ptr
11173: doc---object-asptr
11174: doc---object-[]
11175:
1.1 anton 11176:
1.26 crook 11177: @item
1.78 anton 11178: @code{::} and @code{super} for explicit scoping. You should use explicit
11179: scoping only for super classes or classes with the same set of instance
11180: variables. Explicitly-scoped selectors use early binding.
1.21 crook 11181:
1.78 anton 11182: doc---object-::
11183: doc---object-super
1.21 crook 11184:
11185:
1.26 crook 11186: @item
1.78 anton 11187: @code{self} to get the address of the object
1.21 crook 11188:
1.78 anton 11189: doc---object-self
1.21 crook 11190:
11191:
1.78 anton 11192: @item
11193: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
11194: pointers and instance defers.
1.21 crook 11195:
1.78 anton 11196: doc---object-bind
11197: doc---object-bound
11198: doc---object-link
11199: doc---object-is
1.21 crook 11200:
11201:
1.78 anton 11202: @item
11203: @code{'} to obtain selector tokens, @code{send} to invocate selectors
11204: form the stack, and @code{postpone} to generate selector invocation code.
1.21 crook 11205:
1.78 anton 11206: doc---object-'
11207: doc---object-postpone
1.21 crook 11208:
11209:
1.78 anton 11210: @item
11211: @code{with} and @code{endwith} to select the active object from the
11212: stack, and enable its scope. Using @code{with} and @code{endwith}
11213: also allows you to create code using selector @code{postpone} without being
11214: trapped by the state-smart objects.
1.21 crook 11215:
1.78 anton 11216: doc---object-with
11217: doc---object-endwith
1.21 crook 11218:
11219:
1.78 anton 11220: @end itemize
1.21 crook 11221:
1.78 anton 11222: @node Class Declaration, Class Implementation, The OOF base class, OOF
11223: @subsubsection Class Declaration
11224: @cindex class declaration
1.21 crook 11225:
1.78 anton 11226: @itemize @bullet
11227: @item
11228: Instance variables
1.21 crook 11229:
1.78 anton 11230: doc---oof-var
1.21 crook 11231:
11232:
1.78 anton 11233: @item
11234: Object pointers
1.21 crook 11235:
1.78 anton 11236: doc---oof-ptr
11237: doc---oof-asptr
1.21 crook 11238:
11239:
1.78 anton 11240: @item
11241: Instance defers
1.21 crook 11242:
1.78 anton 11243: doc---oof-defer
1.21 crook 11244:
11245:
1.78 anton 11246: @item
11247: Method selectors
1.21 crook 11248:
1.78 anton 11249: doc---oof-early
11250: doc---oof-method
1.21 crook 11251:
11252:
1.78 anton 11253: @item
11254: Class-wide variables
1.21 crook 11255:
1.78 anton 11256: doc---oof-static
1.21 crook 11257:
11258:
1.78 anton 11259: @item
11260: End declaration
1.1 anton 11261:
1.78 anton 11262: doc---oof-how:
11263: doc---oof-class;
1.21 crook 11264:
11265:
1.78 anton 11266: @end itemize
1.21 crook 11267:
1.78 anton 11268: @c -------------------------------------------------------------
11269: @node Class Implementation, , Class Declaration, OOF
11270: @subsubsection Class Implementation
11271: @cindex class implementation
1.21 crook 11272:
1.78 anton 11273: @c -------------------------------------------------------------
11274: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
11275: @subsection The @file{mini-oof.fs} model
11276: @cindex mini-oof
1.21 crook 11277:
1.78 anton 11278: Gforth's third object oriented Forth package is a 12-liner. It uses a
1.79 anton 11279: mixture of the @file{objects.fs} and the @file{oof.fs} syntax,
1.78 anton 11280: and reduces to the bare minimum of features. This is based on a posting
11281: of Bernd Paysan in comp.lang.forth.
1.21 crook 11282:
1.78 anton 11283: @menu
11284: * Basic Mini-OOF Usage::
11285: * Mini-OOF Example::
11286: * Mini-OOF Implementation::
11287: @end menu
1.21 crook 11288:
1.78 anton 11289: @c -------------------------------------------------------------
11290: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
11291: @subsubsection Basic @file{mini-oof.fs} Usage
11292: @cindex mini-oof usage
1.21 crook 11293:
1.78 anton 11294: There is a base class (@code{class}, which allocates one cell for the
11295: object pointer) plus seven other words: to define a method, a variable,
11296: a class; to end a class, to resolve binding, to allocate an object and
11297: to compile a class method.
11298: @comment TODO better description of the last one
1.26 crook 11299:
1.21 crook 11300:
1.78 anton 11301: doc-object
11302: doc-method
11303: doc-var
11304: doc-class
11305: doc-end-class
11306: doc-defines
11307: doc-new
11308: doc-::
1.21 crook 11309:
11310:
11311:
1.78 anton 11312: @c -------------------------------------------------------------
11313: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
11314: @subsubsection Mini-OOF Example
11315: @cindex mini-oof example
1.1 anton 11316:
1.78 anton 11317: A short example shows how to use this package. This example, in slightly
11318: extended form, is supplied as @file{moof-exm.fs}
11319: @comment TODO could flesh this out with some comments from the Forthwrite article
1.20 pazsan 11320:
1.26 crook 11321: @example
1.78 anton 11322: object class
11323: method init
11324: method draw
11325: end-class graphical
1.26 crook 11326: @end example
1.20 pazsan 11327:
1.78 anton 11328: This code defines a class @code{graphical} with an
11329: operation @code{draw}. We can perform the operation
11330: @code{draw} on any @code{graphical} object, e.g.:
1.20 pazsan 11331:
1.26 crook 11332: @example
1.78 anton 11333: 100 100 t-rex draw
1.26 crook 11334: @end example
1.12 anton 11335:
1.78 anton 11336: where @code{t-rex} is an object or object pointer, created with e.g.
11337: @code{graphical new Constant t-rex}.
1.12 anton 11338:
1.78 anton 11339: For concrete graphical objects, we define child classes of the
11340: class @code{graphical}, e.g.:
1.12 anton 11341:
1.26 crook 11342: @example
11343: graphical class
1.78 anton 11344: cell var circle-radius
11345: end-class circle \ "graphical" is the parent class
1.12 anton 11346:
1.78 anton 11347: :noname ( x y -- )
11348: circle-radius @@ draw-circle ; circle defines draw
11349: :noname ( r -- )
11350: circle-radius ! ; circle defines init
11351: @end example
1.12 anton 11352:
1.78 anton 11353: There is no implicit init method, so we have to define one. The creation
11354: code of the object now has to call init explicitely.
1.21 crook 11355:
1.78 anton 11356: @example
11357: circle new Constant my-circle
11358: 50 my-circle init
1.12 anton 11359: @end example
11360:
1.78 anton 11361: It is also possible to add a function to create named objects with
11362: automatic call of @code{init}, given that all objects have @code{init}
11363: on the same place:
1.38 anton 11364:
1.78 anton 11365: @example
11366: : new: ( .. o "name" -- )
11367: new dup Constant init ;
11368: 80 circle new: large-circle
11369: @end example
1.12 anton 11370:
1.78 anton 11371: We can draw this new circle at (100,100) with:
1.12 anton 11372:
1.78 anton 11373: @example
11374: 100 100 my-circle draw
11375: @end example
1.12 anton 11376:
1.78 anton 11377: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
11378: @subsubsection @file{mini-oof.fs} Implementation
1.12 anton 11379:
1.78 anton 11380: Object-oriented systems with late binding typically use a
11381: ``vtable''-approach: the first variable in each object is a pointer to a
11382: table, which contains the methods as function pointers. The vtable
11383: may also contain other information.
1.12 anton 11384:
1.79 anton 11385: So first, let's declare selectors:
1.37 anton 11386:
11387: @example
1.79 anton 11388: : method ( m v "name" -- m' v ) Create over , swap cell+ swap
1.78 anton 11389: DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
11390: @end example
1.37 anton 11391:
1.79 anton 11392: During selector declaration, the number of selectors and instance
11393: variables is on the stack (in address units). @code{method} creates one
11394: selector and increments the selector number. To execute a selector, it
1.78 anton 11395: takes the object, fetches the vtable pointer, adds the offset, and
1.79 anton 11396: executes the method @i{xt} stored there. Each selector takes the object
11397: it is invoked with as top of stack parameter; it passes the parameters
11398: (including the object) unchanged to the appropriate method which should
1.78 anton 11399: consume that object.
1.37 anton 11400:
1.78 anton 11401: Now, we also have to declare instance variables
1.37 anton 11402:
1.78 anton 11403: @example
1.79 anton 11404: : var ( m v size "name" -- m v' ) Create over , +
1.78 anton 11405: DOES> ( o -- addr ) @@ + ;
1.37 anton 11406: @end example
11407:
1.78 anton 11408: As before, a word is created with the current offset. Instance
11409: variables can have different sizes (cells, floats, doubles, chars), so
11410: all we do is take the size and add it to the offset. If your machine
11411: has alignment restrictions, put the proper @code{aligned} or
11412: @code{faligned} before the variable, to adjust the variable
11413: offset. That's why it is on the top of stack.
1.37 anton 11414:
1.78 anton 11415: We need a starting point (the base object) and some syntactic sugar:
1.37 anton 11416:
1.78 anton 11417: @example
11418: Create object 1 cells , 2 cells ,
1.79 anton 11419: : class ( class -- class selectors vars ) dup 2@@ ;
1.78 anton 11420: @end example
1.12 anton 11421:
1.78 anton 11422: For inheritance, the vtable of the parent object has to be
11423: copied when a new, derived class is declared. This gives all the
11424: methods of the parent class, which can be overridden, though.
1.12 anton 11425:
1.78 anton 11426: @example
1.79 anton 11427: : end-class ( class selectors vars "name" -- )
1.78 anton 11428: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
11429: cell+ dup cell+ r> rot @@ 2 cells /string move ;
11430: @end example
1.12 anton 11431:
1.78 anton 11432: The first line creates the vtable, initialized with
11433: @code{noop}s. The second line is the inheritance mechanism, it
11434: copies the xts from the parent vtable.
1.12 anton 11435:
1.78 anton 11436: We still have no way to define new methods, let's do that now:
1.12 anton 11437:
1.26 crook 11438: @example
1.79 anton 11439: : defines ( xt class "name" -- ) ' >body @@ + ! ;
1.78 anton 11440: @end example
1.12 anton 11441:
1.78 anton 11442: To allocate a new object, we need a word, too:
1.12 anton 11443:
1.78 anton 11444: @example
11445: : new ( class -- o ) here over @@ allot swap over ! ;
1.12 anton 11446: @end example
11447:
1.78 anton 11448: Sometimes derived classes want to access the method of the
11449: parent object. There are two ways to achieve this with Mini-OOF:
11450: first, you could use named words, and second, you could look up the
11451: vtable of the parent object.
1.12 anton 11452:
1.78 anton 11453: @example
11454: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
11455: @end example
1.12 anton 11456:
11457:
1.78 anton 11458: Nothing can be more confusing than a good example, so here is
11459: one. First let's declare a text object (called
11460: @code{button}), that stores text and position:
1.12 anton 11461:
1.78 anton 11462: @example
11463: object class
11464: cell var text
11465: cell var len
11466: cell var x
11467: cell var y
11468: method init
11469: method draw
11470: end-class button
11471: @end example
1.12 anton 11472:
1.78 anton 11473: @noindent
11474: Now, implement the two methods, @code{draw} and @code{init}:
1.21 crook 11475:
1.26 crook 11476: @example
1.78 anton 11477: :noname ( o -- )
11478: >r r@@ x @@ r@@ y @@ at-xy r@@ text @@ r> len @@ type ;
11479: button defines draw
11480: :noname ( addr u o -- )
11481: >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
11482: button defines init
1.26 crook 11483: @end example
1.12 anton 11484:
1.78 anton 11485: @noindent
11486: To demonstrate inheritance, we define a class @code{bold-button}, with no
1.79 anton 11487: new data and no new selectors:
1.78 anton 11488:
11489: @example
11490: button class
11491: end-class bold-button
1.12 anton 11492:
1.78 anton 11493: : bold 27 emit ." [1m" ;
11494: : normal 27 emit ." [0m" ;
11495: @end example
1.1 anton 11496:
1.78 anton 11497: @noindent
11498: The class @code{bold-button} has a different draw method to
11499: @code{button}, but the new method is defined in terms of the draw method
11500: for @code{button}:
1.20 pazsan 11501:
1.78 anton 11502: @example
11503: :noname bold [ button :: draw ] normal ; bold-button defines draw
11504: @end example
1.21 crook 11505:
1.78 anton 11506: @noindent
1.79 anton 11507: Finally, create two objects and apply selectors:
1.21 crook 11508:
1.26 crook 11509: @example
1.78 anton 11510: button new Constant foo
11511: s" thin foo" foo init
11512: page
11513: foo draw
11514: bold-button new Constant bar
11515: s" fat bar" bar init
11516: 1 bar y !
11517: bar draw
1.26 crook 11518: @end example
1.21 crook 11519:
11520:
1.78 anton 11521: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11522: @subsection Comparison with other object models
11523: @cindex comparison of object models
11524: @cindex object models, comparison
11525:
11526: Many object-oriented Forth extensions have been proposed (@cite{A survey
11527: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11528: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11529: relation of the object models described here to two well-known and two
11530: closely-related (by the use of method maps) models. Andras Zsoter
11531: helped us with this section.
11532:
11533: @cindex Neon model
11534: The most popular model currently seems to be the Neon model (see
11535: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11536: 1997) by Andrew McKewan) but this model has a number of limitations
11537: @footnote{A longer version of this critique can be
11538: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11539: Dimensions, May 1997) by Anton Ertl.}:
11540:
11541: @itemize @bullet
11542: @item
11543: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11544: to pass objects on the stack.
1.21 crook 11545:
1.78 anton 11546: @item
11547: It requires that the selector parses the input stream (at
1.79 anton 11548: compile time); this leads to reduced extensibility and to bugs that are
1.78 anton 11549: hard to find.
1.21 crook 11550:
1.78 anton 11551: @item
1.79 anton 11552: It allows using every selector on every object; this eliminates the
11553: need for interfaces, but makes it harder to create efficient
11554: implementations.
1.78 anton 11555: @end itemize
1.21 crook 11556:
1.78 anton 11557: @cindex Pountain's object-oriented model
11558: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11559: Press, London, 1987) by Dick Pountain. However, it is not really about
11560: object-oriented programming, because it hardly deals with late
11561: binding. Instead, it focuses on features like information hiding and
11562: overloading that are characteristic of modular languages like Ada (83).
1.26 crook 11563:
1.78 anton 11564: @cindex Zsoter's object-oriented model
1.79 anton 11565: In @uref{http://www.forth.org/oopf.html, Does late binding have to be
11566: slow?} (Forth Dimensions 18(1) 1996, pages 31-35) Andras Zsoter
11567: describes a model that makes heavy use of an active object (like
11568: @code{this} in @file{objects.fs}): The active object is not only used
11569: for accessing all fields, but also specifies the receiving object of
11570: every selector invocation; you have to change the active object
11571: explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
11572: changes more or less implicitly at @code{m: ... ;m}. Such a change at
11573: the method entry point is unnecessary with Zsoter's model, because the
11574: receiving object is the active object already. On the other hand, the
11575: explicit change is absolutely necessary in that model, because otherwise
11576: no one could ever change the active object. An ANS Forth implementation
11577: of this model is available through
11578: @uref{http://www.forth.org/oopf.html}.
1.21 crook 11579:
1.78 anton 11580: @cindex @file{oof.fs}, differences to other models
11581: The @file{oof.fs} model combines information hiding and overloading
11582: resolution (by keeping names in various word lists) with object-oriented
11583: programming. It sets the active object implicitly on method entry, but
11584: also allows explicit changing (with @code{>o...o>} or with
11585: @code{with...endwith}). It uses parsing and state-smart objects and
11586: classes for resolving overloading and for early binding: the object or
11587: class parses the selector and determines the method from this. If the
11588: selector is not parsed by an object or class, it performs a call to the
11589: selector for the active object (late binding), like Zsoter's model.
11590: Fields are always accessed through the active object. The big
11591: disadvantage of this model is the parsing and the state-smartness, which
11592: reduces extensibility and increases the opportunities for subtle bugs;
11593: essentially, you are only safe if you never tick or @code{postpone} an
11594: object or class (Bernd disagrees, but I (Anton) am not convinced).
1.21 crook 11595:
1.78 anton 11596: @cindex @file{mini-oof.fs}, differences to other models
11597: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11598: version of the @file{objects.fs} model, but syntactically it is a
11599: mixture of the @file{objects.fs} and @file{oof.fs} models.
1.21 crook 11600:
11601:
1.78 anton 11602: @c -------------------------------------------------------------
1.150 anton 11603: @node Programming Tools, C Interface, Object-oriented Forth, Words
1.78 anton 11604: @section Programming Tools
11605: @cindex programming tools
1.21 crook 11606:
1.78 anton 11607: @c !! move this and assembler down below OO stuff.
1.21 crook 11608:
1.78 anton 11609: @menu
1.150 anton 11610: * Examining:: Data and Code.
11611: * Forgetting words:: Usually before reloading.
1.78 anton 11612: * Debugging:: Simple and quick.
11613: * Assertions:: Making your programs self-checking.
11614: * Singlestep Debugger:: Executing your program word by word.
11615: @end menu
1.21 crook 11616:
1.78 anton 11617: @node Examining, Forgetting words, Programming Tools, Programming Tools
11618: @subsection Examining data and code
11619: @cindex examining data and code
11620: @cindex data examination
11621: @cindex code examination
1.44 crook 11622:
1.78 anton 11623: The following words inspect the stack non-destructively:
1.21 crook 11624:
1.78 anton 11625: doc-.s
11626: doc-f.s
1.158 anton 11627: doc-maxdepth-.s
1.44 crook 11628:
1.78 anton 11629: There is a word @code{.r} but it does @i{not} display the return stack!
11630: It is used for formatted numeric output (@pxref{Simple numeric output}).
1.21 crook 11631:
1.78 anton 11632: doc-depth
11633: doc-fdepth
11634: doc-clearstack
1.124 anton 11635: doc-clearstacks
1.21 crook 11636:
1.78 anton 11637: The following words inspect memory.
1.21 crook 11638:
1.78 anton 11639: doc-?
11640: doc-dump
1.21 crook 11641:
1.78 anton 11642: And finally, @code{see} allows to inspect code:
1.21 crook 11643:
1.78 anton 11644: doc-see
11645: doc-xt-see
1.111 anton 11646: doc-simple-see
11647: doc-simple-see-range
1.21 crook 11648:
1.78 anton 11649: @node Forgetting words, Debugging, Examining, Programming Tools
11650: @subsection Forgetting words
11651: @cindex words, forgetting
11652: @cindex forgeting words
1.21 crook 11653:
1.78 anton 11654: @c anton: other, maybe better places for this subsection: Defining Words;
11655: @c Dictionary allocation. At least a reference should be there.
1.21 crook 11656:
1.78 anton 11657: Forth allows you to forget words (and everything that was alloted in the
11658: dictonary after them) in a LIFO manner.
1.21 crook 11659:
1.78 anton 11660: doc-marker
1.21 crook 11661:
1.78 anton 11662: The most common use of this feature is during progam development: when
11663: you change a source file, forget all the words it defined and load it
11664: again (since you also forget everything defined after the source file
11665: was loaded, you have to reload that, too). Note that effects like
11666: storing to variables and destroyed system words are not undone when you
11667: forget words. With a system like Gforth, that is fast enough at
11668: starting up and compiling, I find it more convenient to exit and restart
11669: Gforth, as this gives me a clean slate.
1.21 crook 11670:
1.78 anton 11671: Here's an example of using @code{marker} at the start of a source file
11672: that you are debugging; it ensures that you only ever have one copy of
11673: the file's definitions compiled at any time:
1.21 crook 11674:
1.78 anton 11675: @example
11676: [IFDEF] my-code
11677: my-code
11678: [ENDIF]
1.26 crook 11679:
1.78 anton 11680: marker my-code
11681: init-included-files
1.21 crook 11682:
1.78 anton 11683: \ .. definitions start here
11684: \ .
11685: \ .
11686: \ end
11687: @end example
1.21 crook 11688:
1.26 crook 11689:
1.78 anton 11690: @node Debugging, Assertions, Forgetting words, Programming Tools
11691: @subsection Debugging
11692: @cindex debugging
1.21 crook 11693:
1.78 anton 11694: Languages with a slow edit/compile/link/test development loop tend to
11695: require sophisticated tracing/stepping debuggers to facilate debugging.
1.21 crook 11696:
1.78 anton 11697: A much better (faster) way in fast-compiling languages is to add
11698: printing code at well-selected places, let the program run, look at
11699: the output, see where things went wrong, add more printing code, etc.,
11700: until the bug is found.
1.21 crook 11701:
1.78 anton 11702: The simple debugging aids provided in @file{debugs.fs}
11703: are meant to support this style of debugging.
1.21 crook 11704:
1.78 anton 11705: The word @code{~~} prints debugging information (by default the source
11706: location and the stack contents). It is easy to insert. If you use Emacs
11707: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
11708: query-replace them with nothing). The deferred words
1.101 anton 11709: @code{printdebugdata} and @code{.debugline} control the output of
1.78 anton 11710: @code{~~}. The default source location output format works well with
11711: Emacs' compilation mode, so you can step through the program at the
11712: source level using @kbd{C-x `} (the advantage over a stepping debugger
11713: is that you can step in any direction and you know where the crash has
11714: happened or where the strange data has occurred).
1.21 crook 11715:
1.78 anton 11716: doc-~~
11717: doc-printdebugdata
1.101 anton 11718: doc-.debugline
1.21 crook 11719:
1.106 anton 11720: @cindex filenames in @code{~~} output
11721: @code{~~} (and assertions) will usually print the wrong file name if a
11722: marker is executed in the same file after their occurance. They will
11723: print @samp{*somewhere*} as file name if a marker is executed in the
11724: same file before their occurance.
11725:
11726:
1.78 anton 11727: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
11728: @subsection Assertions
11729: @cindex assertions
1.21 crook 11730:
1.78 anton 11731: It is a good idea to make your programs self-checking, especially if you
11732: make an assumption that may become invalid during maintenance (for
11733: example, that a certain field of a data structure is never zero). Gforth
11734: supports @dfn{assertions} for this purpose. They are used like this:
1.21 crook 11735:
11736: @example
1.78 anton 11737: assert( @i{flag} )
1.26 crook 11738: @end example
11739:
1.78 anton 11740: The code between @code{assert(} and @code{)} should compute a flag, that
11741: should be true if everything is alright and false otherwise. It should
11742: not change anything else on the stack. The overall stack effect of the
11743: assertion is @code{( -- )}. E.g.
1.21 crook 11744:
1.26 crook 11745: @example
1.78 anton 11746: assert( 1 1 + 2 = ) \ what we learn in school
11747: assert( dup 0<> ) \ assert that the top of stack is not zero
11748: assert( false ) \ this code should not be reached
1.21 crook 11749: @end example
11750:
1.78 anton 11751: The need for assertions is different at different times. During
11752: debugging, we want more checking, in production we sometimes care more
11753: for speed. Therefore, assertions can be turned off, i.e., the assertion
11754: becomes a comment. Depending on the importance of an assertion and the
11755: time it takes to check it, you may want to turn off some assertions and
11756: keep others turned on. Gforth provides several levels of assertions for
11757: this purpose:
11758:
11759:
11760: doc-assert0(
11761: doc-assert1(
11762: doc-assert2(
11763: doc-assert3(
11764: doc-assert(
11765: doc-)
1.21 crook 11766:
11767:
1.78 anton 11768: The variable @code{assert-level} specifies the highest assertions that
11769: are turned on. I.e., at the default @code{assert-level} of one,
11770: @code{assert0(} and @code{assert1(} assertions perform checking, while
11771: @code{assert2(} and @code{assert3(} assertions are treated as comments.
1.26 crook 11772:
1.78 anton 11773: The value of @code{assert-level} is evaluated at compile-time, not at
11774: run-time. Therefore you cannot turn assertions on or off at run-time;
11775: you have to set the @code{assert-level} appropriately before compiling a
11776: piece of code. You can compile different pieces of code at different
11777: @code{assert-level}s (e.g., a trusted library at level 1 and
11778: newly-written code at level 3).
1.26 crook 11779:
11780:
1.78 anton 11781: doc-assert-level
1.26 crook 11782:
11783:
1.78 anton 11784: If an assertion fails, a message compatible with Emacs' compilation mode
11785: is produced and the execution is aborted (currently with @code{ABORT"}.
11786: If there is interest, we will introduce a special throw code. But if you
11787: intend to @code{catch} a specific condition, using @code{throw} is
11788: probably more appropriate than an assertion).
1.106 anton 11789:
11790: @cindex filenames in assertion output
11791: Assertions (and @code{~~}) will usually print the wrong file name if a
11792: marker is executed in the same file after their occurance. They will
11793: print @samp{*somewhere*} as file name if a marker is executed in the
11794: same file before their occurance.
1.44 crook 11795:
1.78 anton 11796: Definitions in ANS Forth for these assertion words are provided
11797: in @file{compat/assert.fs}.
1.26 crook 11798:
1.44 crook 11799:
1.78 anton 11800: @node Singlestep Debugger, , Assertions, Programming Tools
11801: @subsection Singlestep Debugger
11802: @cindex singlestep Debugger
11803: @cindex debugging Singlestep
1.44 crook 11804:
1.159 anton 11805: The singlestep debugger works only with the engine @code{gforth-ditc}.
1.112 anton 11806:
1.78 anton 11807: When you create a new word there's often the need to check whether it
11808: behaves correctly or not. You can do this by typing @code{dbg
11809: badword}. A debug session might look like this:
1.26 crook 11810:
1.78 anton 11811: @example
11812: : badword 0 DO i . LOOP ; ok
11813: 2 dbg badword
11814: : badword
11815: Scanning code...
1.44 crook 11816:
1.78 anton 11817: Nesting debugger ready!
1.44 crook 11818:
1.78 anton 11819: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
11820: 400D4740 8049F68 DO -> [ 0 ]
11821: 400D4744 804A0C8 i -> [ 1 ] 00000
11822: 400D4748 400C5E60 . -> 0 [ 0 ]
11823: 400D474C 8049D0C LOOP -> [ 0 ]
11824: 400D4744 804A0C8 i -> [ 1 ] 00001
11825: 400D4748 400C5E60 . -> 1 [ 0 ]
11826: 400D474C 8049D0C LOOP -> [ 0 ]
11827: 400D4758 804B384 ; -> ok
11828: @end example
1.21 crook 11829:
1.78 anton 11830: Each line displayed is one step. You always have to hit return to
11831: execute the next word that is displayed. If you don't want to execute
11832: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
11833: an overview what keys are available:
1.44 crook 11834:
1.78 anton 11835: @table @i
1.44 crook 11836:
1.78 anton 11837: @item @key{RET}
11838: Next; Execute the next word.
1.21 crook 11839:
1.78 anton 11840: @item n
11841: Nest; Single step through next word.
1.44 crook 11842:
1.78 anton 11843: @item u
11844: Unnest; Stop debugging and execute rest of word. If we got to this word
11845: with nest, continue debugging with the calling word.
1.44 crook 11846:
1.78 anton 11847: @item d
11848: Done; Stop debugging and execute rest.
1.21 crook 11849:
1.78 anton 11850: @item s
11851: Stop; Abort immediately.
1.44 crook 11852:
1.78 anton 11853: @end table
1.44 crook 11854:
1.78 anton 11855: Debugging large application with this mechanism is very difficult, because
11856: you have to nest very deeply into the program before the interesting part
11857: begins. This takes a lot of time.
1.26 crook 11858:
1.78 anton 11859: To do it more directly put a @code{BREAK:} command into your source code.
11860: When program execution reaches @code{BREAK:} the single step debugger is
11861: invoked and you have all the features described above.
1.44 crook 11862:
1.78 anton 11863: If you have more than one part to debug it is useful to know where the
11864: program has stopped at the moment. You can do this by the
11865: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
11866: string is typed out when the ``breakpoint'' is reached.
1.44 crook 11867:
1.26 crook 11868:
1.78 anton 11869: doc-dbg
11870: doc-break:
11871: doc-break"
1.44 crook 11872:
1.150 anton 11873: @c ------------------------------------------------------------
11874: @node C Interface, Assembler and Code Words, Programming Tools, Words
11875: @section C Interface
11876: @cindex C interface
11877: @cindex foreign language interface
11878: @cindex interface to C functions
11879:
1.178 anton 11880: Note that the C interface is not yet complete; callbacks are missing,
11881: as well as a way of declaring structs, unions, and their fields.
1.150 anton 11882:
11883: @menu
11884: * Calling C Functions::
11885: * Declaring C Functions::
1.180 anton 11886: * Calling C function pointers::
1.150 anton 11887: * Callbacks::
1.178 anton 11888: * C interface internals::
1.155 anton 11889: * Low-Level C Interface Words::
1.150 anton 11890: @end menu
11891:
1.151 pazsan 11892: @node Calling C Functions, Declaring C Functions, C Interface, C Interface
1.150 anton 11893: @subsection Calling C functions
1.155 anton 11894: @cindex C functions, calls to
11895: @cindex calling C functions
1.150 anton 11896:
1.151 pazsan 11897: Once a C function is declared (see @pxref{Declaring C Functions}), you
1.150 anton 11898: can call it as follows: You push the arguments on the stack(s), and
11899: then call the word for the C function. The arguments have to be
11900: pushed in the same order as the arguments appear in the C
11901: documentation (i.e., the first argument is deepest on the stack).
11902: Integer and pointer arguments have to be pushed on the data stack,
11903: floating-point arguments on the FP stack; these arguments are consumed
1.155 anton 11904: by the called C function.
1.150 anton 11905:
1.155 anton 11906: On returning from the C function, the return value, if any, resides on
11907: the appropriate stack: an integer return value is pushed on the data
11908: stack, an FP return value on the FP stack, and a void return value
11909: results in not pushing anything. Note that most C functions have a
11910: return value, even if that is often not used in C; in Forth, you have
11911: to @code{drop} this return value explicitly if you do not use it.
1.150 anton 11912:
1.177 anton 11913: The C interface automatically converts between the C type and the
11914: Forth type as necessary, on a best-effort basis (in some cases, there
11915: may be some loss).
1.150 anton 11916:
11917: As an example, consider the POSIX function @code{lseek()}:
11918:
11919: @example
11920: off_t lseek(int fd, off_t offset, int whence);
11921: @end example
11922:
11923: This function takes three integer arguments, and returns an integer
11924: argument, so a Forth call for setting the current file offset to the
11925: start of the file could look like this:
11926:
11927: @example
11928: fd @@ 0 SEEK_SET lseek -1 = if
11929: ... \ error handling
11930: then
11931: @end example
11932:
11933: You might be worried that an @code{off_t} does not fit into a cell, so
11934: you could not pass larger offsets to lseek, and might get only a part
1.155 anton 11935: of the return values. In that case, in your declaration of the
11936: function (@pxref{Declaring C Functions}) you should declare it to use
11937: double-cells for the off_t argument and return value, and maybe give
11938: the resulting Forth word a different name, like @code{dlseek}; the
11939: result could be called like this:
1.150 anton 11940:
11941: @example
11942: fd @@ 0. SEEK_SET dlseek -1. d= if
11943: ... \ error handling
11944: then
11945: @end example
11946:
11947: Passing and returning structs or unions is currently not supported by
11948: our interface@footnote{If you know the calling convention of your C
11949: compiler, you usually can call such functions in some way, but that
11950: way is usually not portable between platforms, and sometimes not even
11951: between C compilers.}.
11952:
1.177 anton 11953: Calling functions with a variable number of arguments (@emph{variadic}
11954: functions, e.g., @code{printf()}) is only supported by having you
11955: declare one function-calling word for each argument pattern, and
11956: calling the appropriate word for the desired pattern.
11957:
1.150 anton 11958:
1.155 anton 11959:
1.180 anton 11960: @node Declaring C Functions, Calling C function pointers, Calling C Functions, C Interface
1.150 anton 11961: @subsection Declaring C Functions
1.155 anton 11962: @cindex C functions, declarations
11963: @cindex declaring C functions
1.150 anton 11964:
11965: Before you can call @code{lseek} or @code{dlseek}, you have to declare
1.177 anton 11966: it. The declaration consists of two parts:
11967:
11968: @table @b
11969:
11970: @item The C part
1.179 anton 11971: is the C declaration of the function, or more typically and portably,
11972: a C-style @code{#include} of a file that contains the declaration of
11973: the C function.
1.177 anton 11974:
11975: @item The Forth part
11976: declares the Forth types of the parameters and the Forth word name
11977: corresponding to the C function.
11978:
11979: @end table
11980:
11981: For the words @code{lseek} and @code{dlseek} mentioned earlier, the
11982: declarations are:
11983:
11984: @example
11985: \c #define _FILE_OFFSET_BITS 64
11986: \c #include <sys/types.h>
11987: \c #include <unistd.h>
11988: c-function lseek lseek n n n -- n
11989: c-function dlseek lseek n d n -- d
11990: @end example
11991:
1.178 anton 11992: The C part of the declarations is prefixed by @code{\c}, and the rest
1.177 anton 11993: of the line is ordinary C code. You can use as many lines of C
11994: declarations as you like, and they are visible for all further
11995: function declarations.
11996:
11997: The Forth part declares each interface word with @code{c-function},
11998: followed by the Forth name of the word, the C name of the called
11999: function, and the stack effect of the word. The stack effect contains
1.178 anton 12000: an arbitrary number of types of parameters, then @code{--}, and then
1.177 anton 12001: exactly one type for the return value. The possible types are:
12002:
12003: @table @code
12004:
12005: @item n
12006: single-cell integer
12007:
12008: @item a
12009: address (single-cell)
12010:
12011: @item d
12012: double-cell integer
12013:
12014: @item r
12015: floating-point value
12016:
12017: @item func
12018: C function pointer
12019:
12020: @item void
12021: no value (used as return type for void functions)
12022:
12023: @end table
12024:
12025: @cindex variadic C functions
12026:
12027: To deal with variadic C functions, you can declare one Forth word for
12028: every pattern you want to use, e.g.:
12029:
12030: @example
12031: \c #include <stdio.h>
12032: c-function printf-nr printf a n r -- n
12033: c-function printf-rn printf a r n -- n
12034: @end example
12035:
12036: Note that with C functions declared as variadic (or if you don't
12037: provide a prototype), the C interface has no C type to convert to, so
12038: no automatic conversion happens, which may lead to portability
12039: problems in some cases. In such cases you can perform the conversion
12040: explicitly on the C level, e.g., as follows:
12041:
12042: @example
1.178 anton 12043: \c #define printfll(s,ll) printf(s,(long long)ll)
12044: c-function printfll printfll a n -- n
1.177 anton 12045: @end example
12046:
12047: Here, instead of calling @code{printf()} directly, we define a macro
1.178 anton 12048: that casts (converts) the Forth single-cell integer into a
12049: C @code{long long} before calling @code{printf()}.
1.177 anton 12050:
12051: doc-\c
12052: doc-c-function
12053:
12054: In order to work, this C interface invokes GCC at run-time and uses
1.178 anton 12055: dynamic linking. If these features are not available, there are
12056: other, less convenient and less portable C interfaces in @file{lib.fs}
12057: and @file{oldlib.fs}. These interfaces are mostly undocumented and
12058: mostly incompatible with each other and with the documented C
12059: interface; you can find some examples for the @file{lib.fs} interface
12060: in @file{lib.fs}.
1.177 anton 12061:
12062:
1.180 anton 12063: @node Calling C function pointers, Callbacks, Declaring C Functions, C Interface
12064: @subsection Calling C function pointers from Forth
12065: @cindex C function pointers, calling from Forth
1.177 anton 12066:
1.180 anton 12067: If you come across a C function pointer (e.g., in some C-constructed
12068: structure) and want to call it from your Forth program, you can also
12069: use the features explained until now to achieve that, as follows:
1.150 anton 12070:
1.180 anton 12071: Let us assume that there is a C function pointer type @code{func1}
12072: defined in some header file @file{func1.h}, and you know that these
12073: functions take one integer argument and return an integer result; and
12074: you want to call functions through such pointers. Just define
1.155 anton 12075:
1.180 anton 12076: @example
12077: \c #include <func1.h>
12078: \c #define call_func1(par1,fptr) ((func1)fptr)(par1)
12079: c-function call-func1 call_func1 n func -- n
12080: @end example
12081:
12082: and then you can call a function pointed to by, say @code{func1a} as
12083: follows:
12084:
12085: @example
12086: -5 func1a call-func1 .
12087: @end example
12088:
12089: In the C part, @code{call_func} is defined as a macro to avoid having
12090: to declare the exact parameter and return types, so the C compiler
12091: knows them from the declaration of @code{func1}.
12092:
12093: The Forth word @code{call-func1} is similar to @code{execute}, except
12094: that it takes a C @code{func1} pointer instead of a Forth execution
12095: token, and it is specific to @code{func1} pointers. For each type of
12096: function pointer you want to call from Forth, you have to define
12097: a separate calling word.
12098:
12099:
12100: @node Callbacks, C interface internals, Calling C function pointers, C Interface
1.150 anton 12101: @subsection Callbacks
1.155 anton 12102: @cindex Callback functions written in Forth
12103: @cindex C function pointers to Forth words
12104:
1.177 anton 12105: Callbacks are not yet supported by the documented C interface. You
12106: can use the undocumented @file{lib.fs} interface for callbacks.
12107:
1.155 anton 12108: In some cases you have to pass a function pointer to a C function,
12109: i.e., the library wants to call back to your application (and the
12110: pointed-to function is called a callback function). You can pass the
12111: address of an existing C function (that you get with @code{lib-sym},
12112: @pxref{Low-Level C Interface Words}), but if there is no appropriate C
12113: function, you probably want to define the function as a Forth word.
12114:
12115: @c I don't understand the existing callback interface from the example - anton
12116:
1.165 anton 12117:
12118: @c > > Und dann gibt's noch die fptr-Deklaration, die einem
12119: @c > > C-Funktionspointer entspricht (Deklaration gleich wie bei
12120: @c > > Library-Funktionen, nur ohne den C-Namen, Aufruf mit der
12121: @c > > C-Funktionsadresse auf dem TOS).
12122: @c >
12123: @c > Ja, da bin ich dann ausgestiegen, weil ich aus dem Beispiel nicht
12124: @c > gesehen habe, wozu das gut ist.
12125: @c
12126: @c Irgendwie muss ich den Callback ja testen. Und es soll ja auch
12127: @c vorkommen, dass man von irgendwelchen kranken Interfaces einen
12128: @c Funktionspointer übergeben bekommt, den man dann bei Gelegenheit
12129: @c aufrufen muss. Also kann man den deklarieren, und das damit deklarierte
12130: @c Wort verhält sich dann wie ein EXECUTE für alle C-Funktionen mit
12131: @c demselben Prototyp.
12132:
12133:
1.178 anton 12134: @node C interface internals, Low-Level C Interface Words, Callbacks, C Interface
1.177 anton 12135: @subsection How the C interface works
12136:
12137: The documented C interface works by generating a C code out of the
12138: declarations.
12139:
12140: In particular, for every Forth word declared with @code{c-function},
12141: it generates a wrapper function in C that takes the Forth data from
12142: the Forth stacks, and calls the target C function with these data as
12143: arguments. The C compiler then performs an implicit conversion
12144: between the Forth type from the stack, and the C type for the
12145: parameter, which is given by the C function prototype. After the C
12146: function returns, the return value is likewise implicitly converted to
12147: a Forth type and written back on the stack.
12148:
12149: The @code{\c} lines are literally included in the C code (but without
12150: the @code{\c}), and provide the necessary declarations so that the C
12151: compiler knows the C types and has enough information to perform the
12152: conversion.
12153:
12154: These wrapper functions are eventually compiled and dynamically linked
12155: into Gforth, and then they can be called.
12156:
12157:
1.178 anton 12158: @node Low-Level C Interface Words, , C interface internals, C Interface
1.155 anton 12159: @subsection Low-Level C Interface Words
1.44 crook 12160:
1.155 anton 12161: doc-open-lib
12162: doc-lib-sym
1.177 anton 12163: doc-call-c
1.26 crook 12164:
1.78 anton 12165: @c -------------------------------------------------------------
1.150 anton 12166: @node Assembler and Code Words, Threading Words, C Interface, Words
1.78 anton 12167: @section Assembler and Code Words
12168: @cindex assembler
12169: @cindex code words
1.44 crook 12170:
1.78 anton 12171: @menu
12172: * Code and ;code::
12173: * Common Assembler:: Assembler Syntax
12174: * Common Disassembler::
12175: * 386 Assembler:: Deviations and special cases
12176: * Alpha Assembler:: Deviations and special cases
12177: * MIPS assembler:: Deviations and special cases
1.161 anton 12178: * PowerPC assembler:: Deviations and special cases
1.78 anton 12179: * Other assemblers:: How to write them
12180: @end menu
1.21 crook 12181:
1.78 anton 12182: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
12183: @subsection @code{Code} and @code{;code}
1.26 crook 12184:
1.78 anton 12185: Gforth provides some words for defining primitives (words written in
12186: machine code), and for defining the machine-code equivalent of
12187: @code{DOES>}-based defining words. However, the machine-independent
12188: nature of Gforth poses a few problems: First of all, Gforth runs on
12189: several architectures, so it can provide no standard assembler. What's
12190: worse is that the register allocation not only depends on the processor,
12191: but also on the @code{gcc} version and options used.
1.44 crook 12192:
1.78 anton 12193: The words that Gforth offers encapsulate some system dependences (e.g.,
12194: the header structure), so a system-independent assembler may be used in
12195: Gforth. If you do not have an assembler, you can compile machine code
12196: directly with @code{,} and @code{c,}@footnote{This isn't portable,
12197: because these words emit stuff in @i{data} space; it works because
12198: Gforth has unified code/data spaces. Assembler isn't likely to be
12199: portable anyway.}.
1.21 crook 12200:
1.44 crook 12201:
1.78 anton 12202: doc-assembler
12203: doc-init-asm
12204: doc-code
12205: doc-end-code
12206: doc-;code
12207: doc-flush-icache
1.44 crook 12208:
1.21 crook 12209:
1.78 anton 12210: If @code{flush-icache} does not work correctly, @code{code} words
12211: etc. will not work (reliably), either.
1.44 crook 12212:
1.78 anton 12213: The typical usage of these @code{code} words can be shown most easily by
12214: analogy to the equivalent high-level defining words:
1.44 crook 12215:
1.78 anton 12216: @example
12217: : foo code foo
12218: <high-level Forth words> <assembler>
12219: ; end-code
12220:
12221: : bar : bar
12222: <high-level Forth words> <high-level Forth words>
12223: CREATE CREATE
12224: <high-level Forth words> <high-level Forth words>
12225: DOES> ;code
12226: <high-level Forth words> <assembler>
12227: ; end-code
12228: @end example
1.21 crook 12229:
1.78 anton 12230: @c anton: the following stuff is also in "Common Assembler", in less detail.
1.44 crook 12231:
1.78 anton 12232: @cindex registers of the inner interpreter
12233: In the assembly code you will want to refer to the inner interpreter's
12234: registers (e.g., the data stack pointer) and you may want to use other
12235: registers for temporary storage. Unfortunately, the register allocation
12236: is installation-dependent.
1.44 crook 12237:
1.78 anton 12238: In particular, @code{ip} (Forth instruction pointer) and @code{rp}
1.100 anton 12239: (return stack pointer) may be in different places in @code{gforth} and
12240: @code{gforth-fast}, or different installations. This means that you
12241: cannot write a @code{NEXT} routine that works reliably on both versions
12242: or different installations; so for doing @code{NEXT}, I recommend
12243: jumping to @code{' noop >code-address}, which contains nothing but a
12244: @code{NEXT}.
1.21 crook 12245:
1.78 anton 12246: For general accesses to the inner interpreter's registers, the easiest
12247: solution is to use explicit register declarations (@pxref{Explicit Reg
12248: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) for
12249: all of the inner interpreter's registers: You have to compile Gforth
12250: with @code{-DFORCE_REG} (configure option @code{--enable-force-reg}) and
12251: the appropriate declarations must be present in the @code{machine.h}
12252: file (see @code{mips.h} for an example; you can find a full list of all
12253: declarable register symbols with @code{grep register engine.c}). If you
12254: give explicit registers to all variables that are declared at the
12255: beginning of @code{engine()}, you should be able to use the other
12256: caller-saved registers for temporary storage. Alternatively, you can use
12257: the @code{gcc} option @code{-ffixed-REG} (@pxref{Code Gen Options, ,
12258: Options for Code Generation Conventions, gcc.info, GNU C Manual}) to
12259: reserve a register (however, this restriction on register allocation may
12260: slow Gforth significantly).
1.44 crook 12261:
1.78 anton 12262: If this solution is not viable (e.g., because @code{gcc} does not allow
12263: you to explicitly declare all the registers you need), you have to find
12264: out by looking at the code where the inner interpreter's registers
12265: reside and which registers can be used for temporary storage. You can
12266: get an assembly listing of the engine's code with @code{make engine.s}.
1.44 crook 12267:
1.78 anton 12268: In any case, it is good practice to abstract your assembly code from the
12269: actual register allocation. E.g., if the data stack pointer resides in
12270: register @code{$17}, create an alias for this register called @code{sp},
12271: and use that in your assembly code.
1.21 crook 12272:
1.78 anton 12273: @cindex code words, portable
12274: Another option for implementing normal and defining words efficiently
12275: is to add the desired functionality to the source of Gforth. For normal
12276: words you just have to edit @file{primitives} (@pxref{Automatic
12277: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
12278: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
12279: @file{prims2x.fs}, and possibly @file{cross.fs}.
1.44 crook 12280:
1.78 anton 12281: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
12282: @subsection Common Assembler
1.44 crook 12283:
1.78 anton 12284: The assemblers in Gforth generally use a postfix syntax, i.e., the
12285: instruction name follows the operands.
1.21 crook 12286:
1.78 anton 12287: The operands are passed in the usual order (the same that is used in the
12288: manual of the architecture). Since they all are Forth words, they have
12289: to be separated by spaces; you can also use Forth words to compute the
12290: operands.
1.44 crook 12291:
1.78 anton 12292: The instruction names usually end with a @code{,}. This makes it easier
12293: to visually separate instructions if you put several of them on one
12294: line; it also avoids shadowing other Forth words (e.g., @code{and}).
1.21 crook 12295:
1.78 anton 12296: Registers are usually specified by number; e.g., (decimal) @code{11}
12297: specifies registers R11 and F11 on the Alpha architecture (which one,
12298: depends on the instruction). The usual names are also available, e.g.,
12299: @code{s2} for R11 on Alpha.
1.21 crook 12300:
1.78 anton 12301: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
12302: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
12303: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
12304: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
12305: conditions are specified in a way specific to each assembler.
1.1 anton 12306:
1.78 anton 12307: Note that the register assignments of the Gforth engine can change
12308: between Gforth versions, or even between different compilations of the
12309: same Gforth version (e.g., if you use a different GCC version). So if
12310: you want to refer to Gforth's registers (e.g., the stack pointer or
12311: TOS), I recommend defining your own words for refering to these
12312: registers, and using them later on; then you can easily adapt to a
12313: changed register assignment. The stability of the register assignment
12314: is usually better if you build Gforth with @code{--enable-force-reg}.
1.1 anton 12315:
1.100 anton 12316: The most common use of these registers is to dispatch to the next word
12317: (the @code{next} routine). A portable way to do this is to jump to
12318: @code{' noop >code-address} (of course, this is less efficient than
12319: integrating the @code{next} code and scheduling it well).
1.1 anton 12320:
1.96 anton 12321: Another difference between Gforth version is that the top of stack is
12322: kept in memory in @code{gforth} and, on most platforms, in a register in
12323: @code{gforth-fast}.
12324:
1.78 anton 12325: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
12326: @subsection Common Disassembler
1.127 anton 12327: @cindex disassembler, general
12328: @cindex gdb disassembler
1.1 anton 12329:
1.78 anton 12330: You can disassemble a @code{code} word with @code{see}
12331: (@pxref{Debugging}). You can disassemble a section of memory with
1.1 anton 12332:
1.127 anton 12333: doc-discode
1.44 crook 12334:
1.127 anton 12335: There are two kinds of disassembler for Gforth: The Forth disassembler
12336: (available on some CPUs) and the gdb disassembler (available on
12337: platforms with @command{gdb} and @command{mktemp}). If both are
12338: available, the Forth disassembler is used by default. If you prefer
12339: the gdb disassembler, say
12340:
12341: @example
12342: ' disasm-gdb is discode
12343: @end example
12344:
12345: If neither is available, @code{discode} performs @code{dump}.
12346:
12347: The Forth disassembler generally produces output that can be fed into the
1.78 anton 12348: assembler (i.e., same syntax, etc.). It also includes additional
12349: information in comments. In particular, the address of the instruction
12350: is given in a comment before the instruction.
1.1 anton 12351:
1.127 anton 12352: The gdb disassembler produces output in the same format as the gdb
12353: @code{disassemble} command (@pxref{Machine Code,,Source and machine
12354: code,gdb,Debugging with GDB}), in the default flavour (AT&T syntax for
12355: the 386 and AMD64 architectures).
12356:
1.78 anton 12357: @code{See} may display more or less than the actual code of the word,
12358: because the recognition of the end of the code is unreliable. You can
1.127 anton 12359: use @code{discode} if it did not display enough. It may display more, if
1.78 anton 12360: the code word is not immediately followed by a named word. If you have
1.116 anton 12361: something else there, you can follow the word with @code{align latest ,}
1.78 anton 12362: to ensure that the end is recognized.
1.21 crook 12363:
1.78 anton 12364: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
12365: @subsection 386 Assembler
1.44 crook 12366:
1.78 anton 12367: The 386 assembler included in Gforth was written by Bernd Paysan, it's
12368: available under GPL, and originally part of bigFORTH.
1.21 crook 12369:
1.78 anton 12370: The 386 disassembler included in Gforth was written by Andrew McKewan
12371: and is in the public domain.
1.21 crook 12372:
1.91 anton 12373: The disassembler displays code in an Intel-like prefix syntax.
1.21 crook 12374:
1.78 anton 12375: The assembler uses a postfix syntax with reversed parameters.
1.1 anton 12376:
1.78 anton 12377: The assembler includes all instruction of the Athlon, i.e. 486 core
12378: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
12379: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
12380: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
1.1 anton 12381:
1.78 anton 12382: There are several prefixes to switch between different operation sizes,
12383: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
12384: double-word accesses. Addressing modes can be switched with @code{.wa}
12385: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
12386: need a prefix for byte register names (@code{AL} et al).
1.1 anton 12387:
1.78 anton 12388: For floating point operations, the prefixes are @code{.fs} (IEEE
12389: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
12390: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
1.21 crook 12391:
1.78 anton 12392: The MMX opcodes don't have size prefixes, they are spelled out like in
12393: the Intel assembler. Instead of move from and to memory, there are
12394: PLDQ/PLDD and PSTQ/PSTD.
1.21 crook 12395:
1.78 anton 12396: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
12397: ax. Immediate values are indicated by postfixing them with @code{#},
1.91 anton 12398: e.g., @code{3 #}. Here are some examples of addressing modes in various
12399: syntaxes:
1.21 crook 12400:
1.26 crook 12401: @example
1.91 anton 12402: Gforth Intel (NASM) AT&T (gas) Name
12403: .w ax ax %ax register (16 bit)
12404: ax eax %eax register (32 bit)
12405: 3 # offset 3 $3 immediate
12406: 1000 #) byte ptr 1000 1000 displacement
12407: bx ) [ebx] (%ebx) base
12408: 100 di d) 100[edi] 100(%edi) base+displacement
12409: 20 ax *4 i#) 20[eax*4] 20(,%eax,4) (index*scale)+displacement
12410: di ax *4 i) [edi][eax*4] (%edi,%eax,4) base+(index*scale)
12411: 4 bx cx di) 4[ebx][ecx] 4(%ebx,%ecx) base+index+displacement
12412: 12 sp ax *2 di) 12[esp][eax*2] 12(%esp,%eax,2) base+(index*scale)+displacement
12413: @end example
12414:
12415: You can use @code{L)} and @code{LI)} instead of @code{D)} and
12416: @code{DI)} to enforce 32-bit displacement fields (useful for
12417: later patching).
1.21 crook 12418:
1.78 anton 12419: Some example of instructions are:
1.1 anton 12420:
12421: @example
1.78 anton 12422: ax bx mov \ move ebx,eax
12423: 3 # ax mov \ mov eax,3
1.137 pazsan 12424: 100 di d) ax mov \ mov eax,100[edi]
1.78 anton 12425: 4 bx cx di) ax mov \ mov eax,4[ebx][ecx]
12426: .w ax bx mov \ mov bx,ax
1.1 anton 12427: @end example
12428:
1.78 anton 12429: The following forms are supported for binary instructions:
1.1 anton 12430:
12431: @example
1.78 anton 12432: <reg> <reg> <inst>
12433: <n> # <reg> <inst>
12434: <mem> <reg> <inst>
12435: <reg> <mem> <inst>
1.136 pazsan 12436: <n> # <mem> <inst>
1.1 anton 12437: @end example
12438:
1.136 pazsan 12439: The shift/rotate syntax is:
1.1 anton 12440:
1.26 crook 12441: @example
1.78 anton 12442: <reg/mem> 1 # shl \ shortens to shift without immediate
12443: <reg/mem> 4 # shl
12444: <reg/mem> cl shl
1.26 crook 12445: @end example
1.1 anton 12446:
1.78 anton 12447: Precede string instructions (@code{movs} etc.) with @code{.b} to get
12448: the byte version.
1.1 anton 12449:
1.78 anton 12450: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
12451: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
12452: pc < >= <= >}. (Note that most of these words shadow some Forth words
12453: when @code{assembler} is in front of @code{forth} in the search path,
12454: e.g., in @code{code} words). Currently the control structure words use
12455: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
12456: to shuffle them (you can also use @code{swap} etc.).
1.21 crook 12457:
1.78 anton 12458: Here is an example of a @code{code} word (assumes that the stack pointer
12459: is in esi and the TOS is in ebx):
1.21 crook 12460:
1.26 crook 12461: @example
1.78 anton 12462: code my+ ( n1 n2 -- n )
12463: 4 si D) bx add
12464: 4 # si add
12465: Next
12466: end-code
1.26 crook 12467: @end example
1.21 crook 12468:
1.161 anton 12469:
1.78 anton 12470: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
12471: @subsection Alpha Assembler
1.21 crook 12472:
1.78 anton 12473: The Alpha assembler and disassembler were originally written by Bernd
12474: Thallner.
1.26 crook 12475:
1.78 anton 12476: The register names @code{a0}--@code{a5} are not available to avoid
12477: shadowing hex numbers.
1.2 jwilke 12478:
1.78 anton 12479: Immediate forms of arithmetic instructions are distinguished by a
12480: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
12481: does not count as arithmetic instruction).
1.2 jwilke 12482:
1.78 anton 12483: You have to specify all operands to an instruction, even those that
12484: other assemblers consider optional, e.g., the destination register for
12485: @code{br,}, or the destination register and hint for @code{jmp,}.
1.2 jwilke 12486:
1.78 anton 12487: You can specify conditions for @code{if,} by removing the first @code{b}
12488: and the trailing @code{,} from a branch with a corresponding name; e.g.,
1.2 jwilke 12489:
1.26 crook 12490: @example
1.78 anton 12491: 11 fgt if, \ if F11>0e
12492: ...
12493: endif,
1.26 crook 12494: @end example
1.2 jwilke 12495:
1.78 anton 12496: @code{fbgt,} gives @code{fgt}.
12497:
1.161 anton 12498: @node MIPS assembler, PowerPC assembler, Alpha Assembler, Assembler and Code Words
1.78 anton 12499: @subsection MIPS assembler
1.2 jwilke 12500:
1.78 anton 12501: The MIPS assembler was originally written by Christian Pirker.
1.2 jwilke 12502:
1.78 anton 12503: Currently the assembler and disassembler only cover the MIPS-I
12504: architecture (R3000), and don't support FP instructions.
1.2 jwilke 12505:
1.78 anton 12506: The register names @code{$a0}--@code{$a3} are not available to avoid
12507: shadowing hex numbers.
1.2 jwilke 12508:
1.78 anton 12509: Because there is no way to distinguish registers from immediate values,
12510: you have to explicitly use the immediate forms of instructions, i.e.,
12511: @code{addiu,}, not just @code{addu,} (@command{as} does this
12512: implicitly).
1.2 jwilke 12513:
1.78 anton 12514: If the architecture manual specifies several formats for the instruction
12515: (e.g., for @code{jalr,}), you usually have to use the one with more
12516: arguments (i.e., two for @code{jalr,}). When in doubt, see
12517: @code{arch/mips/testasm.fs} for an example of correct use.
1.2 jwilke 12518:
1.78 anton 12519: Branches and jumps in the MIPS architecture have a delay slot. You have
12520: to fill it yourself (the simplest way is to use @code{nop,}), the
12521: assembler does not do it for you (unlike @command{as}). Even
12522: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
12523: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
12524: and @code{then,} just specify branch targets, they are not affected.
1.2 jwilke 12525:
1.78 anton 12526: Note that you must not put branches, jumps, or @code{li,} into the delay
12527: slot: @code{li,} may expand to several instructions, and control flow
12528: instructions may not be put into the branch delay slot in any case.
1.2 jwilke 12529:
1.78 anton 12530: For branches the argument specifying the target is a relative address;
12531: You have to add the address of the delay slot to get the absolute
12532: address.
1.1 anton 12533:
1.78 anton 12534: The MIPS architecture also has load delay slots and restrictions on
12535: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
12536: yourself to satisfy these restrictions, the assembler does not do it for
12537: you.
1.1 anton 12538:
1.78 anton 12539: You can specify the conditions for @code{if,} etc. by taking a
12540: conditional branch and leaving away the @code{b} at the start and the
12541: @code{,} at the end. E.g.,
1.1 anton 12542:
1.26 crook 12543: @example
1.78 anton 12544: 4 5 eq if,
12545: ... \ do something if $4 equals $5
12546: then,
1.26 crook 12547: @end example
1.1 anton 12548:
1.161 anton 12549:
12550: @node PowerPC assembler, Other assemblers, MIPS assembler, Assembler and Code Words
12551: @subsection PowerPC assembler
12552:
1.162 anton 12553: The PowerPC assembler and disassembler were contributed by Michal
1.161 anton 12554: Revucky.
12555:
1.162 anton 12556: This assembler does not follow the convention of ending mnemonic names
12557: with a ``,'', so some mnemonic names shadow regular Forth words (in
12558: particular: @code{and or xor fabs}); so if you want to use the Forth
12559: words, you have to make them visible first, e.g., with @code{also
12560: forth}.
12561:
1.161 anton 12562: Registers are referred to by their number, e.g., @code{9} means the
12563: integer register 9 or the FP register 9 (depending on the
12564: instruction).
12565:
12566: Because there is no way to distinguish registers from immediate values,
12567: you have to explicitly use the immediate forms of instructions, i.e.,
1.162 anton 12568: @code{addi,}, not just @code{add,}.
1.161 anton 12569:
1.162 anton 12570: The assembler and disassembler usually support the most general form
1.161 anton 12571: of an instruction, but usually not the shorter forms (especially for
12572: branches).
12573:
12574:
12575:
12576: @node Other assemblers, , PowerPC assembler, Assembler and Code Words
1.78 anton 12577: @subsection Other assemblers
12578:
12579: If you want to contribute another assembler/disassembler, please contact
1.103 anton 12580: us (@email{anton@@mips.complang.tuwien.ac.at}) to check if we have such
12581: an assembler already. If you are writing them from scratch, please use
12582: a similar syntax style as the one we use (i.e., postfix, commas at the
12583: end of the instruction names, @pxref{Common Assembler}); make the output
12584: of the disassembler be valid input for the assembler, and keep the style
1.78 anton 12585: similar to the style we used.
12586:
12587: Hints on implementation: The most important part is to have a good test
12588: suite that contains all instructions. Once you have that, the rest is
12589: easy. For actual coding you can take a look at
12590: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
12591: the assembler and disassembler, avoiding redundancy and some potential
12592: bugs. You can also look at that file (and @pxref{Advanced does> usage
12593: example}) to get ideas how to factor a disassembler.
12594:
12595: Start with the disassembler, because it's easier to reuse data from the
12596: disassembler for the assembler than the other way round.
1.1 anton 12597:
1.78 anton 12598: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
12599: how simple it can be.
1.1 anton 12600:
1.161 anton 12601:
12602:
12603:
1.78 anton 12604: @c -------------------------------------------------------------
12605: @node Threading Words, Passing Commands to the OS, Assembler and Code Words, Words
12606: @section Threading Words
12607: @cindex threading words
1.1 anton 12608:
1.78 anton 12609: @cindex code address
12610: These words provide access to code addresses and other threading stuff
12611: in Gforth (and, possibly, other interpretive Forths). It more or less
12612: abstracts away the differences between direct and indirect threading
12613: (and, for direct threading, the machine dependences). However, at
12614: present this wordset is still incomplete. It is also pretty low-level;
12615: some day it will hopefully be made unnecessary by an internals wordset
12616: that abstracts implementation details away completely.
1.1 anton 12617:
1.78 anton 12618: The terminology used here stems from indirect threaded Forth systems; in
12619: such a system, the XT of a word is represented by the CFA (code field
12620: address) of a word; the CFA points to a cell that contains the code
12621: address. The code address is the address of some machine code that
12622: performs the run-time action of invoking the word (e.g., the
12623: @code{dovar:} routine pushes the address of the body of the word (a
12624: variable) on the stack
12625: ).
1.1 anton 12626:
1.78 anton 12627: @cindex code address
12628: @cindex code field address
12629: In an indirect threaded Forth, you can get the code address of @i{name}
12630: with @code{' @i{name} @@}; in Gforth you can get it with @code{' @i{name}
12631: >code-address}, independent of the threading method.
1.1 anton 12632:
1.78 anton 12633: doc-threading-method
12634: doc->code-address
12635: doc-code-address!
1.1 anton 12636:
1.78 anton 12637: @cindex @code{does>}-handler
12638: @cindex @code{does>}-code
12639: For a word defined with @code{DOES>}, the code address usually points to
12640: a jump instruction (the @dfn{does-handler}) that jumps to the dodoes
12641: routine (in Gforth on some platforms, it can also point to the dodoes
12642: routine itself). What you are typically interested in, though, is
12643: whether a word is a @code{DOES>}-defined word, and what Forth code it
12644: executes; @code{>does-code} tells you that.
1.1 anton 12645:
1.78 anton 12646: doc->does-code
1.1 anton 12647:
1.78 anton 12648: To create a @code{DOES>}-defined word with the following basic words,
12649: you have to set up a @code{DOES>}-handler with @code{does-handler!};
12650: @code{/does-handler} aus behind you have to place your executable Forth
12651: code. Finally you have to create a word and modify its behaviour with
12652: @code{does-handler!}.
1.1 anton 12653:
1.78 anton 12654: doc-does-code!
12655: doc-does-handler!
12656: doc-/does-handler
1.1 anton 12657:
1.78 anton 12658: The code addresses produced by various defining words are produced by
12659: the following words:
1.1 anton 12660:
1.78 anton 12661: doc-docol:
12662: doc-docon:
12663: doc-dovar:
12664: doc-douser:
12665: doc-dodefer:
12666: doc-dofield:
1.1 anton 12667:
1.99 anton 12668: @cindex definer
12669: The following two words generalize @code{>code-address},
12670: @code{>does-code}, @code{code-address!}, and @code{does-code!}:
12671:
12672: doc->definer
12673: doc-definer!
12674:
1.26 crook 12675: @c -------------------------------------------------------------
1.78 anton 12676: @node Passing Commands to the OS, Keeping track of Time, Threading Words, Words
1.21 crook 12677: @section Passing Commands to the Operating System
12678: @cindex operating system - passing commands
12679: @cindex shell commands
12680:
12681: Gforth allows you to pass an arbitrary string to the host operating
12682: system shell (if such a thing exists) for execution.
12683:
12684: doc-sh
12685: doc-system
12686: doc-$?
1.23 crook 12687: doc-getenv
1.44 crook 12688:
1.26 crook 12689: @c -------------------------------------------------------------
1.47 crook 12690: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
12691: @section Keeping track of Time
12692: @cindex time-related words
12693:
12694: doc-ms
12695: doc-time&date
1.79 anton 12696: doc-utime
12697: doc-cputime
1.47 crook 12698:
12699:
12700: @c -------------------------------------------------------------
12701: @node Miscellaneous Words, , Keeping track of Time, Words
1.21 crook 12702: @section Miscellaneous Words
12703: @cindex miscellaneous words
12704:
1.29 crook 12705: @comment TODO find homes for these
12706:
1.26 crook 12707: These section lists the ANS Forth words that are not documented
1.21 crook 12708: elsewhere in this manual. Ultimately, they all need proper homes.
12709:
1.68 anton 12710: doc-quit
1.44 crook 12711:
1.26 crook 12712: The following ANS Forth words are not currently supported by Gforth
1.27 crook 12713: (@pxref{ANS conformance}):
1.21 crook 12714:
12715: @code{EDITOR}
12716: @code{EMIT?}
12717: @code{FORGET}
12718:
1.24 anton 12719: @c ******************************************************************
12720: @node Error messages, Tools, Words, Top
12721: @chapter Error messages
12722: @cindex error messages
12723: @cindex backtrace
12724:
12725: A typical Gforth error message looks like this:
12726:
12727: @example
1.86 anton 12728: in file included from \evaluated string/:-1
1.24 anton 12729: in file included from ./yyy.fs:1
12730: ./xxx.fs:4: Invalid memory address
1.134 anton 12731: >>>bar<<<
1.79 anton 12732: Backtrace:
1.25 anton 12733: $400E664C @@
12734: $400E6664 foo
1.24 anton 12735: @end example
12736:
12737: The message identifying the error is @code{Invalid memory address}. The
12738: error happened when text-interpreting line 4 of the file
12739: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
12740: word on the line where the error happened, is pointed out (with
1.134 anton 12741: @code{>>>} and @code{<<<}).
1.24 anton 12742:
12743: The file containing the error was included in line 1 of @file{./yyy.fs},
12744: and @file{yyy.fs} was included from a non-file (in this case, by giving
12745: @file{yyy.fs} as command-line parameter to Gforth).
12746:
12747: At the end of the error message you find a return stack dump that can be
12748: interpreted as a backtrace (possibly empty). On top you find the top of
12749: the return stack when the @code{throw} happened, and at the bottom you
12750: find the return stack entry just above the return stack of the topmost
12751: text interpreter.
12752:
12753: To the right of most return stack entries you see a guess for the word
12754: that pushed that return stack entry as its return address. This gives a
12755: backtrace. In our case we see that @code{bar} called @code{foo}, and
12756: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
12757: address} exception).
12758:
12759: Note that the backtrace is not perfect: We don't know which return stack
12760: entries are return addresses (so we may get false positives); and in
12761: some cases (e.g., for @code{abort"}) we cannot determine from the return
12762: address the word that pushed the return address, so for some return
12763: addresses you see no names in the return stack dump.
1.25 anton 12764:
12765: @cindex @code{catch} and backtraces
12766: The return stack dump represents the return stack at the time when a
12767: specific @code{throw} was executed. In programs that make use of
12768: @code{catch}, it is not necessarily clear which @code{throw} should be
12769: used for the return stack dump (e.g., consider one @code{throw} that
12770: indicates an error, which is caught, and during recovery another error
1.160 anton 12771: happens; which @code{throw} should be used for the stack dump?).
12772: Gforth presents the return stack dump for the first @code{throw} after
12773: the last executed (not returned-to) @code{catch} or @code{nothrow};
12774: this works well in the usual case. To get the right backtrace, you
12775: usually want to insert @code{nothrow} or @code{['] false catch drop}
12776: after a @code{catch} if the error is not rethrown.
1.25 anton 12777:
12778: @cindex @code{gforth-fast} and backtraces
12779: @cindex @code{gforth-fast}, difference from @code{gforth}
12780: @cindex backtraces with @code{gforth-fast}
12781: @cindex return stack dump with @code{gforth-fast}
1.79 anton 12782: @code{Gforth} is able to do a return stack dump for throws generated
1.25 anton 12783: from primitives (e.g., invalid memory address, stack empty etc.);
12784: @code{gforth-fast} is only able to do a return stack dump from a
1.96 anton 12785: directly called @code{throw} (including @code{abort} etc.). Given an
1.30 anton 12786: exception caused by a primitive in @code{gforth-fast}, you will
12787: typically see no return stack dump at all; however, if the exception is
12788: caught by @code{catch} (e.g., for restoring some state), and then
12789: @code{throw}n again, the return stack dump will be for the first such
12790: @code{throw}.
1.2 jwilke 12791:
1.5 anton 12792: @c ******************************************************************
1.24 anton 12793: @node Tools, ANS conformance, Error messages, Top
1.1 anton 12794: @chapter Tools
12795:
12796: @menu
12797: * ANS Report:: Report the words used, sorted by wordset.
1.127 anton 12798: * Stack depth changes:: Where does this stack item come from?
1.1 anton 12799: @end menu
12800:
12801: See also @ref{Emacs and Gforth}.
12802:
1.126 pazsan 12803: @node ANS Report, Stack depth changes, Tools, Tools
1.1 anton 12804: @section @file{ans-report.fs}: Report the words used, sorted by wordset
12805: @cindex @file{ans-report.fs}
12806: @cindex report the words used in your program
12807: @cindex words used in your program
12808:
12809: If you want to label a Forth program as ANS Forth Program, you must
12810: document which wordsets the program uses; for extension wordsets, it is
12811: helpful to list the words the program requires from these wordsets
12812: (because Forth systems are allowed to provide only some words of them).
12813:
12814: The @file{ans-report.fs} tool makes it easy for you to determine which
12815: words from which wordset and which non-ANS words your application
12816: uses. You simply have to include @file{ans-report.fs} before loading the
12817: program you want to check. After loading your program, you can get the
12818: report with @code{print-ans-report}. A typical use is to run this as
12819: batch job like this:
12820: @example
12821: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
12822: @end example
12823:
12824: The output looks like this (for @file{compat/control.fs}):
12825: @example
12826: The program uses the following words
12827: from CORE :
12828: : POSTPONE THEN ; immediate ?dup IF 0=
12829: from BLOCK-EXT :
12830: \
12831: from FILE :
12832: (
12833: @end example
12834:
12835: @subsection Caveats
12836:
12837: Note that @file{ans-report.fs} just checks which words are used, not whether
12838: they are used in an ANS Forth conforming way!
12839:
12840: Some words are defined in several wordsets in the
12841: standard. @file{ans-report.fs} reports them for only one of the
12842: wordsets, and not necessarily the one you expect. It depends on usage
12843: which wordset is the right one to specify. E.g., if you only use the
12844: compilation semantics of @code{S"}, it is a Core word; if you also use
12845: its interpretation semantics, it is a File word.
1.124 anton 12846:
12847:
1.127 anton 12848: @node Stack depth changes, , ANS Report, Tools
1.124 anton 12849: @section Stack depth changes during interpretation
12850: @cindex @file{depth-changes.fs}
12851: @cindex depth changes during interpretation
12852: @cindex stack depth changes during interpretation
12853: @cindex items on the stack after interpretation
12854:
12855: Sometimes you notice that, after loading a file, there are items left
12856: on the stack. The tool @file{depth-changes.fs} helps you find out
12857: quickly where in the file these stack items are coming from.
12858:
12859: The simplest way of using @file{depth-changes.fs} is to include it
12860: before the file(s) you want to check, e.g.:
12861:
12862: @example
12863: gforth depth-changes.fs my-file.fs
12864: @end example
12865:
12866: This will compare the stack depths of the data and FP stack at every
12867: empty line (in interpretation state) against these depths at the last
12868: empty line (in interpretation state). If the depths are not equal,
12869: the position in the file and the stack contents are printed with
12870: @code{~~} (@pxref{Debugging}). This indicates that a stack depth
12871: change has occured in the paragraph of non-empty lines before the
12872: indicated line. It is a good idea to leave an empty line at the end
12873: of the file, so the last paragraph is checked, too.
12874:
12875: Checking only at empty lines usually works well, but sometimes you
12876: have big blocks of non-empty lines (e.g., when building a big table),
12877: and you want to know where in this block the stack depth changed. You
12878: can check all interpreted lines with
12879:
12880: @example
12881: gforth depth-changes.fs -e "' all-lines is depth-changes-filter" my-file.fs
12882: @end example
12883:
12884: This checks the stack depth at every end-of-line. So the depth change
12885: occured in the line reported by the @code{~~} (not in the line
12886: before).
12887:
12888: Note that, while this offers better accuracy in indicating where the
12889: stack depth changes, it will often report many intentional stack depth
12890: changes (e.g., when an interpreted computation stretches across
12891: several lines). You can suppress the checking of some lines by
12892: putting backslashes at the end of these lines (not followed by white
12893: space), and using
12894:
12895: @example
12896: gforth depth-changes.fs -e "' most-lines is depth-changes-filter" my-file.fs
12897: @end example
1.1 anton 12898:
12899: @c ******************************************************************
1.65 anton 12900: @node ANS conformance, Standard vs Extensions, Tools, Top
1.1 anton 12901: @chapter ANS conformance
12902: @cindex ANS conformance of Gforth
12903:
12904: To the best of our knowledge, Gforth is an
12905:
12906: ANS Forth System
12907: @itemize @bullet
12908: @item providing the Core Extensions word set
12909: @item providing the Block word set
12910: @item providing the Block Extensions word set
12911: @item providing the Double-Number word set
12912: @item providing the Double-Number Extensions word set
12913: @item providing the Exception word set
12914: @item providing the Exception Extensions word set
12915: @item providing the Facility word set
1.40 anton 12916: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
1.1 anton 12917: @item providing the File Access word set
12918: @item providing the File Access Extensions word set
12919: @item providing the Floating-Point word set
12920: @item providing the Floating-Point Extensions word set
12921: @item providing the Locals word set
12922: @item providing the Locals Extensions word set
12923: @item providing the Memory-Allocation word set
12924: @item providing the Memory-Allocation Extensions word set (that one's easy)
12925: @item providing the Programming-Tools word set
12926: @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
12927: @item providing the Search-Order word set
12928: @item providing the Search-Order Extensions word set
12929: @item providing the String word set
12930: @item providing the String Extensions word set (another easy one)
12931: @end itemize
12932:
1.118 anton 12933: Gforth has the following environmental restrictions:
12934:
12935: @cindex environmental restrictions
12936: @itemize @bullet
12937: @item
12938: While processing the OS command line, if an exception is not caught,
12939: Gforth exits with a non-zero exit code instyead of performing QUIT.
12940:
12941: @item
12942: When an @code{throw} is performed after a @code{query}, Gforth does not
12943: allways restore the input source specification in effect at the
12944: corresponding catch.
12945:
12946: @end itemize
12947:
12948:
1.1 anton 12949: @cindex system documentation
12950: In addition, ANS Forth systems are required to document certain
12951: implementation choices. This chapter tries to meet these
12952: requirements. In many cases it gives a way to ask the system for the
12953: information instead of providing the information directly, in
12954: particular, if the information depends on the processor, the operating
12955: system or the installation options chosen, or if they are likely to
12956: change during the maintenance of Gforth.
12957:
12958: @comment The framework for the rest has been taken from pfe.
12959:
12960: @menu
12961: * The Core Words::
12962: * The optional Block word set::
12963: * The optional Double Number word set::
12964: * The optional Exception word set::
12965: * The optional Facility word set::
12966: * The optional File-Access word set::
12967: * The optional Floating-Point word set::
12968: * The optional Locals word set::
12969: * The optional Memory-Allocation word set::
12970: * The optional Programming-Tools word set::
12971: * The optional Search-Order word set::
12972: @end menu
12973:
12974:
12975: @c =====================================================================
12976: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
12977: @comment node-name, next, previous, up
12978: @section The Core Words
12979: @c =====================================================================
12980: @cindex core words, system documentation
12981: @cindex system documentation, core words
12982:
12983: @menu
12984: * core-idef:: Implementation Defined Options
12985: * core-ambcond:: Ambiguous Conditions
12986: * core-other:: Other System Documentation
12987: @end menu
12988:
12989: @c ---------------------------------------------------------------------
12990: @node core-idef, core-ambcond, The Core Words, The Core Words
12991: @subsection Implementation Defined Options
12992: @c ---------------------------------------------------------------------
12993: @cindex core words, implementation-defined options
12994: @cindex implementation-defined options, core words
12995:
12996:
12997: @table @i
12998: @item (Cell) aligned addresses:
12999: @cindex cell-aligned addresses
13000: @cindex aligned addresses
13001: processor-dependent. Gforth's alignment words perform natural alignment
13002: (e.g., an address aligned for a datum of size 8 is divisible by
13003: 8). Unaligned accesses usually result in a @code{-23 THROW}.
13004:
13005: @item @code{EMIT} and non-graphic characters:
13006: @cindex @code{EMIT} and non-graphic characters
13007: @cindex non-graphic characters and @code{EMIT}
13008: The character is output using the C library function (actually, macro)
13009: @code{putc}.
13010:
13011: @item character editing of @code{ACCEPT} and @code{EXPECT}:
13012: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
13013: @cindex editing in @code{ACCEPT} and @code{EXPECT}
13014: @cindex @code{ACCEPT}, editing
13015: @cindex @code{EXPECT}, editing
13016: This is modeled on the GNU readline library (@pxref{Readline
13017: Interaction, , Command Line Editing, readline, The GNU Readline
13018: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
13019: producing a full word completion every time you type it (instead of
1.28 crook 13020: producing the common prefix of all completions). @xref{Command-line editing}.
1.1 anton 13021:
13022: @item character set:
13023: @cindex character set
13024: The character set of your computer and display device. Gforth is
13025: 8-bit-clean (but some other component in your system may make trouble).
13026:
13027: @item Character-aligned address requirements:
13028: @cindex character-aligned address requirements
13029: installation-dependent. Currently a character is represented by a C
13030: @code{unsigned char}; in the future we might switch to @code{wchar_t}
13031: (Comments on that requested).
13032:
13033: @item character-set extensions and matching of names:
13034: @cindex character-set extensions and matching of names
1.26 crook 13035: @cindex case-sensitivity for name lookup
13036: @cindex name lookup, case-sensitivity
13037: @cindex locale and case-sensitivity
1.21 crook 13038: Any character except the ASCII NUL character can be used in a
1.1 anton 13039: name. Matching is case-insensitive (except in @code{TABLE}s). The
1.47 crook 13040: matching is performed using the C library function @code{strncasecmp}, whose
1.1 anton 13041: function is probably influenced by the locale. E.g., the @code{C} locale
13042: does not know about accents and umlauts, so they are matched
13043: case-sensitively in that locale. For portability reasons it is best to
13044: write programs such that they work in the @code{C} locale. Then one can
13045: use libraries written by a Polish programmer (who might use words
13046: containing ISO Latin-2 encoded characters) and by a French programmer
13047: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
13048: funny results for some of the words (which ones, depends on the font you
13049: are using)). Also, the locale you prefer may not be available in other
13050: operating systems. Hopefully, Unicode will solve these problems one day.
13051:
13052: @item conditions under which control characters match a space delimiter:
13053: @cindex space delimiters
13054: @cindex control characters as delimiters
1.117 anton 13055: If @code{word} is called with the space character as a delimiter, all
1.1 anton 13056: white-space characters (as identified by the C macro @code{isspace()})
1.117 anton 13057: are delimiters. @code{Parse}, on the other hand, treats space like other
1.138 anton 13058: delimiters. @code{Parse-name}, which is used by the outer
1.1 anton 13059: interpreter (aka text interpreter) by default, treats all white-space
13060: characters as delimiters.
13061:
1.26 crook 13062: @item format of the control-flow stack:
13063: @cindex control-flow stack, format
13064: The data stack is used as control-flow stack. The size of a control-flow
1.1 anton 13065: stack item in cells is given by the constant @code{cs-item-size}. At the
13066: time of this writing, an item consists of a (pointer to a) locals list
13067: (third), an address in the code (second), and a tag for identifying the
13068: item (TOS). The following tags are used: @code{defstart},
13069: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
13070: @code{scopestart}.
13071:
13072: @item conversion of digits > 35
13073: @cindex digits > 35
13074: The characters @code{[\]^_'} are the digits with the decimal value
13075: 36@minus{}41. There is no way to input many of the larger digits.
13076:
13077: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
13078: @cindex @code{EXPECT}, display after end of input
13079: @cindex @code{ACCEPT}, display after end of input
13080: The cursor is moved to the end of the entered string. If the input is
13081: terminated using the @kbd{Return} key, a space is typed.
13082:
13083: @item exception abort sequence of @code{ABORT"}:
13084: @cindex exception abort sequence of @code{ABORT"}
13085: @cindex @code{ABORT"}, exception abort sequence
13086: The error string is stored into the variable @code{"error} and a
13087: @code{-2 throw} is performed.
13088:
13089: @item input line terminator:
13090: @cindex input line terminator
13091: @cindex line terminator on input
1.26 crook 13092: @cindex newline character on input
1.1 anton 13093: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
13094: lines. One of these characters is typically produced when you type the
13095: @kbd{Enter} or @kbd{Return} key.
13096:
13097: @item maximum size of a counted string:
13098: @cindex maximum size of a counted string
13099: @cindex counted string, maximum size
13100: @code{s" /counted-string" environment? drop .}. Currently 255 characters
1.79 anton 13101: on all platforms, but this may change.
1.1 anton 13102:
13103: @item maximum size of a parsed string:
13104: @cindex maximum size of a parsed string
13105: @cindex parsed string, maximum size
13106: Given by the constant @code{/line}. Currently 255 characters.
13107:
13108: @item maximum size of a definition name, in characters:
13109: @cindex maximum size of a definition name, in characters
13110: @cindex name, maximum length
1.113 anton 13111: MAXU/8
1.1 anton 13112:
13113: @item maximum string length for @code{ENVIRONMENT?}, in characters:
13114: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
13115: @cindex @code{ENVIRONMENT?} string length, maximum
1.113 anton 13116: MAXU/8
1.1 anton 13117:
13118: @item method of selecting the user input device:
13119: @cindex user input device, method of selecting
13120: The user input device is the standard input. There is currently no way to
13121: change it from within Gforth. However, the input can typically be
13122: redirected in the command line that starts Gforth.
13123:
13124: @item method of selecting the user output device:
13125: @cindex user output device, method of selecting
13126: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
1.10 anton 13127: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
13128: output when the user output device is a terminal, otherwise the output
13129: is buffered.
1.1 anton 13130:
13131: @item methods of dictionary compilation:
13132: What are we expected to document here?
13133:
13134: @item number of bits in one address unit:
13135: @cindex number of bits in one address unit
13136: @cindex address unit, size in bits
13137: @code{s" address-units-bits" environment? drop .}. 8 in all current
1.79 anton 13138: platforms.
1.1 anton 13139:
13140: @item number representation and arithmetic:
13141: @cindex number representation and arithmetic
1.79 anton 13142: Processor-dependent. Binary two's complement on all current platforms.
1.1 anton 13143:
13144: @item ranges for integer types:
13145: @cindex ranges for integer types
13146: @cindex integer types, ranges
13147: Installation-dependent. Make environmental queries for @code{MAX-N},
13148: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
13149: unsigned (and positive) types is 0. The lower bound for signed types on
13150: two's complement and one's complement machines machines can be computed
13151: by adding 1 to the upper bound.
13152:
13153: @item read-only data space regions:
13154: @cindex read-only data space regions
13155: @cindex data-space, read-only regions
13156: The whole Forth data space is writable.
13157:
13158: @item size of buffer at @code{WORD}:
13159: @cindex size of buffer at @code{WORD}
13160: @cindex @code{WORD} buffer size
13161: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
13162: shared with the pictured numeric output string. If overwriting
13163: @code{PAD} is acceptable, it is as large as the remaining dictionary
13164: space, although only as much can be sensibly used as fits in a counted
13165: string.
13166:
13167: @item size of one cell in address units:
13168: @cindex cell size
13169: @code{1 cells .}.
13170:
13171: @item size of one character in address units:
13172: @cindex char size
1.79 anton 13173: @code{1 chars .}. 1 on all current platforms.
1.1 anton 13174:
13175: @item size of the keyboard terminal buffer:
13176: @cindex size of the keyboard terminal buffer
13177: @cindex terminal buffer, size
13178: Varies. You can determine the size at a specific time using @code{lp@@
13179: tib - .}. It is shared with the locals stack and TIBs of files that
13180: include the current file. You can change the amount of space for TIBs
13181: and locals stack at Gforth startup with the command line option
13182: @code{-l}.
13183:
13184: @item size of the pictured numeric output buffer:
13185: @cindex size of the pictured numeric output buffer
13186: @cindex pictured numeric output buffer, size
13187: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
13188: shared with @code{WORD}.
13189:
13190: @item size of the scratch area returned by @code{PAD}:
13191: @cindex size of the scratch area returned by @code{PAD}
13192: @cindex @code{PAD} size
13193: The remainder of dictionary space. @code{unused pad here - - .}.
13194:
13195: @item system case-sensitivity characteristics:
13196: @cindex case-sensitivity characteristics
1.26 crook 13197: Dictionary searches are case-insensitive (except in
1.1 anton 13198: @code{TABLE}s). However, as explained above under @i{character-set
13199: extensions}, the matching for non-ASCII characters is determined by the
13200: locale you are using. In the default @code{C} locale all non-ASCII
13201: characters are matched case-sensitively.
13202:
13203: @item system prompt:
13204: @cindex system prompt
13205: @cindex prompt
13206: @code{ ok} in interpret state, @code{ compiled} in compile state.
13207:
13208: @item division rounding:
13209: @cindex division rounding
1.166 anton 13210: The ordinary division words @code{/ mod /mod */ */mod} perform floored
13211: division (with the default installation of Gforth). You can check
13212: this with @code{s" floored" environment? drop .}. If you write
13213: programs that need a specific division rounding, best use
13214: @code{fm/mod} or @code{sm/rem} for portability.
1.1 anton 13215:
13216: @item values of @code{STATE} when true:
13217: @cindex @code{STATE} values
13218: -1.
13219:
13220: @item values returned after arithmetic overflow:
13221: On two's complement machines, arithmetic is performed modulo
13222: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
1.164 anton 13223: arithmetic (with appropriate mapping for signed types). Division by
13224: zero typically results in a @code{-55 throw} (Floating-point
13225: unidentified fault) or @code{-10 throw} (divide by zero). Integer
1.166 anton 13226: division overflow can result in these throws, or in @code{-11 throw};
13227: in @code{gforth-fast} division overflow and divide by zero may also
13228: result in returning bogus results without producing an exception.
1.1 anton 13229:
13230: @item whether the current definition can be found after @t{DOES>}:
13231: @cindex @t{DOES>}, visibility of current definition
13232: No.
13233:
13234: @end table
13235:
13236: @c ---------------------------------------------------------------------
13237: @node core-ambcond, core-other, core-idef, The Core Words
13238: @subsection Ambiguous conditions
13239: @c ---------------------------------------------------------------------
13240: @cindex core words, ambiguous conditions
13241: @cindex ambiguous conditions, core words
13242:
13243: @table @i
13244:
13245: @item a name is neither a word nor a number:
13246: @cindex name not found
1.26 crook 13247: @cindex undefined word
1.80 anton 13248: @code{-13 throw} (Undefined word).
1.1 anton 13249:
13250: @item a definition name exceeds the maximum length allowed:
1.26 crook 13251: @cindex word name too long
1.1 anton 13252: @code{-19 throw} (Word name too long)
13253:
13254: @item addressing a region not inside the various data spaces of the forth system:
13255: @cindex Invalid memory address
1.32 anton 13256: The stacks, code space and header space are accessible. Machine code space is
1.1 anton 13257: typically readable. Accessing other addresses gives results dependent on
13258: the operating system. On decent systems: @code{-9 throw} (Invalid memory
13259: address).
13260:
13261: @item argument type incompatible with parameter:
1.26 crook 13262: @cindex argument type mismatch
1.1 anton 13263: This is usually not caught. Some words perform checks, e.g., the control
13264: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
13265: mismatch).
13266:
13267: @item attempting to obtain the execution token of a word with undefined execution semantics:
13268: @cindex Interpreting a compile-only word, for @code{'} etc.
13269: @cindex execution token of words with undefined execution semantics
13270: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
13271: get an execution token for @code{compile-only-error} (which performs a
13272: @code{-14 throw} when executed).
13273:
13274: @item dividing by zero:
13275: @cindex dividing by zero
13276: @cindex floating point unidentified fault, integer division
1.80 anton 13277: On some platforms, this produces a @code{-10 throw} (Division by
1.24 anton 13278: zero); on other systems, this typically results in a @code{-55 throw}
13279: (Floating-point unidentified fault).
1.1 anton 13280:
13281: @item insufficient data stack or return stack space:
13282: @cindex insufficient data stack or return stack space
13283: @cindex stack overflow
1.26 crook 13284: @cindex address alignment exception, stack overflow
1.1 anton 13285: @cindex Invalid memory address, stack overflow
13286: Depending on the operating system, the installation, and the invocation
13287: of Gforth, this is either checked by the memory management hardware, or
1.24 anton 13288: it is not checked. If it is checked, you typically get a @code{-3 throw}
13289: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
13290: throw} (Invalid memory address) (depending on the platform and how you
13291: achieved the overflow) as soon as the overflow happens. If it is not
13292: checked, overflows typically result in mysterious illegal memory
13293: accesses, producing @code{-9 throw} (Invalid memory address) or
13294: @code{-23 throw} (Address alignment exception); they might also destroy
13295: the internal data structure of @code{ALLOCATE} and friends, resulting in
13296: various errors in these words.
1.1 anton 13297:
13298: @item insufficient space for loop control parameters:
13299: @cindex insufficient space for loop control parameters
1.80 anton 13300: Like other return stack overflows.
1.1 anton 13301:
13302: @item insufficient space in the dictionary:
13303: @cindex insufficient space in the dictionary
13304: @cindex dictionary overflow
1.12 anton 13305: If you try to allot (either directly with @code{allot}, or indirectly
13306: with @code{,}, @code{create} etc.) more memory than available in the
13307: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
13308: to access memory beyond the end of the dictionary, the results are
13309: similar to stack overflows.
1.1 anton 13310:
13311: @item interpreting a word with undefined interpretation semantics:
13312: @cindex interpreting a word with undefined interpretation semantics
13313: @cindex Interpreting a compile-only word
13314: For some words, we have defined interpretation semantics. For the
13315: others: @code{-14 throw} (Interpreting a compile-only word).
13316:
13317: @item modifying the contents of the input buffer or a string literal:
13318: @cindex modifying the contents of the input buffer or a string literal
13319: These are located in writable memory and can be modified.
13320:
13321: @item overflow of the pictured numeric output string:
13322: @cindex overflow of the pictured numeric output string
13323: @cindex pictured numeric output string, overflow
1.24 anton 13324: @code{-17 throw} (Pictured numeric ouput string overflow).
1.1 anton 13325:
13326: @item parsed string overflow:
13327: @cindex parsed string overflow
13328: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
13329:
13330: @item producing a result out of range:
13331: @cindex result out of range
13332: On two's complement machines, arithmetic is performed modulo
13333: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
1.166 anton 13334: arithmetic (with appropriate mapping for signed types). Division by
13335: zero typically results in a @code{-10 throw} (divide by zero) or
13336: @code{-55 throw} (floating point unidentified fault). Overflow on
13337: division may result in these errors or in @code{-11 throw} (result out
13338: of range). @code{Gforth-fast} may silently produce bogus results on
13339: division overflow or division by zero. @code{Convert} and
1.24 anton 13340: @code{>number} currently overflow silently.
1.1 anton 13341:
13342: @item reading from an empty data or return stack:
13343: @cindex stack empty
13344: @cindex stack underflow
1.24 anton 13345: @cindex return stack underflow
1.1 anton 13346: The data stack is checked by the outer (aka text) interpreter after
13347: every word executed. If it has underflowed, a @code{-4 throw} (Stack
13348: underflow) is performed. Apart from that, stacks may be checked or not,
1.24 anton 13349: depending on operating system, installation, and invocation. If they are
13350: caught by a check, they typically result in @code{-4 throw} (Stack
13351: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
13352: (Invalid memory address), depending on the platform and which stack
13353: underflows and by how much. Note that even if the system uses checking
13354: (through the MMU), your program may have to underflow by a significant
13355: number of stack items to trigger the reaction (the reason for this is
13356: that the MMU, and therefore the checking, works with a page-size
13357: granularity). If there is no checking, the symptoms resulting from an
13358: underflow are similar to those from an overflow. Unbalanced return
1.80 anton 13359: stack errors can result in a variety of symptoms, including @code{-9 throw}
1.24 anton 13360: (Invalid memory address) and Illegal Instruction (typically @code{-260
13361: throw}).
1.1 anton 13362:
13363: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
13364: @cindex unexpected end of the input buffer
13365: @cindex zero-length string as a name
13366: @cindex Attempt to use zero-length string as a name
13367: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
13368: use zero-length string as a name). Words like @code{'} probably will not
13369: find what they search. Note that it is possible to create zero-length
13370: names with @code{nextname} (should it not?).
13371:
13372: @item @code{>IN} greater than input buffer:
13373: @cindex @code{>IN} greater than input buffer
13374: The next invocation of a parsing word returns a string with length 0.
13375:
13376: @item @code{RECURSE} appears after @code{DOES>}:
13377: @cindex @code{RECURSE} appears after @code{DOES>}
13378: Compiles a recursive call to the defining word, not to the defined word.
13379:
13380: @item argument input source different than current input source for @code{RESTORE-INPUT}:
13381: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
1.26 crook 13382: @cindex argument type mismatch, @code{RESTORE-INPUT}
1.1 anton 13383: @cindex @code{RESTORE-INPUT}, Argument type mismatch
13384: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
13385: the end of the file was reached), its source-id may be
13386: reused. Therefore, restoring an input source specification referencing a
13387: closed file may lead to unpredictable results instead of a @code{-12
13388: THROW}.
13389:
13390: In the future, Gforth may be able to restore input source specifications
13391: from other than the current input source.
13392:
13393: @item data space containing definitions gets de-allocated:
13394: @cindex data space containing definitions gets de-allocated
13395: Deallocation with @code{allot} is not checked. This typically results in
13396: memory access faults or execution of illegal instructions.
13397:
13398: @item data space read/write with incorrect alignment:
13399: @cindex data space read/write with incorrect alignment
13400: @cindex alignment faults
1.26 crook 13401: @cindex address alignment exception
1.1 anton 13402: Processor-dependent. Typically results in a @code{-23 throw} (Address
1.12 anton 13403: alignment exception). Under Linux-Intel on a 486 or later processor with
1.1 anton 13404: alignment turned on, incorrect alignment results in a @code{-9 throw}
13405: (Invalid memory address). There are reportedly some processors with
1.12 anton 13406: alignment restrictions that do not report violations.
1.1 anton 13407:
13408: @item data space pointer not properly aligned, @code{,}, @code{C,}:
13409: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
13410: Like other alignment errors.
13411:
13412: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
13413: Like other stack underflows.
13414:
13415: @item loop control parameters not available:
13416: @cindex loop control parameters not available
13417: Not checked. The counted loop words simply assume that the top of return
13418: stack items are loop control parameters and behave accordingly.
13419:
13420: @item most recent definition does not have a name (@code{IMMEDIATE}):
13421: @cindex most recent definition does not have a name (@code{IMMEDIATE})
13422: @cindex last word was headerless
13423: @code{abort" last word was headerless"}.
13424:
13425: @item name not defined by @code{VALUE} used by @code{TO}:
13426: @cindex name not defined by @code{VALUE} used by @code{TO}
13427: @cindex @code{TO} on non-@code{VALUE}s
13428: @cindex Invalid name argument, @code{TO}
13429: @code{-32 throw} (Invalid name argument) (unless name is a local or was
13430: defined by @code{CONSTANT}; in the latter case it just changes the constant).
13431:
13432: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
13433: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
1.26 crook 13434: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
1.1 anton 13435: @code{-13 throw} (Undefined word)
13436:
13437: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
13438: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
13439: Gforth behaves as if they were of the same type. I.e., you can predict
13440: the behaviour by interpreting all parameters as, e.g., signed.
13441:
13442: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
13443: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
13444: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
13445: compilation semantics of @code{TO}.
13446:
13447: @item String longer than a counted string returned by @code{WORD}:
1.26 crook 13448: @cindex string longer than a counted string returned by @code{WORD}
1.1 anton 13449: @cindex @code{WORD}, string overflow
13450: Not checked. The string will be ok, but the count will, of course,
13451: contain only the least significant bits of the length.
13452:
13453: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
13454: @cindex @code{LSHIFT}, large shift counts
13455: @cindex @code{RSHIFT}, large shift counts
13456: Processor-dependent. Typical behaviours are returning 0 and using only
13457: the low bits of the shift count.
13458:
13459: @item word not defined via @code{CREATE}:
13460: @cindex @code{>BODY} of non-@code{CREATE}d words
13461: @code{>BODY} produces the PFA of the word no matter how it was defined.
13462:
13463: @cindex @code{DOES>} of non-@code{CREATE}d words
13464: @code{DOES>} changes the execution semantics of the last defined word no
13465: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
13466: @code{CREATE , DOES>}.
13467:
13468: @item words improperly used outside @code{<#} and @code{#>}:
13469: Not checked. As usual, you can expect memory faults.
13470:
13471: @end table
13472:
13473:
13474: @c ---------------------------------------------------------------------
13475: @node core-other, , core-ambcond, The Core Words
13476: @subsection Other system documentation
13477: @c ---------------------------------------------------------------------
13478: @cindex other system documentation, core words
13479: @cindex core words, other system documentation
13480:
13481: @table @i
13482: @item nonstandard words using @code{PAD}:
13483: @cindex @code{PAD} use by nonstandard words
13484: None.
13485:
13486: @item operator's terminal facilities available:
13487: @cindex operator's terminal facilities available
1.80 anton 13488: After processing the OS's command line, Gforth goes into interactive mode,
1.1 anton 13489: and you can give commands to Gforth interactively. The actual facilities
13490: available depend on how you invoke Gforth.
13491:
13492: @item program data space available:
13493: @cindex program data space available
13494: @cindex data space available
13495: @code{UNUSED .} gives the remaining dictionary space. The total
13496: dictionary space can be specified with the @code{-m} switch
13497: (@pxref{Invoking Gforth}) when Gforth starts up.
13498:
13499: @item return stack space available:
13500: @cindex return stack space available
13501: You can compute the total return stack space in cells with
13502: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
13503: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
13504:
13505: @item stack space available:
13506: @cindex stack space available
13507: You can compute the total data stack space in cells with
13508: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
13509: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
13510:
13511: @item system dictionary space required, in address units:
13512: @cindex system dictionary space required, in address units
13513: Type @code{here forthstart - .} after startup. At the time of this
13514: writing, this gives 80080 (bytes) on a 32-bit system.
13515: @end table
13516:
13517:
13518: @c =====================================================================
13519: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
13520: @section The optional Block word set
13521: @c =====================================================================
13522: @cindex system documentation, block words
13523: @cindex block words, system documentation
13524:
13525: @menu
13526: * block-idef:: Implementation Defined Options
13527: * block-ambcond:: Ambiguous Conditions
13528: * block-other:: Other System Documentation
13529: @end menu
13530:
13531:
13532: @c ---------------------------------------------------------------------
13533: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
13534: @subsection Implementation Defined Options
13535: @c ---------------------------------------------------------------------
13536: @cindex implementation-defined options, block words
13537: @cindex block words, implementation-defined options
13538:
13539: @table @i
13540: @item the format for display by @code{LIST}:
13541: @cindex @code{LIST} display format
13542: First the screen number is displayed, then 16 lines of 64 characters,
13543: each line preceded by the line number.
13544:
13545: @item the length of a line affected by @code{\}:
13546: @cindex length of a line affected by @code{\}
13547: @cindex @code{\}, line length in blocks
13548: 64 characters.
13549: @end table
13550:
13551:
13552: @c ---------------------------------------------------------------------
13553: @node block-ambcond, block-other, block-idef, The optional Block word set
13554: @subsection Ambiguous conditions
13555: @c ---------------------------------------------------------------------
13556: @cindex block words, ambiguous conditions
13557: @cindex ambiguous conditions, block words
13558:
13559: @table @i
13560: @item correct block read was not possible:
13561: @cindex block read not possible
13562: Typically results in a @code{throw} of some OS-derived value (between
13563: -512 and -2048). If the blocks file was just not long enough, blanks are
13564: supplied for the missing portion.
13565:
13566: @item I/O exception in block transfer:
13567: @cindex I/O exception in block transfer
13568: @cindex block transfer, I/O exception
13569: Typically results in a @code{throw} of some OS-derived value (between
13570: -512 and -2048).
13571:
13572: @item invalid block number:
13573: @cindex invalid block number
13574: @cindex block number invalid
13575: @code{-35 throw} (Invalid block number)
13576:
13577: @item a program directly alters the contents of @code{BLK}:
13578: @cindex @code{BLK}, altering @code{BLK}
13579: The input stream is switched to that other block, at the same
13580: position. If the storing to @code{BLK} happens when interpreting
13581: non-block input, the system will get quite confused when the block ends.
13582:
13583: @item no current block buffer for @code{UPDATE}:
13584: @cindex @code{UPDATE}, no current block buffer
13585: @code{UPDATE} has no effect.
13586:
13587: @end table
13588:
13589: @c ---------------------------------------------------------------------
13590: @node block-other, , block-ambcond, The optional Block word set
13591: @subsection Other system documentation
13592: @c ---------------------------------------------------------------------
13593: @cindex other system documentation, block words
13594: @cindex block words, other system documentation
13595:
13596: @table @i
13597: @item any restrictions a multiprogramming system places on the use of buffer addresses:
13598: No restrictions (yet).
13599:
13600: @item the number of blocks available for source and data:
13601: depends on your disk space.
13602:
13603: @end table
13604:
13605:
13606: @c =====================================================================
13607: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
13608: @section The optional Double Number word set
13609: @c =====================================================================
13610: @cindex system documentation, double words
13611: @cindex double words, system documentation
13612:
13613: @menu
13614: * double-ambcond:: Ambiguous Conditions
13615: @end menu
13616:
13617:
13618: @c ---------------------------------------------------------------------
13619: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
13620: @subsection Ambiguous conditions
13621: @c ---------------------------------------------------------------------
13622: @cindex double words, ambiguous conditions
13623: @cindex ambiguous conditions, double words
13624:
13625: @table @i
1.29 crook 13626: @item @i{d} outside of range of @i{n} in @code{D>S}:
13627: @cindex @code{D>S}, @i{d} out of range of @i{n}
13628: The least significant cell of @i{d} is produced.
1.1 anton 13629:
13630: @end table
13631:
13632:
13633: @c =====================================================================
13634: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
13635: @section The optional Exception word set
13636: @c =====================================================================
13637: @cindex system documentation, exception words
13638: @cindex exception words, system documentation
13639:
13640: @menu
13641: * exception-idef:: Implementation Defined Options
13642: @end menu
13643:
13644:
13645: @c ---------------------------------------------------------------------
13646: @node exception-idef, , The optional Exception word set, The optional Exception word set
13647: @subsection Implementation Defined Options
13648: @c ---------------------------------------------------------------------
13649: @cindex implementation-defined options, exception words
13650: @cindex exception words, implementation-defined options
13651:
13652: @table @i
13653: @item @code{THROW}-codes used in the system:
13654: @cindex @code{THROW}-codes used in the system
13655: The codes -256@minus{}-511 are used for reporting signals. The mapping
1.29 crook 13656: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
1.1 anton 13657: codes -512@minus{}-2047 are used for OS errors (for file and memory
13658: allocation operations). The mapping from OS error numbers to throw codes
13659: is -512@minus{}@code{errno}. One side effect of this mapping is that
13660: undefined OS errors produce a message with a strange number; e.g.,
13661: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
13662: @end table
13663:
13664: @c =====================================================================
13665: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
13666: @section The optional Facility word set
13667: @c =====================================================================
13668: @cindex system documentation, facility words
13669: @cindex facility words, system documentation
13670:
13671: @menu
13672: * facility-idef:: Implementation Defined Options
13673: * facility-ambcond:: Ambiguous Conditions
13674: @end menu
13675:
13676:
13677: @c ---------------------------------------------------------------------
13678: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
13679: @subsection Implementation Defined Options
13680: @c ---------------------------------------------------------------------
13681: @cindex implementation-defined options, facility words
13682: @cindex facility words, implementation-defined options
13683:
13684: @table @i
13685: @item encoding of keyboard events (@code{EKEY}):
13686: @cindex keyboard events, encoding in @code{EKEY}
13687: @cindex @code{EKEY}, encoding of keyboard events
1.40 anton 13688: Keys corresponding to ASCII characters are encoded as ASCII characters.
1.41 anton 13689: Other keys are encoded with the constants @code{k-left}, @code{k-right},
13690: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
13691: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
13692: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
1.40 anton 13693:
1.1 anton 13694:
13695: @item duration of a system clock tick:
13696: @cindex duration of a system clock tick
13697: @cindex clock tick duration
13698: System dependent. With respect to @code{MS}, the time is specified in
13699: microseconds. How well the OS and the hardware implement this, is
13700: another question.
13701:
13702: @item repeatability to be expected from the execution of @code{MS}:
13703: @cindex repeatability to be expected from the execution of @code{MS}
13704: @cindex @code{MS}, repeatability to be expected
13705: System dependent. On Unix, a lot depends on load. If the system is
13706: lightly loaded, and the delay is short enough that Gforth does not get
13707: swapped out, the performance should be acceptable. Under MS-DOS and
13708: other single-tasking systems, it should be good.
13709:
13710: @end table
13711:
13712:
13713: @c ---------------------------------------------------------------------
13714: @node facility-ambcond, , facility-idef, The optional Facility word set
13715: @subsection Ambiguous conditions
13716: @c ---------------------------------------------------------------------
13717: @cindex facility words, ambiguous conditions
13718: @cindex ambiguous conditions, facility words
13719:
13720: @table @i
13721: @item @code{AT-XY} can't be performed on user output device:
13722: @cindex @code{AT-XY} can't be performed on user output device
13723: Largely terminal dependent. No range checks are done on the arguments.
13724: No errors are reported. You may see some garbage appearing, you may see
13725: simply nothing happen.
13726:
13727: @end table
13728:
13729:
13730: @c =====================================================================
13731: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
13732: @section The optional File-Access word set
13733: @c =====================================================================
13734: @cindex system documentation, file words
13735: @cindex file words, system documentation
13736:
13737: @menu
13738: * file-idef:: Implementation Defined Options
13739: * file-ambcond:: Ambiguous Conditions
13740: @end menu
13741:
13742: @c ---------------------------------------------------------------------
13743: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
13744: @subsection Implementation Defined Options
13745: @c ---------------------------------------------------------------------
13746: @cindex implementation-defined options, file words
13747: @cindex file words, implementation-defined options
13748:
13749: @table @i
13750: @item file access methods used:
13751: @cindex file access methods used
13752: @code{R/O}, @code{R/W} and @code{BIN} work as you would
13753: expect. @code{W/O} translates into the C file opening mode @code{w} (or
13754: @code{wb}): The file is cleared, if it exists, and created, if it does
13755: not (with both @code{open-file} and @code{create-file}). Under Unix
13756: @code{create-file} creates a file with 666 permissions modified by your
13757: umask.
13758:
13759: @item file exceptions:
13760: @cindex file exceptions
13761: The file words do not raise exceptions (except, perhaps, memory access
13762: faults when you pass illegal addresses or file-ids).
13763:
13764: @item file line terminator:
13765: @cindex file line terminator
13766: System-dependent. Gforth uses C's newline character as line
13767: terminator. What the actual character code(s) of this are is
13768: system-dependent.
13769:
13770: @item file name format:
13771: @cindex file name format
13772: System dependent. Gforth just uses the file name format of your OS.
13773:
13774: @item information returned by @code{FILE-STATUS}:
13775: @cindex @code{FILE-STATUS}, returned information
13776: @code{FILE-STATUS} returns the most powerful file access mode allowed
13777: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
13778: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
13779: along with the returned mode.
13780:
13781: @item input file state after an exception when including source:
13782: @cindex exception when including source
13783: All files that are left via the exception are closed.
13784:
1.29 crook 13785: @item @i{ior} values and meaning:
13786: @cindex @i{ior} values and meaning
1.68 anton 13787: @cindex @i{wior} values and meaning
1.29 crook 13788: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 13789: intended as throw codes. They typically are in the range
13790: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 13791: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 13792:
13793: @item maximum depth of file input nesting:
13794: @cindex maximum depth of file input nesting
13795: @cindex file input nesting, maximum depth
13796: limited by the amount of return stack, locals/TIB stack, and the number
13797: of open files available. This should not give you troubles.
13798:
13799: @item maximum size of input line:
13800: @cindex maximum size of input line
13801: @cindex input line size, maximum
13802: @code{/line}. Currently 255.
13803:
13804: @item methods of mapping block ranges to files:
13805: @cindex mapping block ranges to files
13806: @cindex files containing blocks
13807: @cindex blocks in files
13808: By default, blocks are accessed in the file @file{blocks.fb} in the
13809: current working directory. The file can be switched with @code{USE}.
13810:
13811: @item number of string buffers provided by @code{S"}:
13812: @cindex @code{S"}, number of string buffers
13813: 1
13814:
13815: @item size of string buffer used by @code{S"}:
13816: @cindex @code{S"}, size of string buffer
13817: @code{/line}. currently 255.
13818:
13819: @end table
13820:
13821: @c ---------------------------------------------------------------------
13822: @node file-ambcond, , file-idef, The optional File-Access word set
13823: @subsection Ambiguous conditions
13824: @c ---------------------------------------------------------------------
13825: @cindex file words, ambiguous conditions
13826: @cindex ambiguous conditions, file words
13827:
13828: @table @i
13829: @item attempting to position a file outside its boundaries:
13830: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
13831: @code{REPOSITION-FILE} is performed as usual: Afterwards,
13832: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
13833:
13834: @item attempting to read from file positions not yet written:
13835: @cindex reading from file positions not yet written
13836: End-of-file, i.e., zero characters are read and no error is reported.
13837:
1.29 crook 13838: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
13839: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
1.1 anton 13840: An appropriate exception may be thrown, but a memory fault or other
13841: problem is more probable.
13842:
1.29 crook 13843: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
13844: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
13845: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
13846: The @i{ior} produced by the operation, that discovered the problem, is
1.1 anton 13847: thrown.
13848:
13849: @item named file cannot be opened (@code{INCLUDED}):
13850: @cindex @code{INCLUDED}, named file cannot be opened
1.29 crook 13851: The @i{ior} produced by @code{open-file} is thrown.
1.1 anton 13852:
13853: @item requesting an unmapped block number:
13854: @cindex unmapped block numbers
13855: There are no unmapped legal block numbers. On some operating systems,
13856: writing a block with a large number may overflow the file system and
13857: have an error message as consequence.
13858:
13859: @item using @code{source-id} when @code{blk} is non-zero:
13860: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
13861: @code{source-id} performs its function. Typically it will give the id of
13862: the source which loaded the block. (Better ideas?)
13863:
13864: @end table
13865:
13866:
13867: @c =====================================================================
13868: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
13869: @section The optional Floating-Point word set
13870: @c =====================================================================
13871: @cindex system documentation, floating-point words
13872: @cindex floating-point words, system documentation
13873:
13874: @menu
13875: * floating-idef:: Implementation Defined Options
13876: * floating-ambcond:: Ambiguous Conditions
13877: @end menu
13878:
13879:
13880: @c ---------------------------------------------------------------------
13881: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
13882: @subsection Implementation Defined Options
13883: @c ---------------------------------------------------------------------
13884: @cindex implementation-defined options, floating-point words
13885: @cindex floating-point words, implementation-defined options
13886:
13887: @table @i
13888: @item format and range of floating point numbers:
13889: @cindex format and range of floating point numbers
13890: @cindex floating point numbers, format and range
13891: System-dependent; the @code{double} type of C.
13892:
1.29 crook 13893: @item results of @code{REPRESENT} when @i{float} is out of range:
13894: @cindex @code{REPRESENT}, results when @i{float} is out of range
1.1 anton 13895: System dependent; @code{REPRESENT} is implemented using the C library
13896: function @code{ecvt()} and inherits its behaviour in this respect.
13897:
13898: @item rounding or truncation of floating-point numbers:
13899: @cindex rounding of floating-point numbers
13900: @cindex truncation of floating-point numbers
13901: @cindex floating-point numbers, rounding or truncation
13902: System dependent; the rounding behaviour is inherited from the hosting C
13903: compiler. IEEE-FP-based (i.e., most) systems by default round to
13904: nearest, and break ties by rounding to even (i.e., such that the last
13905: bit of the mantissa is 0).
13906:
13907: @item size of floating-point stack:
13908: @cindex floating-point stack size
13909: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
13910: the floating-point stack (in floats). You can specify this on startup
13911: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
13912:
13913: @item width of floating-point stack:
13914: @cindex floating-point stack width
13915: @code{1 floats}.
13916:
13917: @end table
13918:
13919:
13920: @c ---------------------------------------------------------------------
13921: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
13922: @subsection Ambiguous conditions
13923: @c ---------------------------------------------------------------------
13924: @cindex floating-point words, ambiguous conditions
13925: @cindex ambiguous conditions, floating-point words
13926:
13927: @table @i
13928: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
13929: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
13930: System-dependent. Typically results in a @code{-23 THROW} like other
13931: alignment violations.
13932:
13933: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
13934: @cindex @code{f@@} used with an address that is not float aligned
13935: @cindex @code{f!} used with an address that is not float aligned
13936: System-dependent. Typically results in a @code{-23 THROW} like other
13937: alignment violations.
13938:
13939: @item floating-point result out of range:
13940: @cindex floating-point result out of range
1.80 anton 13941: System-dependent. Can result in a @code{-43 throw} (floating point
13942: overflow), @code{-54 throw} (floating point underflow), @code{-41 throw}
13943: (floating point inexact result), @code{-55 THROW} (Floating-point
1.1 anton 13944: unidentified fault), or can produce a special value representing, e.g.,
13945: Infinity.
13946:
13947: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
13948: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
13949: System-dependent. Typically results in an alignment fault like other
13950: alignment violations.
13951:
1.35 anton 13952: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
13953: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
1.1 anton 13954: The floating-point number is converted into decimal nonetheless.
13955:
13956: @item Both arguments are equal to zero (@code{FATAN2}):
13957: @cindex @code{FATAN2}, both arguments are equal to zero
13958: System-dependent. @code{FATAN2} is implemented using the C library
13959: function @code{atan2()}.
13960:
1.29 crook 13961: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
13962: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
13963: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
1.1 anton 13964: because of small errors and the tan will be a very large (or very small)
13965: but finite number.
13966:
1.29 crook 13967: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
13968: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
1.1 anton 13969: The result is rounded to the nearest float.
13970:
13971: @item dividing by zero:
13972: @cindex dividing by zero, floating-point
13973: @cindex floating-point dividing by zero
13974: @cindex floating-point unidentified fault, FP divide-by-zero
1.80 anton 13975: Platform-dependent; can produce an Infinity, NaN, @code{-42 throw}
13976: (floating point divide by zero) or @code{-55 throw} (Floating-point
13977: unidentified fault).
1.1 anton 13978:
13979: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
13980: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
13981: System dependent. On IEEE-FP based systems the number is converted into
13982: an infinity.
13983:
1.29 crook 13984: @item @i{float}<1 (@code{FACOSH}):
13985: @cindex @code{FACOSH}, @i{float}<1
1.1 anton 13986: @cindex floating-point unidentified fault, @code{FACOSH}
1.80 anton 13987: Platform-dependent; on IEEE-FP systems typically produces a NaN.
1.1 anton 13988:
1.29 crook 13989: @item @i{float}=<-1 (@code{FLNP1}):
13990: @cindex @code{FLNP1}, @i{float}=<-1
1.1 anton 13991: @cindex floating-point unidentified fault, @code{FLNP1}
1.80 anton 13992: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13993: negative infinity for @i{float}=-1).
1.1 anton 13994:
1.29 crook 13995: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
13996: @cindex @code{FLN}, @i{float}=<0
13997: @cindex @code{FLOG}, @i{float}=<0
1.1 anton 13998: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
1.80 anton 13999: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
14000: negative infinity for @i{float}=0).
1.1 anton 14001:
1.29 crook 14002: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
14003: @cindex @code{FASINH}, @i{float}<0
14004: @cindex @code{FSQRT}, @i{float}<0
1.1 anton 14005: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
1.80 anton 14006: Platform-dependent; for @code{fsqrt} this typically gives a NaN, for
14007: @code{fasinh} some platforms produce a NaN, others a number (bug in the
14008: C library?).
1.1 anton 14009:
1.29 crook 14010: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
14011: @cindex @code{FACOS}, |@i{float}|>1
14012: @cindex @code{FASIN}, |@i{float}|>1
14013: @cindex @code{FATANH}, |@i{float}|>1
1.1 anton 14014: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
1.80 anton 14015: Platform-dependent; IEEE-FP systems typically produce a NaN.
1.1 anton 14016:
1.29 crook 14017: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
14018: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
1.1 anton 14019: @cindex floating-point unidentified fault, @code{F>D}
1.80 anton 14020: Platform-dependent; typically, some double number is produced and no
14021: error is reported.
1.1 anton 14022:
14023: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
14024: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
1.80 anton 14025: @code{Precision} characters of the numeric output area are used. If
14026: @code{precision} is too high, these words will smash the data or code
14027: close to @code{here}.
1.1 anton 14028: @end table
14029:
14030: @c =====================================================================
14031: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
14032: @section The optional Locals word set
14033: @c =====================================================================
14034: @cindex system documentation, locals words
14035: @cindex locals words, system documentation
14036:
14037: @menu
14038: * locals-idef:: Implementation Defined Options
14039: * locals-ambcond:: Ambiguous Conditions
14040: @end menu
14041:
14042:
14043: @c ---------------------------------------------------------------------
14044: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
14045: @subsection Implementation Defined Options
14046: @c ---------------------------------------------------------------------
14047: @cindex implementation-defined options, locals words
14048: @cindex locals words, implementation-defined options
14049:
14050: @table @i
14051: @item maximum number of locals in a definition:
14052: @cindex maximum number of locals in a definition
14053: @cindex locals, maximum number in a definition
14054: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
14055: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
14056: characters. The number of locals in a definition is bounded by the size
14057: of locals-buffer, which contains the names of the locals.
14058:
14059: @end table
14060:
14061:
14062: @c ---------------------------------------------------------------------
14063: @node locals-ambcond, , locals-idef, The optional Locals word set
14064: @subsection Ambiguous conditions
14065: @c ---------------------------------------------------------------------
14066: @cindex locals words, ambiguous conditions
14067: @cindex ambiguous conditions, locals words
14068:
14069: @table @i
14070: @item executing a named local in interpretation state:
14071: @cindex local in interpretation state
14072: @cindex Interpreting a compile-only word, for a local
14073: Locals have no interpretation semantics. If you try to perform the
14074: interpretation semantics, you will get a @code{-14 throw} somewhere
14075: (Interpreting a compile-only word). If you perform the compilation
14076: semantics, the locals access will be compiled (irrespective of state).
14077:
1.29 crook 14078: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
1.1 anton 14079: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
14080: @cindex @code{TO} on non-@code{VALUE}s and non-locals
14081: @cindex Invalid name argument, @code{TO}
14082: @code{-32 throw} (Invalid name argument)
14083:
14084: @end table
14085:
14086:
14087: @c =====================================================================
14088: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
14089: @section The optional Memory-Allocation word set
14090: @c =====================================================================
14091: @cindex system documentation, memory-allocation words
14092: @cindex memory-allocation words, system documentation
14093:
14094: @menu
14095: * memory-idef:: Implementation Defined Options
14096: @end menu
14097:
14098:
14099: @c ---------------------------------------------------------------------
14100: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
14101: @subsection Implementation Defined Options
14102: @c ---------------------------------------------------------------------
14103: @cindex implementation-defined options, memory-allocation words
14104: @cindex memory-allocation words, implementation-defined options
14105:
14106: @table @i
1.29 crook 14107: @item values and meaning of @i{ior}:
14108: @cindex @i{ior} values and meaning
14109: The @i{ior}s returned by the file and memory allocation words are
1.1 anton 14110: intended as throw codes. They typically are in the range
14111: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
1.29 crook 14112: @i{ior}s is -512@minus{}@i{errno}.
1.1 anton 14113:
14114: @end table
14115:
14116: @c =====================================================================
14117: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
14118: @section The optional Programming-Tools word set
14119: @c =====================================================================
14120: @cindex system documentation, programming-tools words
14121: @cindex programming-tools words, system documentation
14122:
14123: @menu
14124: * programming-idef:: Implementation Defined Options
14125: * programming-ambcond:: Ambiguous Conditions
14126: @end menu
14127:
14128:
14129: @c ---------------------------------------------------------------------
14130: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
14131: @subsection Implementation Defined Options
14132: @c ---------------------------------------------------------------------
14133: @cindex implementation-defined options, programming-tools words
14134: @cindex programming-tools words, implementation-defined options
14135:
14136: @table @i
14137: @item ending sequence for input following @code{;CODE} and @code{CODE}:
14138: @cindex @code{;CODE} ending sequence
14139: @cindex @code{CODE} ending sequence
14140: @code{END-CODE}
14141:
14142: @item manner of processing input following @code{;CODE} and @code{CODE}:
14143: @cindex @code{;CODE}, processing input
14144: @cindex @code{CODE}, processing input
14145: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
14146: the input is processed by the text interpreter, (starting) in interpret
14147: state.
14148:
14149: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
14150: @cindex @code{ASSEMBLER}, search order capability
14151: The ANS Forth search order word set.
14152:
14153: @item source and format of display by @code{SEE}:
14154: @cindex @code{SEE}, source and format of output
1.80 anton 14155: The source for @code{see} is the executable code used by the inner
1.1 anton 14156: interpreter. The current @code{see} tries to output Forth source code
1.80 anton 14157: (and on some platforms, assembly code for primitives) as well as
14158: possible.
1.1 anton 14159:
14160: @end table
14161:
14162: @c ---------------------------------------------------------------------
14163: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
14164: @subsection Ambiguous conditions
14165: @c ---------------------------------------------------------------------
14166: @cindex programming-tools words, ambiguous conditions
14167: @cindex ambiguous conditions, programming-tools words
14168:
14169: @table @i
14170:
1.21 crook 14171: @item deleting the compilation word list (@code{FORGET}):
14172: @cindex @code{FORGET}, deleting the compilation word list
1.1 anton 14173: Not implemented (yet).
14174:
1.29 crook 14175: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
14176: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
14177: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
1.1 anton 14178: @cindex control-flow stack underflow
14179: This typically results in an @code{abort"} with a descriptive error
14180: message (may change into a @code{-22 throw} (Control structure mismatch)
14181: in the future). You may also get a memory access error. If you are
14182: unlucky, this ambiguous condition is not caught.
14183:
1.29 crook 14184: @item @i{name} can't be found (@code{FORGET}):
14185: @cindex @code{FORGET}, @i{name} can't be found
1.1 anton 14186: Not implemented (yet).
14187:
1.29 crook 14188: @item @i{name} not defined via @code{CREATE}:
14189: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
1.1 anton 14190: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
14191: the execution semantics of the last defined word no matter how it was
14192: defined.
14193:
14194: @item @code{POSTPONE} applied to @code{[IF]}:
14195: @cindex @code{POSTPONE} applied to @code{[IF]}
14196: @cindex @code{[IF]} and @code{POSTPONE}
14197: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
14198: equivalent to @code{[IF]}.
14199:
14200: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
14201: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
14202: Continue in the same state of conditional compilation in the next outer
14203: input source. Currently there is no warning to the user about this.
14204:
14205: @item removing a needed definition (@code{FORGET}):
14206: @cindex @code{FORGET}, removing a needed definition
14207: Not implemented (yet).
14208:
14209: @end table
14210:
14211:
14212: @c =====================================================================
14213: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
14214: @section The optional Search-Order word set
14215: @c =====================================================================
14216: @cindex system documentation, search-order words
14217: @cindex search-order words, system documentation
14218:
14219: @menu
14220: * search-idef:: Implementation Defined Options
14221: * search-ambcond:: Ambiguous Conditions
14222: @end menu
14223:
14224:
14225: @c ---------------------------------------------------------------------
14226: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
14227: @subsection Implementation Defined Options
14228: @c ---------------------------------------------------------------------
14229: @cindex implementation-defined options, search-order words
14230: @cindex search-order words, implementation-defined options
14231:
14232: @table @i
14233: @item maximum number of word lists in search order:
14234: @cindex maximum number of word lists in search order
14235: @cindex search order, maximum depth
14236: @code{s" wordlists" environment? drop .}. Currently 16.
14237:
14238: @item minimum search order:
14239: @cindex minimum search order
14240: @cindex search order, minimum
14241: @code{root root}.
14242:
14243: @end table
14244:
14245: @c ---------------------------------------------------------------------
14246: @node search-ambcond, , search-idef, The optional Search-Order word set
14247: @subsection Ambiguous conditions
14248: @c ---------------------------------------------------------------------
14249: @cindex search-order words, ambiguous conditions
14250: @cindex ambiguous conditions, search-order words
14251:
14252: @table @i
1.21 crook 14253: @item changing the compilation word list (during compilation):
14254: @cindex changing the compilation word list (during compilation)
14255: @cindex compilation word list, change before definition ends
14256: The word is entered into the word list that was the compilation word list
1.1 anton 14257: at the start of the definition. Any changes to the name field (e.g.,
14258: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
1.116 anton 14259: are applied to the latest defined word (as reported by @code{latest} or
14260: @code{latestxt}), if possible, irrespective of the compilation word list.
1.1 anton 14261:
14262: @item search order empty (@code{previous}):
14263: @cindex @code{previous}, search order empty
1.26 crook 14264: @cindex vocstack empty, @code{previous}
1.1 anton 14265: @code{abort" Vocstack empty"}.
14266:
14267: @item too many word lists in search order (@code{also}):
14268: @cindex @code{also}, too many word lists in search order
1.26 crook 14269: @cindex vocstack full, @code{also}
1.1 anton 14270: @code{abort" Vocstack full"}.
14271:
14272: @end table
14273:
14274: @c ***************************************************************
1.65 anton 14275: @node Standard vs Extensions, Model, ANS conformance, Top
14276: @chapter Should I use Gforth extensions?
14277: @cindex Gforth extensions
14278:
14279: As you read through the rest of this manual, you will see documentation
14280: for @i{Standard} words, and documentation for some appealing Gforth
14281: @i{extensions}. You might ask yourself the question: @i{``Should I
14282: restrict myself to the standard, or should I use the extensions?''}
14283:
14284: The answer depends on the goals you have for the program you are working
14285: on:
14286:
14287: @itemize @bullet
14288:
14289: @item Is it just for yourself or do you want to share it with others?
14290:
14291: @item
14292: If you want to share it, do the others all use Gforth?
14293:
14294: @item
14295: If it is just for yourself, do you want to restrict yourself to Gforth?
14296:
14297: @end itemize
14298:
14299: If restricting the program to Gforth is ok, then there is no reason not
14300: to use extensions. It is still a good idea to keep to the standard
14301: where it is easy, in case you want to reuse these parts in another
14302: program that you want to be portable.
14303:
14304: If you want to be able to port the program to other Forth systems, there
14305: are the following points to consider:
14306:
14307: @itemize @bullet
14308:
14309: @item
14310: Most Forth systems that are being maintained support the ANS Forth
14311: standard. So if your program complies with the standard, it will be
14312: portable among many systems.
14313:
14314: @item
14315: A number of the Gforth extensions can be implemented in ANS Forth using
14316: public-domain files provided in the @file{compat/} directory. These are
14317: mentioned in the text in passing. There is no reason not to use these
14318: extensions, your program will still be ANS Forth compliant; just include
14319: the appropriate compat files with your program.
14320:
14321: @item
14322: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
14323: analyse your program and determine what non-Standard words it relies
14324: upon. However, it does not check whether you use standard words in a
14325: non-standard way.
14326:
14327: @item
14328: Some techniques are not standardized by ANS Forth, and are hard or
14329: impossible to implement in a standard way, but can be implemented in
14330: most Forth systems easily, and usually in similar ways (e.g., accessing
14331: word headers). Forth has a rich historical precedent for programmers
14332: taking advantage of implementation-dependent features of their tools
14333: (for example, relying on a knowledge of the dictionary
14334: structure). Sometimes these techniques are necessary to extract every
14335: last bit of performance from the hardware, sometimes they are just a
14336: programming shorthand.
14337:
14338: @item
14339: Does using a Gforth extension save more work than the porting this part
14340: to other Forth systems (if any) will cost?
14341:
14342: @item
14343: Is the additional functionality worth the reduction in portability and
14344: the additional porting problems?
14345:
14346: @end itemize
14347:
14348: In order to perform these consideratios, you need to know what's
14349: standard and what's not. This manual generally states if something is
1.81 anton 14350: non-standard, but the authoritative source is the
14351: @uref{http://www.taygeta.com/forth/dpans.html,standard document}.
1.65 anton 14352: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
14353: into the thought processes of the technical committee.
14354:
14355: Note also that portability between Forth systems is not the only
14356: portability issue; there is also the issue of portability between
14357: different platforms (processor/OS combinations).
14358:
14359: @c ***************************************************************
14360: @node Model, Integrating Gforth, Standard vs Extensions, Top
1.1 anton 14361: @chapter Model
14362:
14363: This chapter has yet to be written. It will contain information, on
14364: which internal structures you can rely.
14365:
14366: @c ***************************************************************
14367: @node Integrating Gforth, Emacs and Gforth, Model, Top
14368: @chapter Integrating Gforth into C programs
14369:
14370: This is not yet implemented.
14371:
14372: Several people like to use Forth as scripting language for applications
14373: that are otherwise written in C, C++, or some other language.
14374:
14375: The Forth system ATLAST provides facilities for embedding it into
14376: applications; unfortunately it has several disadvantages: most
14377: importantly, it is not based on ANS Forth, and it is apparently dead
14378: (i.e., not developed further and not supported). The facilities
1.21 crook 14379: provided by Gforth in this area are inspired by ATLAST's facilities, so
1.1 anton 14380: making the switch should not be hard.
14381:
14382: We also tried to design the interface such that it can easily be
14383: implemented by other Forth systems, so that we may one day arrive at a
14384: standardized interface. Such a standard interface would allow you to
14385: replace the Forth system without having to rewrite C code.
14386:
14387: You embed the Gforth interpreter by linking with the library
14388: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
14389: global symbols in this library that belong to the interface, have the
14390: prefix @code{forth_}. (Global symbols that are used internally have the
14391: prefix @code{gforth_}).
14392:
14393: You can include the declarations of Forth types and the functions and
14394: variables of the interface with @code{#include <forth.h>}.
14395:
14396: Types.
14397:
14398: Variables.
14399:
14400: Data and FP Stack pointer. Area sizes.
14401:
14402: functions.
14403:
14404: forth_init(imagefile)
14405: forth_evaluate(string) exceptions?
14406: forth_goto(address) (or forth_execute(xt)?)
14407: forth_continue() (a corountining mechanism)
14408:
14409: Adding primitives.
14410:
14411: No checking.
14412:
14413: Signals?
14414:
14415: Accessing the Stacks
14416:
1.26 crook 14417: @c ******************************************************************
1.1 anton 14418: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
14419: @chapter Emacs and Gforth
14420: @cindex Emacs and Gforth
14421:
14422: @cindex @file{gforth.el}
14423: @cindex @file{forth.el}
14424: @cindex Rydqvist, Goran
1.107 dvdkhlng 14425: @cindex Kuehling, David
1.1 anton 14426: @cindex comment editing commands
14427: @cindex @code{\}, editing with Emacs
14428: @cindex debug tracer editing commands
14429: @cindex @code{~~}, removal with Emacs
14430: @cindex Forth mode in Emacs
1.107 dvdkhlng 14431:
1.1 anton 14432: Gforth comes with @file{gforth.el}, an improved version of
14433: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
1.26 crook 14434: improvements are:
14435:
14436: @itemize @bullet
14437: @item
1.107 dvdkhlng 14438: A better handling of indentation.
14439: @item
14440: A custom hilighting engine for Forth-code.
1.26 crook 14441: @item
14442: Comment paragraph filling (@kbd{M-q})
14443: @item
14444: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
14445: @item
14446: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
1.41 anton 14447: @item
14448: Support of the @code{info-lookup} feature for looking up the
14449: documentation of a word.
1.107 dvdkhlng 14450: @item
14451: Support for reading and writing blocks files.
1.26 crook 14452: @end itemize
14453:
1.107 dvdkhlng 14454: To get a basic description of these features, enter Forth mode and
14455: type @kbd{C-h m}.
1.1 anton 14456:
14457: @cindex source location of error or debugging output in Emacs
14458: @cindex error output, finding the source location in Emacs
14459: @cindex debugging output, finding the source location in Emacs
14460: In addition, Gforth supports Emacs quite well: The source code locations
14461: given in error messages, debugging output (from @code{~~}) and failed
14462: assertion messages are in the right format for Emacs' compilation mode
14463: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
14464: Manual}) so the source location corresponding to an error or other
14465: message is only a few keystrokes away (@kbd{C-x `} for the next error,
14466: @kbd{C-c C-c} for the error under the cursor).
14467:
1.107 dvdkhlng 14468: @cindex viewing the documentation of a word in Emacs
14469: @cindex context-sensitive help
14470: Moreover, for words documented in this manual, you can look up the
14471: glossary entry quickly by using @kbd{C-h TAB}
14472: (@code{info-lookup-symbol}, @pxref{Documentation, ,Documentation
14473: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
14474: later and does not work for words containing @code{:}.
14475:
14476: @menu
14477: * Installing gforth.el:: Making Emacs aware of Forth.
14478: * Emacs Tags:: Viewing the source of a word in Emacs.
14479: * Hilighting:: Making Forth code look prettier.
14480: * Auto-Indentation:: Customizing auto-indentation.
14481: * Blocks Files:: Reading and writing blocks files.
14482: @end menu
14483:
14484: @c ----------------------------------
1.109 anton 14485: @node Installing gforth.el, Emacs Tags, Emacs and Gforth, Emacs and Gforth
1.107 dvdkhlng 14486: @section Installing gforth.el
14487: @cindex @file{.emacs}
14488: @cindex @file{gforth.el}, installation
14489: To make the features from @file{gforth.el} available in Emacs, add
14490: the following lines to your @file{.emacs} file:
14491:
14492: @example
14493: (autoload 'forth-mode "gforth.el")
14494: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode)
14495: auto-mode-alist))
14496: (autoload 'forth-block-mode "gforth.el")
14497: (setq auto-mode-alist (cons '("\\.fb\\'" . forth-block-mode)
14498: auto-mode-alist))
14499: (add-hook 'forth-mode-hook (function (lambda ()
14500: ;; customize variables here:
14501: (setq forth-indent-level 4)
14502: (setq forth-minor-indent-level 2)
14503: (setq forth-hilight-level 3)
14504: ;;; ...
14505: )))
14506: @end example
14507:
14508: @c ----------------------------------
14509: @node Emacs Tags, Hilighting, Installing gforth.el, Emacs and Gforth
14510: @section Emacs Tags
1.1 anton 14511: @cindex @file{TAGS} file
14512: @cindex @file{etags.fs}
14513: @cindex viewing the source of a word in Emacs
1.43 anton 14514: @cindex @code{require}, placement in files
14515: @cindex @code{include}, placement in files
1.107 dvdkhlng 14516: If you @code{require} @file{etags.fs}, a new @file{TAGS} file will be
14517: produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
1.1 anton 14518: contains the definitions of all words defined afterwards. You can then
1.107 dvdkhlng 14519: find the source for a word using @kbd{M-.}. Note that Emacs can use
1.1 anton 14520: several tags files at the same time (e.g., one for the Gforth sources
14521: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
14522: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
14523: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
1.43 anton 14524: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
14525: with @file{etags.fs}, you should avoid putting definitions both before
14526: and after @code{require} etc., otherwise you will see the same file
14527: visited several times by commands like @code{tags-search}.
1.1 anton 14528:
1.107 dvdkhlng 14529: @c ----------------------------------
14530: @node Hilighting, Auto-Indentation, Emacs Tags, Emacs and Gforth
14531: @section Hilighting
14532: @cindex hilighting Forth code in Emacs
14533: @cindex highlighting Forth code in Emacs
14534: @file{gforth.el} comes with a custom source hilighting engine. When
14535: you open a file in @code{forth-mode}, it will be completely parsed,
14536: assigning faces to keywords, comments, strings etc. While you edit
14537: the file, modified regions get parsed and updated on-the-fly.
14538:
14539: Use the variable `forth-hilight-level' to change the level of
14540: decoration from 0 (no hilighting at all) to 3 (the default). Even if
14541: you set the hilighting level to 0, the parser will still work in the
14542: background, collecting information about whether regions of text are
14543: ``compiled'' or ``interpreted''. Those information are required for
14544: auto-indentation to work properly. Set `forth-disable-parser' to
14545: non-nil if your computer is too slow to handle parsing. This will
14546: have an impact on the smartness of the auto-indentation engine,
14547: though.
14548:
14549: Sometimes Forth sources define new features that should be hilighted,
14550: new control structures, defining-words etc. You can use the variable
14551: `forth-custom-words' to make @code{forth-mode} hilight additional
14552: words and constructs. See the docstring of `forth-words' for details
14553: (in Emacs, type @kbd{C-h v forth-words}).
14554:
14555: `forth-custom-words' is meant to be customized in your
14556: @file{.emacs} file. To customize hilighing in a file-specific manner,
14557: set `forth-local-words' in a local-variables section at the end of
14558: your source file (@pxref{Local Variables in Files,, Variables, emacs, Emacs Manual}).
14559:
14560: Example:
14561: @example
14562: 0 [IF]
14563: Local Variables:
14564: forth-local-words:
14565: ((("t:") definition-starter (font-lock-keyword-face . 1)
14566: "[ \t\n]" t name (font-lock-function-name-face . 3))
14567: ((";t") definition-ender (font-lock-keyword-face . 1)))
14568: End:
14569: [THEN]
14570: @end example
14571:
14572: @c ----------------------------------
14573: @node Auto-Indentation, Blocks Files, Hilighting, Emacs and Gforth
14574: @section Auto-Indentation
14575: @cindex auto-indentation of Forth code in Emacs
14576: @cindex indentation of Forth code in Emacs
14577: @code{forth-mode} automatically tries to indent lines in a smart way,
14578: whenever you type @key{TAB} or break a line with @kbd{C-m}.
14579:
14580: Simple customization can be achieved by setting
14581: `forth-indent-level' and `forth-minor-indent-level' in your
14582: @file{.emacs} file. For historical reasons @file{gforth.el} indents
14583: per default by multiples of 4 columns. To use the more traditional
14584: 3-column indentation, add the following lines to your @file{.emacs}:
14585:
14586: @example
14587: (add-hook 'forth-mode-hook (function (lambda ()
14588: ;; customize variables here:
14589: (setq forth-indent-level 3)
14590: (setq forth-minor-indent-level 1)
14591: )))
14592: @end example
14593:
14594: If you want indentation to recognize non-default words, customize it
14595: by setting `forth-custom-indent-words' in your @file{.emacs}. See the
14596: docstring of `forth-indent-words' for details (in Emacs, type @kbd{C-h
14597: v forth-indent-words}).
14598:
14599: To customize indentation in a file-specific manner, set
14600: `forth-local-indent-words' in a local-variables section at the end of
14601: your source file (@pxref{Local Variables in Files, Variables,,emacs,
14602: Emacs Manual}).
14603:
14604: Example:
14605: @example
14606: 0 [IF]
14607: Local Variables:
14608: forth-local-indent-words:
14609: ((("t:") (0 . 2) (0 . 2))
14610: ((";t") (-2 . 0) (0 . -2)))
14611: End:
14612: [THEN]
14613: @end example
14614:
14615: @c ----------------------------------
1.109 anton 14616: @node Blocks Files, , Auto-Indentation, Emacs and Gforth
1.107 dvdkhlng 14617: @section Blocks Files
14618: @cindex blocks files, use with Emacs
14619: @code{forth-mode} Autodetects blocks files by checking whether the
14620: length of the first line exceeds 1023 characters. It then tries to
14621: convert the file into normal text format. When you save the file, it
14622: will be written to disk as normal stream-source file.
14623:
14624: If you want to write blocks files, use @code{forth-blocks-mode}. It
14625: inherits all the features from @code{forth-mode}, plus some additions:
1.41 anton 14626:
1.107 dvdkhlng 14627: @itemize @bullet
14628: @item
14629: Files are written to disk in blocks file format.
14630: @item
14631: Screen numbers are displayed in the mode line (enumerated beginning
14632: with the value of `forth-block-base')
14633: @item
14634: Warnings are displayed when lines exceed 64 characters.
14635: @item
14636: The beginning of the currently edited block is marked with an
14637: overlay-arrow.
14638: @end itemize
1.41 anton 14639:
1.107 dvdkhlng 14640: There are some restrictions you should be aware of. When you open a
14641: blocks file that contains tabulator or newline characters, these
14642: characters will be translated into spaces when the file is written
14643: back to disk. If tabs or newlines are encountered during blocks file
14644: reading, an error is output to the echo area. So have a look at the
14645: `*Messages*' buffer, when Emacs' bell rings during reading.
1.1 anton 14646:
1.107 dvdkhlng 14647: Please consult the docstring of @code{forth-blocks-mode} for more
14648: information by typing @kbd{C-h v forth-blocks-mode}).
1.1 anton 14649:
1.26 crook 14650: @c ******************************************************************
1.1 anton 14651: @node Image Files, Engine, Emacs and Gforth, Top
14652: @chapter Image Files
1.26 crook 14653: @cindex image file
14654: @cindex @file{.fi} files
1.1 anton 14655: @cindex precompiled Forth code
14656: @cindex dictionary in persistent form
14657: @cindex persistent form of dictionary
14658:
14659: An image file is a file containing an image of the Forth dictionary,
14660: i.e., compiled Forth code and data residing in the dictionary. By
14661: convention, we use the extension @code{.fi} for image files.
14662:
14663: @menu
1.18 anton 14664: * Image Licensing Issues:: Distribution terms for images.
14665: * Image File Background:: Why have image files?
1.67 anton 14666: * Non-Relocatable Image Files:: don't always work.
1.18 anton 14667: * Data-Relocatable Image Files:: are better.
1.67 anton 14668: * Fully Relocatable Image Files:: better yet.
1.18 anton 14669: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
1.29 crook 14670: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
1.18 anton 14671: * Modifying the Startup Sequence:: and turnkey applications.
1.1 anton 14672: @end menu
14673:
1.18 anton 14674: @node Image Licensing Issues, Image File Background, Image Files, Image Files
14675: @section Image Licensing Issues
14676: @cindex license for images
14677: @cindex image license
14678:
14679: An image created with @code{gforthmi} (@pxref{gforthmi}) or
14680: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
14681: original image; i.e., according to copyright law it is a derived work of
14682: the original image.
14683:
14684: Since Gforth is distributed under the GNU GPL, the newly created image
14685: falls under the GNU GPL, too. In particular, this means that if you
14686: distribute the image, you have to make all of the sources for the image
1.113 anton 14687: available, including those you wrote. For details see @ref{Copying, ,
1.18 anton 14688: GNU General Public License (Section 3)}.
14689:
14690: If you create an image with @code{cross} (@pxref{cross.fs}), the image
14691: contains only code compiled from the sources you gave it; if none of
14692: these sources is under the GPL, the terms discussed above do not apply
14693: to the image. However, if your image needs an engine (a gforth binary)
14694: that is under the GPL, you should make sure that you distribute both in
14695: a way that is at most a @emph{mere aggregation}, if you don't want the
14696: terms of the GPL to apply to the image.
14697:
14698: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
1.1 anton 14699: @section Image File Background
14700: @cindex image file background
14701:
1.80 anton 14702: Gforth consists not only of primitives (in the engine), but also of
1.1 anton 14703: definitions written in Forth. Since the Forth compiler itself belongs to
14704: those definitions, it is not possible to start the system with the
1.80 anton 14705: engine and the Forth source alone. Therefore we provide the Forth
1.26 crook 14706: code as an image file in nearly executable form. When Gforth starts up,
14707: a C routine loads the image file into memory, optionally relocates the
14708: addresses, then sets up the memory (stacks etc.) according to
14709: information in the image file, and (finally) starts executing Forth
14710: code.
1.1 anton 14711:
14712: The image file variants represent different compromises between the
14713: goals of making it easy to generate image files and making them
14714: portable.
14715:
14716: @cindex relocation at run-time
1.26 crook 14717: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
1.1 anton 14718: run-time. This avoids many of the complications discussed below (image
14719: files are data relocatable without further ado), but costs performance
14720: (one addition per memory access).
14721:
14722: @cindex relocation at load-time
1.26 crook 14723: By contrast, the Gforth loader performs relocation at image load time. The
14724: loader also has to replace tokens that represent primitive calls with the
1.1 anton 14725: appropriate code-field addresses (or code addresses in the case of
14726: direct threading).
14727:
14728: There are three kinds of image files, with different degrees of
14729: relocatability: non-relocatable, data-relocatable, and fully relocatable
14730: image files.
14731:
14732: @cindex image file loader
14733: @cindex relocating loader
14734: @cindex loader for image files
14735: These image file variants have several restrictions in common; they are
14736: caused by the design of the image file loader:
14737:
14738: @itemize @bullet
14739: @item
14740: There is only one segment; in particular, this means, that an image file
14741: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
1.26 crook 14742: them). The contents of the stacks are not represented, either.
1.1 anton 14743:
14744: @item
14745: The only kinds of relocation supported are: adding the same offset to
14746: all cells that represent data addresses; and replacing special tokens
14747: with code addresses or with pieces of machine code.
14748:
14749: If any complex computations involving addresses are performed, the
14750: results cannot be represented in the image file. Several applications that
14751: use such computations come to mind:
14752: @itemize @minus
14753: @item
14754: Hashing addresses (or data structures which contain addresses) for table
14755: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
14756: purpose, you will have no problem, because the hash tables are
14757: recomputed automatically when the system is started. If you use your own
14758: hash tables, you will have to do something similar.
14759:
14760: @item
14761: There's a cute implementation of doubly-linked lists that uses
14762: @code{XOR}ed addresses. You could represent such lists as singly-linked
14763: in the image file, and restore the doubly-linked representation on
14764: startup.@footnote{In my opinion, though, you should think thrice before
14765: using a doubly-linked list (whatever implementation).}
14766:
14767: @item
14768: The code addresses of run-time routines like @code{docol:} cannot be
14769: represented in the image file (because their tokens would be replaced by
14770: machine code in direct threaded implementations). As a workaround,
14771: compute these addresses at run-time with @code{>code-address} from the
14772: executions tokens of appropriate words (see the definitions of
1.80 anton 14773: @code{docol:} and friends in @file{kernel/getdoers.fs}).
1.1 anton 14774:
14775: @item
14776: On many architectures addresses are represented in machine code in some
14777: shifted or mangled form. You cannot put @code{CODE} words that contain
14778: absolute addresses in this form in a relocatable image file. Workarounds
14779: are representing the address in some relative form (e.g., relative to
14780: the CFA, which is present in some register), or loading the address from
14781: a place where it is stored in a non-mangled form.
14782: @end itemize
14783: @end itemize
14784:
14785: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
14786: @section Non-Relocatable Image Files
14787: @cindex non-relocatable image files
1.26 crook 14788: @cindex image file, non-relocatable
1.1 anton 14789:
14790: These files are simple memory dumps of the dictionary. They are specific
14791: to the executable (i.e., @file{gforth} file) they were created
14792: with. What's worse, they are specific to the place on which the
14793: dictionary resided when the image was created. Now, there is no
14794: guarantee that the dictionary will reside at the same place the next
14795: time you start Gforth, so there's no guarantee that a non-relocatable
14796: image will work the next time (Gforth will complain instead of crashing,
14797: though).
14798:
14799: You can create a non-relocatable image file with
14800:
1.44 crook 14801:
1.1 anton 14802: doc-savesystem
14803:
1.44 crook 14804:
1.1 anton 14805: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
14806: @section Data-Relocatable Image Files
14807: @cindex data-relocatable image files
1.26 crook 14808: @cindex image file, data-relocatable
1.1 anton 14809:
14810: These files contain relocatable data addresses, but fixed code addresses
14811: (instead of tokens). They are specific to the executable (i.e.,
14812: @file{gforth} file) they were created with. For direct threading on some
14813: architectures (e.g., the i386), data-relocatable images do not work. You
14814: get a data-relocatable image, if you use @file{gforthmi} with a
14815: Gforth binary that is not doubly indirect threaded (@pxref{Fully
14816: Relocatable Image Files}).
14817:
14818: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
14819: @section Fully Relocatable Image Files
14820: @cindex fully relocatable image files
1.26 crook 14821: @cindex image file, fully relocatable
1.1 anton 14822:
14823: @cindex @file{kern*.fi}, relocatability
14824: @cindex @file{gforth.fi}, relocatability
14825: These image files have relocatable data addresses, and tokens for code
14826: addresses. They can be used with different binaries (e.g., with and
14827: without debugging) on the same machine, and even across machines with
14828: the same data formats (byte order, cell size, floating point
14829: format). However, they are usually specific to the version of Gforth
14830: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
14831: are fully relocatable.
14832:
14833: There are two ways to create a fully relocatable image file:
14834:
14835: @menu
1.29 crook 14836: * gforthmi:: The normal way
1.1 anton 14837: * cross.fs:: The hard way
14838: @end menu
14839:
14840: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
14841: @subsection @file{gforthmi}
14842: @cindex @file{comp-i.fs}
14843: @cindex @file{gforthmi}
14844:
14845: You will usually use @file{gforthmi}. If you want to create an
1.29 crook 14846: image @i{file} that contains everything you would load by invoking
14847: Gforth with @code{gforth @i{options}}, you simply say:
1.1 anton 14848: @example
1.29 crook 14849: gforthmi @i{file} @i{options}
1.1 anton 14850: @end example
14851:
14852: E.g., if you want to create an image @file{asm.fi} that has the file
14853: @file{asm.fs} loaded in addition to the usual stuff, you could do it
14854: like this:
14855:
14856: @example
14857: gforthmi asm.fi asm.fs
14858: @end example
14859:
1.27 crook 14860: @file{gforthmi} is implemented as a sh script and works like this: It
14861: produces two non-relocatable images for different addresses and then
14862: compares them. Its output reflects this: first you see the output (if
1.62 crook 14863: any) of the two Gforth invocations that produce the non-relocatable image
1.27 crook 14864: files, then you see the output of the comparing program: It displays the
14865: offset used for data addresses and the offset used for code addresses;
1.1 anton 14866: moreover, for each cell that cannot be represented correctly in the
1.44 crook 14867: image files, it displays a line like this:
1.1 anton 14868:
14869: @example
14870: 78DC BFFFFA50 BFFFFA40
14871: @end example
14872:
14873: This means that at offset $78dc from @code{forthstart}, one input image
14874: contains $bffffa50, and the other contains $bffffa40. Since these cells
14875: cannot be represented correctly in the output image, you should examine
14876: these places in the dictionary and verify that these cells are dead
14877: (i.e., not read before they are written).
1.39 anton 14878:
14879: @cindex --application, @code{gforthmi} option
14880: If you insert the option @code{--application} in front of the image file
14881: name, you will get an image that uses the @code{--appl-image} option
14882: instead of the @code{--image-file} option (@pxref{Invoking
14883: Gforth}). When you execute such an image on Unix (by typing the image
14884: name as command), the Gforth engine will pass all options to the image
14885: instead of trying to interpret them as engine options.
1.1 anton 14886:
1.27 crook 14887: If you type @file{gforthmi} with no arguments, it prints some usage
14888: instructions.
14889:
1.1 anton 14890: @cindex @code{savesystem} during @file{gforthmi}
14891: @cindex @code{bye} during @file{gforthmi}
14892: @cindex doubly indirect threaded code
1.44 crook 14893: @cindex environment variables
14894: @cindex @code{GFORTHD} -- environment variable
14895: @cindex @code{GFORTH} -- environment variable
1.1 anton 14896: @cindex @code{gforth-ditc}
1.29 crook 14897: There are a few wrinkles: After processing the passed @i{options}, the
1.1 anton 14898: words @code{savesystem} and @code{bye} must be visible. A special doubly
14899: indirect threaded version of the @file{gforth} executable is used for
1.62 crook 14900: creating the non-relocatable images; you can pass the exact filename of
1.1 anton 14901: this executable through the environment variable @code{GFORTHD}
14902: (default: @file{gforth-ditc}); if you pass a version that is not doubly
14903: indirect threaded, you will not get a fully relocatable image, but a
1.27 crook 14904: data-relocatable image (because there is no code address offset). The
14905: normal @file{gforth} executable is used for creating the relocatable
14906: image; you can pass the exact filename of this executable through the
14907: environment variable @code{GFORTH}.
1.1 anton 14908:
14909: @node cross.fs, , gforthmi, Fully Relocatable Image Files
14910: @subsection @file{cross.fs}
14911: @cindex @file{cross.fs}
14912: @cindex cross-compiler
14913: @cindex metacompiler
1.47 crook 14914: @cindex target compiler
1.1 anton 14915:
14916: You can also use @code{cross}, a batch compiler that accepts a Forth-like
1.47 crook 14917: programming language (@pxref{Cross Compiler}).
1.1 anton 14918:
1.47 crook 14919: @code{cross} allows you to create image files for machines with
1.1 anton 14920: different data sizes and data formats than the one used for generating
14921: the image file. You can also use it to create an application image that
14922: does not contain a Forth compiler. These features are bought with
14923: restrictions and inconveniences in programming. E.g., addresses have to
14924: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
14925: order to make the code relocatable.
14926:
14927:
14928: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
14929: @section Stack and Dictionary Sizes
14930: @cindex image file, stack and dictionary sizes
14931: @cindex dictionary size default
14932: @cindex stack size default
14933:
14934: If you invoke Gforth with a command line flag for the size
14935: (@pxref{Invoking Gforth}), the size you specify is stored in the
14936: dictionary. If you save the dictionary with @code{savesystem} or create
14937: an image with @file{gforthmi}, this size will become the default
14938: for the resulting image file. E.g., the following will create a
1.21 crook 14939: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
1.1 anton 14940:
14941: @example
14942: gforthmi gforth.fi -m 1M
14943: @end example
14944:
14945: In other words, if you want to set the default size for the dictionary
14946: and the stacks of an image, just invoke @file{gforthmi} with the
14947: appropriate options when creating the image.
14948:
14949: @cindex stack size, cache-friendly
14950: Note: For cache-friendly behaviour (i.e., good performance), you should
14951: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
14952: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
14953: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
14954:
14955: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
14956: @section Running Image Files
14957: @cindex running image files
14958: @cindex invoking image files
14959: @cindex image file invocation
14960:
14961: @cindex -i, invoke image file
14962: @cindex --image file, invoke image file
1.29 crook 14963: You can invoke Gforth with an image file @i{image} instead of the
1.1 anton 14964: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
14965: @example
1.29 crook 14966: gforth -i @i{image}
1.1 anton 14967: @end example
14968:
14969: @cindex executable image file
1.26 crook 14970: @cindex image file, executable
1.1 anton 14971: If your operating system supports starting scripts with a line of the
14972: form @code{#! ...}, you just have to type the image file name to start
14973: Gforth with this image file (note that the file extension @code{.fi} is
1.29 crook 14974: just a convention). I.e., to run Gforth with the image file @i{image},
14975: you can just type @i{image} instead of @code{gforth -i @i{image}}.
1.27 crook 14976: This works because every @code{.fi} file starts with a line of this
14977: format:
14978:
14979: @example
14980: #! /usr/local/bin/gforth-0.4.0 -i
14981: @end example
14982:
14983: The file and pathname for the Gforth engine specified on this line is
14984: the specific Gforth executable that it was built against; i.e. the value
14985: of the environment variable @code{GFORTH} at the time that
14986: @file{gforthmi} was executed.
1.1 anton 14987:
1.27 crook 14988: You can make use of the same shell capability to make a Forth source
14989: file into an executable. For example, if you place this text in a file:
1.26 crook 14990:
14991: @example
14992: #! /usr/local/bin/gforth
14993:
14994: ." Hello, world" CR
14995: bye
14996: @end example
14997:
14998: @noindent
1.27 crook 14999: and then make the file executable (chmod +x in Unix), you can run it
1.26 crook 15000: directly from the command line. The sequence @code{#!} is used in two
15001: ways; firstly, it is recognised as a ``magic sequence'' by the operating
1.29 crook 15002: system@footnote{The Unix kernel actually recognises two types of files:
15003: executable files and files of data, where the data is processed by an
15004: interpreter that is specified on the ``interpreter line'' -- the first
15005: line of the file, starting with the sequence #!. There may be a small
15006: limit (e.g., 32) on the number of characters that may be specified on
15007: the interpreter line.} secondly it is treated as a comment character by
15008: Gforth. Because of the second usage, a space is required between
1.80 anton 15009: @code{#!} and the path to the executable (moreover, some Unixes
15010: require the sequence @code{#! /}).
1.27 crook 15011:
15012: The disadvantage of this latter technique, compared with using
1.80 anton 15013: @file{gforthmi}, is that it is slightly slower; the Forth source code is
15014: compiled on-the-fly, each time the program is invoked.
1.26 crook 15015:
1.1 anton 15016: doc-#!
15017:
1.44 crook 15018:
1.1 anton 15019: @node Modifying the Startup Sequence, , Running Image Files, Image Files
15020: @section Modifying the Startup Sequence
15021: @cindex startup sequence for image file
15022: @cindex image file initialization sequence
15023: @cindex initialization sequence of image file
15024:
1.120 anton 15025: You can add your own initialization to the startup sequence of an image
15026: through the deferred word @code{'cold}. @code{'cold} is invoked just
15027: before the image-specific command line processing (i.e., loading files
15028: and evaluating (@code{-e}) strings) starts.
1.1 anton 15029:
15030: A sequence for adding your initialization usually looks like this:
15031:
15032: @example
15033: :noname
15034: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
15035: ... \ your stuff
15036: ; IS 'cold
15037: @end example
15038:
1.157 anton 15039: After @code{'cold}, Gforth processes the image options
15040: (@pxref{Invoking Gforth}), and then it performs @code{bootmessage},
15041: another deferred word. This normally prints Gforth's startup message
15042: and does nothing else.
15043:
1.1 anton 15044: @cindex turnkey image files
1.26 crook 15045: @cindex image file, turnkey applications
1.157 anton 15046: So, if you want to make a turnkey image (i.e., an image for an
15047: application instead of an extended Forth system), you can do this in
15048: two ways:
15049:
15050: @itemize @bullet
15051:
15052: @item
15053: If you want to do your interpretation of the OS command-line
15054: arguments, hook into @code{'cold}. In that case you probably also
15055: want to build the image with @code{gforthmi --application}
15056: (@pxref{gforthmi}) to keep the engine from processing OS command line
15057: options. You can then do your own command-line processing with
15058: @code{next-arg}
15059:
15060: @item
15061: If you want to have the normal Gforth processing of OS command-line
15062: arguments, hook into @code{bootmessage}.
15063:
15064: @end itemize
15065:
15066: In either case, you probably do not want the word that you execute in
15067: these hooks to exit normally, but use @code{bye} or @code{throw}.
15068: Otherwise the Gforth startup process would continue and eventually
15069: present the Forth command line to the user.
1.26 crook 15070:
15071: doc-'cold
1.157 anton 15072: doc-bootmessage
1.44 crook 15073:
1.1 anton 15074: @c ******************************************************************
1.113 anton 15075: @node Engine, Cross Compiler, Image Files, Top
1.1 anton 15076: @chapter Engine
15077: @cindex engine
15078: @cindex virtual machine
15079:
1.26 crook 15080: Reading this chapter is not necessary for programming with Gforth. It
1.1 anton 15081: may be helpful for finding your way in the Gforth sources.
15082:
1.109 anton 15083: The ideas in this section have also been published in the following
15084: papers: Bernd Paysan, @cite{ANS fig/GNU/??? Forth} (in German),
15085: Forth-Tagung '93; M. Anton Ertl,
15086: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
15087: Portable Forth Engine}}, EuroForth '93; M. Anton Ertl,
15088: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl02.ps.gz,
15089: Threaded code variations and optimizations (extended version)}},
15090: Forth-Tagung '02.
1.1 anton 15091:
15092: @menu
15093: * Portability::
15094: * Threading::
15095: * Primitives::
15096: * Performance::
15097: @end menu
15098:
15099: @node Portability, Threading, Engine, Engine
15100: @section Portability
15101: @cindex engine portability
15102:
1.26 crook 15103: An important goal of the Gforth Project is availability across a wide
15104: range of personal machines. fig-Forth, and, to a lesser extent, F83,
15105: achieved this goal by manually coding the engine in assembly language
15106: for several then-popular processors. This approach is very
15107: labor-intensive and the results are short-lived due to progress in
15108: computer architecture.
1.1 anton 15109:
15110: @cindex C, using C for the engine
15111: Others have avoided this problem by coding in C, e.g., Mitch Bradley
15112: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
15113: particularly popular for UNIX-based Forths due to the large variety of
15114: architectures of UNIX machines. Unfortunately an implementation in C
15115: does not mix well with the goals of efficiency and with using
15116: traditional techniques: Indirect or direct threading cannot be expressed
15117: in C, and switch threading, the fastest technique available in C, is
15118: significantly slower. Another problem with C is that it is very
15119: cumbersome to express double integer arithmetic.
15120:
15121: @cindex GNU C for the engine
15122: @cindex long long
15123: Fortunately, there is a portable language that does not have these
15124: limitations: GNU C, the version of C processed by the GNU C compiler
15125: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
15126: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
15127: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
15128: threading possible, its @code{long long} type (@pxref{Long Long, ,
15129: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
1.109 anton 15130: double numbers on many systems. GNU C is freely available on all
1.1 anton 15131: important (and many unimportant) UNIX machines, VMS, 80386s running
15132: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
15133: on all these machines.
15134:
15135: Writing in a portable language has the reputation of producing code that
15136: is slower than assembly. For our Forth engine we repeatedly looked at
15137: the code produced by the compiler and eliminated most compiler-induced
15138: inefficiencies by appropriate changes in the source code.
15139:
15140: @cindex explicit register declarations
15141: @cindex --enable-force-reg, configuration flag
15142: @cindex -DFORCE_REG
15143: However, register allocation cannot be portably influenced by the
15144: programmer, leading to some inefficiencies on register-starved
15145: machines. We use explicit register declarations (@pxref{Explicit Reg
15146: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
15147: improve the speed on some machines. They are turned on by using the
15148: configuration flag @code{--enable-force-reg} (@code{gcc} switch
15149: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
15150: machine, but also on the compiler version: On some machines some
15151: compiler versions produce incorrect code when certain explicit register
15152: declarations are used. So by default @code{-DFORCE_REG} is not used.
15153:
15154: @node Threading, Primitives, Portability, Engine
15155: @section Threading
15156: @cindex inner interpreter implementation
15157: @cindex threaded code implementation
15158:
15159: @cindex labels as values
15160: GNU C's labels as values extension (available since @code{gcc-2.0},
15161: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
1.29 crook 15162: makes it possible to take the address of @i{label} by writing
15163: @code{&&@i{label}}. This address can then be used in a statement like
15164: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
1.1 anton 15165: @code{goto x}.
15166:
1.26 crook 15167: @cindex @code{NEXT}, indirect threaded
1.1 anton 15168: @cindex indirect threaded inner interpreter
15169: @cindex inner interpreter, indirect threaded
1.26 crook 15170: With this feature an indirect threaded @code{NEXT} looks like:
1.1 anton 15171: @example
15172: cfa = *ip++;
15173: ca = *cfa;
15174: goto *ca;
15175: @end example
15176: @cindex instruction pointer
15177: For those unfamiliar with the names: @code{ip} is the Forth instruction
15178: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
15179: execution token and points to the code field of the next word to be
15180: executed; The @code{ca} (code address) fetched from there points to some
15181: executable code, e.g., a primitive or the colon definition handler
15182: @code{docol}.
15183:
1.26 crook 15184: @cindex @code{NEXT}, direct threaded
1.1 anton 15185: @cindex direct threaded inner interpreter
15186: @cindex inner interpreter, direct threaded
15187: Direct threading is even simpler:
15188: @example
15189: ca = *ip++;
15190: goto *ca;
15191: @end example
15192:
15193: Of course we have packaged the whole thing neatly in macros called
1.26 crook 15194: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
1.1 anton 15195:
15196: @menu
15197: * Scheduling::
15198: * Direct or Indirect Threaded?::
1.109 anton 15199: * Dynamic Superinstructions::
1.1 anton 15200: * DOES>::
15201: @end menu
15202:
15203: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
15204: @subsection Scheduling
15205: @cindex inner interpreter optimization
15206:
15207: There is a little complication: Pipelined and superscalar processors,
15208: i.e., RISC and some modern CISC machines can process independent
15209: instructions while waiting for the results of an instruction. The
15210: compiler usually reorders (schedules) the instructions in a way that
15211: achieves good usage of these delay slots. However, on our first tries
15212: the compiler did not do well on scheduling primitives. E.g., for
15213: @code{+} implemented as
15214: @example
15215: n=sp[0]+sp[1];
15216: sp++;
15217: sp[0]=n;
15218: NEXT;
15219: @end example
1.81 anton 15220: the @code{NEXT} comes strictly after the other code, i.e., there is
15221: nearly no scheduling. After a little thought the problem becomes clear:
15222: The compiler cannot know that @code{sp} and @code{ip} point to different
1.21 crook 15223: addresses (and the version of @code{gcc} we used would not know it even
15224: if it was possible), so it could not move the load of the cfa above the
15225: store to the TOS. Indeed the pointers could be the same, if code on or
15226: very near the top of stack were executed. In the interest of speed we
15227: chose to forbid this probably unused ``feature'' and helped the compiler
1.81 anton 15228: in scheduling: @code{NEXT} is divided into several parts:
15229: @code{NEXT_P0}, @code{NEXT_P1} and @code{NEXT_P2}). @code{+} now looks
15230: like:
1.1 anton 15231: @example
1.81 anton 15232: NEXT_P0;
1.1 anton 15233: n=sp[0]+sp[1];
15234: sp++;
15235: NEXT_P1;
15236: sp[0]=n;
15237: NEXT_P2;
15238: @end example
15239:
1.81 anton 15240: There are various schemes that distribute the different operations of
15241: NEXT between these parts in several ways; in general, different schemes
15242: perform best on different processors. We use a scheme for most
15243: architectures that performs well for most processors of this
1.109 anton 15244: architecture; in the future we may switch to benchmarking and chosing
1.81 anton 15245: the scheme on installation time.
15246:
1.1 anton 15247:
1.109 anton 15248: @node Direct or Indirect Threaded?, Dynamic Superinstructions, Scheduling, Threading
1.1 anton 15249: @subsection Direct or Indirect Threaded?
15250: @cindex threading, direct or indirect?
15251:
1.109 anton 15252: Threaded forth code consists of references to primitives (simple machine
15253: code routines like @code{+}) and to non-primitives (e.g., colon
15254: definitions, variables, constants); for a specific class of
15255: non-primitives (e.g., variables) there is one code routine (e.g.,
15256: @code{dovar}), but each variable needs a separate reference to its data.
15257:
15258: Traditionally Forth has been implemented as indirect threaded code,
15259: because this allows to use only one cell to reference a non-primitive
15260: (basically you point to the data, and find the code address there).
15261:
15262: @cindex primitive-centric threaded code
15263: However, threaded code in Gforth (since 0.6.0) uses two cells for
15264: non-primitives, one for the code address, and one for the data address;
15265: the data pointer is an immediate argument for the virtual machine
15266: instruction represented by the code address. We call this
15267: @emph{primitive-centric} threaded code, because all code addresses point
15268: to simple primitives. E.g., for a variable, the code address is for
15269: @code{lit} (also used for integer literals like @code{99}).
15270:
15271: Primitive-centric threaded code allows us to use (faster) direct
15272: threading as dispatch method, completely portably (direct threaded code
15273: in Gforth before 0.6.0 required architecture-specific code). It also
15274: eliminates the performance problems related to I-cache consistency that
15275: 386 implementations have with direct threaded code, and allows
15276: additional optimizations.
15277:
15278: @cindex hybrid direct/indirect threaded code
15279: There is a catch, however: the @var{xt} parameter of @code{execute} can
15280: occupy only one cell, so how do we pass non-primitives with their code
15281: @emph{and} data addresses to them? Our answer is to use indirect
15282: threaded dispatch for @code{execute} and other words that use a
15283: single-cell xt. So, normal threaded code in colon definitions uses
15284: direct threading, and @code{execute} and similar words, which dispatch
15285: to xts on the data stack, use indirect threaded code. We call this
15286: @emph{hybrid direct/indirect} threaded code.
15287:
15288: @cindex engines, gforth vs. gforth-fast vs. gforth-itc
15289: @cindex gforth engine
15290: @cindex gforth-fast engine
15291: The engines @command{gforth} and @command{gforth-fast} use hybrid
15292: direct/indirect threaded code. This means that with these engines you
15293: cannot use @code{,} to compile an xt. Instead, you have to use
15294: @code{compile,}.
15295:
15296: @cindex gforth-itc engine
1.115 anton 15297: If you want to compile xts with @code{,}, use @command{gforth-itc}.
15298: This engine uses plain old indirect threaded code. It still compiles in
15299: a primitive-centric style, so you cannot use @code{compile,} instead of
1.109 anton 15300: @code{,} (e.g., for producing tables of xts with @code{] word1 word2
1.115 anton 15301: ... [}). If you want to do that, you have to use @command{gforth-itc}
1.109 anton 15302: and execute @code{' , is compile,}. Your program can check if it is
15303: running on a hybrid direct/indirect threaded engine or a pure indirect
15304: threaded engine with @code{threading-method} (@pxref{Threading Words}).
15305:
15306:
15307: @node Dynamic Superinstructions, DOES>, Direct or Indirect Threaded?, Threading
15308: @subsection Dynamic Superinstructions
15309: @cindex Dynamic superinstructions with replication
15310: @cindex Superinstructions
15311: @cindex Replication
15312:
15313: The engines @command{gforth} and @command{gforth-fast} use another
15314: optimization: Dynamic superinstructions with replication. As an
15315: example, consider the following colon definition:
15316:
15317: @example
15318: : squared ( n1 -- n2 )
15319: dup * ;
15320: @end example
15321:
15322: Gforth compiles this into the threaded code sequence
15323:
15324: @example
15325: dup
15326: *
15327: ;s
15328: @end example
15329:
15330: In normal direct threaded code there is a code address occupying one
15331: cell for each of these primitives. Each code address points to a
15332: machine code routine, and the interpreter jumps to this machine code in
15333: order to execute the primitive. The routines for these three
15334: primitives are (in @command{gforth-fast} on the 386):
15335:
15336: @example
15337: Code dup
15338: ( $804B950 ) add esi , # -4 \ $83 $C6 $FC
15339: ( $804B953 ) add ebx , # 4 \ $83 $C3 $4
15340: ( $804B956 ) mov dword ptr 4 [esi] , ecx \ $89 $4E $4
15341: ( $804B959 ) jmp dword ptr FC [ebx] \ $FF $63 $FC
15342: end-code
15343: Code *
15344: ( $804ACC4 ) mov eax , dword ptr 4 [esi] \ $8B $46 $4
15345: ( $804ACC7 ) add esi , # 4 \ $83 $C6 $4
15346: ( $804ACCA ) add ebx , # 4 \ $83 $C3 $4
15347: ( $804ACCD ) imul ecx , eax \ $F $AF $C8
15348: ( $804ACD0 ) jmp dword ptr FC [ebx] \ $FF $63 $FC
15349: end-code
15350: Code ;s
15351: ( $804A693 ) mov eax , dword ptr [edi] \ $8B $7
15352: ( $804A695 ) add edi , # 4 \ $83 $C7 $4
15353: ( $804A698 ) lea ebx , dword ptr 4 [eax] \ $8D $58 $4
15354: ( $804A69B ) jmp dword ptr FC [ebx] \ $FF $63 $FC
15355: end-code
15356: @end example
15357:
15358: With dynamic superinstructions and replication the compiler does not
15359: just lay down the threaded code, but also copies the machine code
15360: fragments, usually without the jump at the end.
15361:
15362: @example
15363: ( $4057D27D ) add esi , # -4 \ $83 $C6 $FC
15364: ( $4057D280 ) add ebx , # 4 \ $83 $C3 $4
15365: ( $4057D283 ) mov dword ptr 4 [esi] , ecx \ $89 $4E $4
15366: ( $4057D286 ) mov eax , dword ptr 4 [esi] \ $8B $46 $4
15367: ( $4057D289 ) add esi , # 4 \ $83 $C6 $4
15368: ( $4057D28C ) add ebx , # 4 \ $83 $C3 $4
15369: ( $4057D28F ) imul ecx , eax \ $F $AF $C8
15370: ( $4057D292 ) mov eax , dword ptr [edi] \ $8B $7
15371: ( $4057D294 ) add edi , # 4 \ $83 $C7 $4
15372: ( $4057D297 ) lea ebx , dword ptr 4 [eax] \ $8D $58 $4
15373: ( $4057D29A ) jmp dword ptr FC [ebx] \ $FF $63 $FC
15374: @end example
15375:
15376: Only when a threaded-code control-flow change happens (e.g., in
15377: @code{;s}), the jump is appended. This optimization eliminates many of
15378: these jumps and makes the rest much more predictable. The speedup
15379: depends on the processor and the application; on the Athlon and Pentium
15380: III this optimization typically produces a speedup by a factor of 2.
15381:
15382: The code addresses in the direct-threaded code are set to point to the
15383: appropriate points in the copied machine code, in this example like
15384: this:
1.1 anton 15385:
1.109 anton 15386: @example
15387: primitive code address
15388: dup $4057D27D
15389: * $4057D286
15390: ;s $4057D292
15391: @end example
15392:
15393: Thus there can be threaded-code jumps to any place in this piece of
15394: code. This also simplifies decompilation quite a bit.
15395:
15396: @cindex --no-dynamic command-line option
15397: @cindex --no-super command-line option
15398: You can disable this optimization with @option{--no-dynamic}. You can
15399: use the copying without eliminating the jumps (i.e., dynamic
15400: replication, but without superinstructions) with @option{--no-super};
15401: this gives the branch prediction benefit alone; the effect on
1.110 anton 15402: performance depends on the CPU; on the Athlon and Pentium III the
15403: speedup is a little less than for dynamic superinstructions with
15404: replication.
15405:
15406: @cindex patching threaded code
15407: One use of these options is if you want to patch the threaded code.
15408: With superinstructions, many of the dispatch jumps are eliminated, so
15409: patching often has no effect. These options preserve all the dispatch
15410: jumps.
1.109 anton 15411:
15412: @cindex --dynamic command-line option
1.110 anton 15413: On some machines dynamic superinstructions are disabled by default,
15414: because it is unsafe on these machines. However, if you feel
15415: adventurous, you can enable it with @option{--dynamic}.
1.109 anton 15416:
15417: @node DOES>, , Dynamic Superinstructions, Threading
1.1 anton 15418: @subsection DOES>
15419: @cindex @code{DOES>} implementation
15420:
1.26 crook 15421: @cindex @code{dodoes} routine
15422: @cindex @code{DOES>}-code
1.1 anton 15423: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
15424: the chunk of code executed by every word defined by a
1.109 anton 15425: @code{CREATE}...@code{DOES>} pair; actually with primitive-centric code,
15426: this is only needed if the xt of the word is @code{execute}d. The main
15427: problem here is: How to find the Forth code to be executed, i.e. the
15428: code after the @code{DOES>} (the @code{DOES>}-code)? There are two
15429: solutions:
1.1 anton 15430:
1.21 crook 15431: In fig-Forth the code field points directly to the @code{dodoes} and the
1.109 anton 15432: @code{DOES>}-code address is stored in the cell after the code address
15433: (i.e. at @code{@i{CFA} cell+}). It may seem that this solution is
15434: illegal in the Forth-79 and all later standards, because in fig-Forth
15435: this address lies in the body (which is illegal in these
15436: standards). However, by making the code field larger for all words this
15437: solution becomes legal again. We use this approach. Leaving a cell
15438: unused in most words is a bit wasteful, but on the machines we are
15439: targeting this is hardly a problem.
15440:
1.1 anton 15441:
15442: @node Primitives, Performance, Threading, Engine
15443: @section Primitives
15444: @cindex primitives, implementation
15445: @cindex virtual machine instructions, implementation
15446:
15447: @menu
15448: * Automatic Generation::
15449: * TOS Optimization::
15450: * Produced code::
15451: @end menu
15452:
15453: @node Automatic Generation, TOS Optimization, Primitives, Primitives
15454: @subsection Automatic Generation
15455: @cindex primitives, automatic generation
15456:
15457: @cindex @file{prims2x.fs}
1.109 anton 15458:
1.1 anton 15459: Since the primitives are implemented in a portable language, there is no
15460: longer any need to minimize the number of primitives. On the contrary,
15461: having many primitives has an advantage: speed. In order to reduce the
15462: number of errors in primitives and to make programming them easier, we
1.109 anton 15463: provide a tool, the primitive generator (@file{prims2x.fs} aka Vmgen,
15464: @pxref{Top, Vmgen, Introduction, vmgen, Vmgen}), that automatically
15465: generates most (and sometimes all) of the C code for a primitive from
15466: the stack effect notation. The source for a primitive has the following
15467: form:
1.1 anton 15468:
15469: @cindex primitive source format
15470: @format
1.58 anton 15471: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
1.29 crook 15472: [@code{""}@i{glossary entry}@code{""}]
15473: @i{C code}
1.1 anton 15474: [@code{:}
1.29 crook 15475: @i{Forth code}]
1.1 anton 15476: @end format
15477:
15478: The items in brackets are optional. The category and glossary fields
15479: are there for generating the documentation, the Forth code is there
15480: for manual implementations on machines without GNU C. E.g., the source
15481: for the primitive @code{+} is:
15482: @example
1.58 anton 15483: + ( n1 n2 -- n ) core plus
1.1 anton 15484: n = n1+n2;
15485: @end example
15486:
15487: This looks like a specification, but in fact @code{n = n1+n2} is C
15488: code. Our primitive generation tool extracts a lot of information from
15489: the stack effect notations@footnote{We use a one-stack notation, even
15490: though we have separate data and floating-point stacks; The separate
15491: notation can be generated easily from the unified notation.}: The number
15492: of items popped from and pushed on the stack, their type, and by what
15493: name they are referred to in the C code. It then generates a C code
15494: prelude and postlude for each primitive. The final C code for @code{+}
15495: looks like this:
15496:
15497: @example
1.46 pazsan 15498: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
1.1 anton 15499: /* */ /* documentation */
1.81 anton 15500: NAME("+") /* debugging output (with -DDEBUG) */
1.1 anton 15501: @{
15502: DEF_CA /* definition of variable ca (indirect threading) */
15503: Cell n1; /* definitions of variables */
15504: Cell n2;
15505: Cell n;
1.81 anton 15506: NEXT_P0; /* NEXT part 0 */
1.1 anton 15507: n1 = (Cell) sp[1]; /* input */
15508: n2 = (Cell) TOS;
15509: sp += 1; /* stack adjustment */
15510: @{
15511: n = n1+n2; /* C code taken from the source */
15512: @}
15513: NEXT_P1; /* NEXT part 1 */
15514: TOS = (Cell)n; /* output */
15515: NEXT_P2; /* NEXT part 2 */
15516: @}
15517: @end example
15518:
15519: This looks long and inefficient, but the GNU C compiler optimizes quite
15520: well and produces optimal code for @code{+} on, e.g., the R3000 and the
15521: HP RISC machines: Defining the @code{n}s does not produce any code, and
15522: using them as intermediate storage also adds no cost.
15523:
1.26 crook 15524: There are also other optimizations that are not illustrated by this
15525: example: assignments between simple variables are usually for free (copy
1.1 anton 15526: propagation). If one of the stack items is not used by the primitive
15527: (e.g. in @code{drop}), the compiler eliminates the load from the stack
15528: (dead code elimination). On the other hand, there are some things that
15529: the compiler does not do, therefore they are performed by
15530: @file{prims2x.fs}: The compiler does not optimize code away that stores
15531: a stack item to the place where it just came from (e.g., @code{over}).
15532:
15533: While programming a primitive is usually easy, there are a few cases
15534: where the programmer has to take the actions of the generator into
15535: account, most notably @code{?dup}, but also words that do not (always)
1.26 crook 15536: fall through to @code{NEXT}.
1.109 anton 15537:
15538: For more information
1.1 anton 15539:
15540: @node TOS Optimization, Produced code, Automatic Generation, Primitives
15541: @subsection TOS Optimization
15542: @cindex TOS optimization for primitives
15543: @cindex primitives, keeping the TOS in a register
15544:
15545: An important optimization for stack machine emulators, e.g., Forth
15546: engines, is keeping one or more of the top stack items in
1.29 crook 15547: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
15548: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
1.1 anton 15549: @itemize @bullet
15550: @item
1.29 crook 15551: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
1.1 anton 15552: due to fewer loads from and stores to the stack.
1.29 crook 15553: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
15554: @i{y<n}, due to additional moves between registers.
1.1 anton 15555: @end itemize
15556:
15557: @cindex -DUSE_TOS
15558: @cindex -DUSE_NO_TOS
15559: In particular, keeping one item in a register is never a disadvantage,
15560: if there are enough registers. Keeping two items in registers is a
15561: disadvantage for frequent words like @code{?branch}, constants,
15562: variables, literals and @code{i}. Therefore our generator only produces
15563: code that keeps zero or one items in registers. The generated C code
15564: covers both cases; the selection between these alternatives is made at
15565: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
15566: code for @code{+} is just a simple variable name in the one-item case,
15567: otherwise it is a macro that expands into @code{sp[0]}. Note that the
15568: GNU C compiler tries to keep simple variables like @code{TOS} in
15569: registers, and it usually succeeds, if there are enough registers.
15570:
15571: @cindex -DUSE_FTOS
15572: @cindex -DUSE_NO_FTOS
15573: The primitive generator performs the TOS optimization for the
15574: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
15575: operations the benefit of this optimization is even larger:
15576: floating-point operations take quite long on most processors, but can be
15577: performed in parallel with other operations as long as their results are
15578: not used. If the FP-TOS is kept in a register, this works. If
15579: it is kept on the stack, i.e., in memory, the store into memory has to
15580: wait for the result of the floating-point operation, lengthening the
15581: execution time of the primitive considerably.
15582:
15583: The TOS optimization makes the automatic generation of primitives a
15584: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
15585: @code{TOS} is not sufficient. There are some special cases to
15586: consider:
15587: @itemize @bullet
15588: @item In the case of @code{dup ( w -- w w )} the generator must not
15589: eliminate the store to the original location of the item on the stack,
15590: if the TOS optimization is turned on.
15591: @item Primitives with stack effects of the form @code{--}
1.29 crook 15592: @i{out1}...@i{outy} must store the TOS to the stack at the start.
15593: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
1.1 anton 15594: must load the TOS from the stack at the end. But for the null stack
15595: effect @code{--} no stores or loads should be generated.
15596: @end itemize
15597:
15598: @node Produced code, , TOS Optimization, Primitives
15599: @subsection Produced code
15600: @cindex primitives, assembly code listing
15601:
15602: @cindex @file{engine.s}
15603: To see what assembly code is produced for the primitives on your machine
15604: with your compiler and your flag settings, type @code{make engine.s} and
1.81 anton 15605: look at the resulting file @file{engine.s}. Alternatively, you can also
15606: disassemble the code of primitives with @code{see} on some architectures.
1.1 anton 15607:
15608: @node Performance, , Primitives, Engine
15609: @section Performance
15610: @cindex performance of some Forth interpreters
15611: @cindex engine performance
15612: @cindex benchmarking Forth systems
15613: @cindex Gforth performance
15614:
15615: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
1.112 anton 15616: impossible to write a significantly faster threaded-code engine.
1.1 anton 15617:
15618: On register-starved machines like the 386 architecture processors
15619: improvements are possible, because @code{gcc} does not utilize the
15620: registers as well as a human, even with explicit register declarations;
15621: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
15622: and hand-tuned it for the 486; this system is 1.19 times faster on the
15623: Sieve benchmark on a 486DX2/66 than Gforth compiled with
1.40 anton 15624: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
15625: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
15626: registers fit in real registers (and we can even afford to use the TOS
15627: optimization), resulting in a speedup of 1.14 on the sieve over the
1.112 anton 15628: earlier results. And dynamic superinstructions provide another speedup
15629: (but only around a factor 1.2 on the 486).
1.1 anton 15630:
15631: @cindex Win32Forth performance
15632: @cindex NT Forth performance
15633: @cindex eforth performance
15634: @cindex ThisForth performance
15635: @cindex PFE performance
15636: @cindex TILE performance
1.81 anton 15637: The potential advantage of assembly language implementations is not
1.112 anton 15638: necessarily realized in complete Forth systems: We compared Gforth-0.5.9
1.81 anton 15639: (direct threaded, compiled with @code{gcc-2.95.1} and
15640: @code{-DFORCE_REG}) with Win32Forth 1.2093 (newer versions are
15641: reportedly much faster), LMI's NT Forth (Beta, May 1994) and Eforth
15642: (with and without peephole (aka pinhole) optimization of the threaded
15643: code); all these systems were written in assembly language. We also
15644: compared Gforth with three systems written in C: PFE-0.9.14 (compiled
15645: with @code{gcc-2.6.3} with the default configuration for Linux:
15646: @code{-O2 -fomit-frame-pointer -DUSE_REGS -DUNROLL_NEXT}), ThisForth
15647: Beta (compiled with @code{gcc-2.6.3 -O3 -fomit-frame-pointer}; ThisForth
15648: employs peephole optimization of the threaded code) and TILE (compiled
15649: with @code{make opt}). We benchmarked Gforth, PFE, ThisForth and TILE on
15650: a 486DX2/66 under Linux. Kenneth O'Heskin kindly provided the results
15651: for Win32Forth and NT Forth on a 486DX2/66 with similar memory
15652: performance under Windows NT. Marcel Hendrix ported Eforth to Linux,
15653: then extended it to run the benchmarks, added the peephole optimizer,
15654: ran the benchmarks and reported the results.
1.40 anton 15655:
1.1 anton 15656: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
15657: matrix multiplication come from the Stanford integer benchmarks and have
15658: been translated into Forth by Martin Fraeman; we used the versions
15659: included in the TILE Forth package, but with bigger data set sizes; and
15660: a recursive Fibonacci number computation for benchmarking calling
15661: performance. The following table shows the time taken for the benchmarks
15662: scaled by the time taken by Gforth (in other words, it shows the speedup
15663: factor that Gforth achieved over the other systems).
15664:
15665: @example
1.112 anton 15666: relative Win32- NT eforth This-
15667: time Gforth Forth Forth eforth +opt PFE Forth TILE
15668: sieve 1.00 2.16 1.78 2.16 1.32 2.46 4.96 13.37
15669: bubble 1.00 1.93 2.07 2.18 1.29 2.21 5.70
15670: matmul 1.00 1.92 1.76 1.90 0.96 2.06 5.32
15671: fib 1.00 2.32 2.03 1.86 1.31 2.64 4.55 6.54
1.1 anton 15672: @end example
15673:
1.26 crook 15674: You may be quite surprised by the good performance of Gforth when
15675: compared with systems written in assembly language. One important reason
15676: for the disappointing performance of these other systems is probably
15677: that they are not written optimally for the 486 (e.g., they use the
15678: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
15679: but costly method for relocating the Forth image: like @code{cforth}, it
15680: computes the actual addresses at run time, resulting in two address
15681: computations per @code{NEXT} (@pxref{Image File Background}).
15682:
1.1 anton 15683: The speedup of Gforth over PFE, ThisForth and TILE can be easily
15684: explained with the self-imposed restriction of the latter systems to
15685: standard C, which makes efficient threading impossible (however, the
1.4 anton 15686: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
1.1 anton 15687: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
15688: Moreover, current C compilers have a hard time optimizing other aspects
15689: of the ThisForth and the TILE source.
15690:
1.26 crook 15691: The performance of Gforth on 386 architecture processors varies widely
15692: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
15693: allocate any of the virtual machine registers into real machine
15694: registers by itself and would not work correctly with explicit register
1.112 anton 15695: declarations, giving a significantly slower engine (on a 486DX2/66
15696: running the Sieve) than the one measured above.
1.1 anton 15697:
1.26 crook 15698: Note that there have been several releases of Win32Forth since the
15699: release presented here, so the results presented above may have little
1.40 anton 15700: predictive value for the performance of Win32Forth today (results for
15701: the current release on an i486DX2/66 are welcome).
1.1 anton 15702:
15703: @cindex @file{Benchres}
1.66 anton 15704: In
15705: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
15706: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
1.1 anton 15707: Maierhofer (presented at EuroForth '95), an indirect threaded version of
1.66 anton 15708: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
15709: several native code systems; that version of Gforth is slower on a 486
1.112 anton 15710: than the version used here. You can find a newer version of these
15711: measurements at
1.47 crook 15712: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
1.1 anton 15713: find numbers for Gforth on various machines in @file{Benchres}.
15714:
1.26 crook 15715: @c ******************************************************************
1.113 anton 15716: @c @node Binding to System Library, Cross Compiler, Engine, Top
15717: @c @chapter Binding to System Library
1.13 pazsan 15718:
1.113 anton 15719: @c ****************************************************************
15720: @node Cross Compiler, Bugs, Engine, Top
1.14 pazsan 15721: @chapter Cross Compiler
1.47 crook 15722: @cindex @file{cross.fs}
15723: @cindex cross-compiler
15724: @cindex metacompiler
15725: @cindex target compiler
1.13 pazsan 15726:
1.46 pazsan 15727: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
15728: mostly written in Forth, including crucial parts like the outer
15729: interpreter and compiler, it needs compiled Forth code to get
15730: started. The cross compiler allows to create new images for other
15731: architectures, even running under another Forth system.
1.13 pazsan 15732:
15733: @menu
1.67 anton 15734: * Using the Cross Compiler::
15735: * How the Cross Compiler Works::
1.13 pazsan 15736: @end menu
15737:
1.21 crook 15738: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
1.14 pazsan 15739: @section Using the Cross Compiler
1.46 pazsan 15740:
15741: The cross compiler uses a language that resembles Forth, but isn't. The
15742: main difference is that you can execute Forth code after definition,
15743: while you usually can't execute the code compiled by cross, because the
15744: code you are compiling is typically for a different computer than the
15745: one you are compiling on.
15746:
1.81 anton 15747: @c anton: This chapter is somewhat different from waht I would expect: I
15748: @c would expect an explanation of the cross language and how to create an
15749: @c application image with it. The section explains some aspects of
15750: @c creating a Gforth kernel.
15751:
1.46 pazsan 15752: The Makefile is already set up to allow you to create kernels for new
15753: architectures with a simple make command. The generic kernels using the
15754: GCC compiled virtual machine are created in the normal build process
15755: with @code{make}. To create a embedded Gforth executable for e.g. the
15756: 8086 processor (running on a DOS machine), type
15757:
15758: @example
15759: make kernl-8086.fi
15760: @end example
15761:
15762: This will use the machine description from the @file{arch/8086}
15763: directory to create a new kernel. A machine file may look like that:
15764:
15765: @example
15766: \ Parameter for target systems 06oct92py
15767:
15768: 4 Constant cell \ cell size in bytes
15769: 2 Constant cell<< \ cell shift to bytes
15770: 5 Constant cell>bit \ cell shift to bits
15771: 8 Constant bits/char \ bits per character
15772: 8 Constant bits/byte \ bits per byte [default: 8]
15773: 8 Constant float \ bytes per float
15774: 8 Constant /maxalign \ maximum alignment in bytes
15775: false Constant bigendian \ byte order
15776: ( true=big, false=little )
15777:
15778: include machpc.fs \ feature list
15779: @end example
15780:
15781: This part is obligatory for the cross compiler itself, the feature list
15782: is used by the kernel to conditionally compile some features in and out,
15783: depending on whether the target supports these features.
15784:
15785: There are some optional features, if you define your own primitives,
15786: have an assembler, or need special, nonstandard preparation to make the
1.81 anton 15787: boot process work. @code{asm-include} includes an assembler,
1.46 pazsan 15788: @code{prims-include} includes primitives, and @code{>boot} prepares for
15789: booting.
15790:
15791: @example
15792: : asm-include ." Include assembler" cr
15793: s" arch/8086/asm.fs" included ;
15794:
15795: : prims-include ." Include primitives" cr
15796: s" arch/8086/prim.fs" included ;
15797:
15798: : >boot ." Prepare booting" cr
15799: s" ' boot >body into-forth 1+ !" evaluate ;
15800: @end example
15801:
15802: These words are used as sort of macro during the cross compilation in
1.81 anton 15803: the file @file{kernel/main.fs}. Instead of using these macros, it would
1.46 pazsan 15804: be possible --- but more complicated --- to write a new kernel project
15805: file, too.
15806:
15807: @file{kernel/main.fs} expects the machine description file name on the
15808: stack; the cross compiler itself (@file{cross.fs}) assumes that either
15809: @code{mach-file} leaves a counted string on the stack, or
15810: @code{machine-file} leaves an address, count pair of the filename on the
15811: stack.
15812:
15813: The feature list is typically controlled using @code{SetValue}, generic
15814: files that are used by several projects can use @code{DefaultValue}
15815: instead. Both functions work like @code{Value}, when the value isn't
15816: defined, but @code{SetValue} works like @code{to} if the value is
15817: defined, and @code{DefaultValue} doesn't set anything, if the value is
15818: defined.
15819:
15820: @example
15821: \ generic mach file for pc gforth 03sep97jaw
15822:
15823: true DefaultValue NIL \ relocating
15824:
15825: >ENVIRON
15826:
15827: true DefaultValue file \ controls the presence of the
15828: \ file access wordset
15829: true DefaultValue OS \ flag to indicate a operating system
15830:
15831: true DefaultValue prims \ true: primitives are c-code
15832:
15833: true DefaultValue floating \ floating point wordset is present
15834:
15835: true DefaultValue glocals \ gforth locals are present
15836: \ will be loaded
15837: true DefaultValue dcomps \ double number comparisons
15838:
15839: true DefaultValue hash \ hashing primitives are loaded/present
15840:
15841: true DefaultValue xconds \ used together with glocals,
15842: \ special conditionals supporting gforths'
15843: \ local variables
15844: true DefaultValue header \ save a header information
15845:
15846: true DefaultValue backtrace \ enables backtrace code
15847:
15848: false DefaultValue ec
15849: false DefaultValue crlf
15850:
15851: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
15852:
15853: &16 KB DefaultValue stack-size
15854: &15 KB &512 + DefaultValue fstack-size
15855: &15 KB DefaultValue rstack-size
15856: &14 KB &512 + DefaultValue lstack-size
15857: @end example
1.13 pazsan 15858:
1.48 anton 15859: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
1.14 pazsan 15860: @section How the Cross Compiler Works
1.13 pazsan 15861:
15862: @node Bugs, Origin, Cross Compiler, Top
1.21 crook 15863: @appendix Bugs
1.1 anton 15864: @cindex bug reporting
15865:
1.21 crook 15866: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
1.1 anton 15867:
1.103 anton 15868: If you find a bug, please submit a bug report through
15869: @uref{https://savannah.gnu.org/bugs/?func=addbug&group=gforth}.
1.21 crook 15870:
15871: @itemize @bullet
15872: @item
1.81 anton 15873: A program (or a sequence of keyboard commands) that reproduces the bug.
15874: @item
15875: A description of what you think constitutes the buggy behaviour.
15876: @item
1.21 crook 15877: The Gforth version used (it is announced at the start of an
15878: interactive Gforth session).
15879: @item
15880: The machine and operating system (on Unix
15881: systems @code{uname -a} will report this information).
15882: @item
1.81 anton 15883: The installation options (you can find the configure options at the
15884: start of @file{config.status}) and configuration (@code{configure}
15885: output or @file{config.cache}).
1.21 crook 15886: @item
15887: A complete list of changes (if any) you (or your installer) have made to the
15888: Gforth sources.
15889: @end itemize
1.1 anton 15890:
15891: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
15892: to Report Bugs, gcc.info, GNU C Manual}.
15893:
15894:
1.21 crook 15895: @node Origin, Forth-related information, Bugs, Top
15896: @appendix Authors and Ancestors of Gforth
1.1 anton 15897:
15898: @section Authors and Contributors
15899: @cindex authors of Gforth
15900: @cindex contributors to Gforth
15901:
15902: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
1.81 anton 15903: Ertl. The third major author was Jens Wilke. Neal Crook contributed a
15904: lot to the manual. Assemblers and disassemblers were contributed by
1.161 anton 15905: Andrew McKewan, Christian Pirker, Bernd Thallner, and Michal Revucky.
15906: Lennart Benschop (who was one of Gforth's first users, in mid-1993)
15907: and Stuart Ramsden inspired us with their continuous feedback. Lennart
15908: Benshop contributed @file{glosgen.fs}, while Stuart Ramsden has been
15909: working on automatic support for calling C libraries. Helpful comments
15910: also came from Paul Kleinrubatscher, Christian Pirker, Dirk Zoller,
15911: Marcel Hendrix, John Wavrik, Barrie Stott, Marc de Groot, Jorge
15912: Acerada, Bruce Hoyt, Robert Epprecht, Dennis Ruffer and David
15913: N. Williams. Since the release of Gforth-0.2.1 there were also helpful
15914: comments from many others; thank you all, sorry for not listing you
15915: here (but digging through my mailbox to extract your names is on my
15916: to-do list).
1.1 anton 15917:
15918: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
15919: and autoconf, among others), and to the creators of the Internet: Gforth
1.21 crook 15920: was developed across the Internet, and its authors did not meet
1.20 pazsan 15921: physically for the first 4 years of development.
1.1 anton 15922:
15923: @section Pedigree
1.26 crook 15924: @cindex pedigree of Gforth
1.1 anton 15925:
1.81 anton 15926: Gforth descends from bigFORTH (1993) and fig-Forth. Of course, a
15927: significant part of the design of Gforth was prescribed by ANS Forth.
1.1 anton 15928:
1.20 pazsan 15929: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
1.1 anton 15930: 32 bit native code version of VolksForth for the Atari ST, written
15931: mostly by Dietrich Weineck.
15932:
1.81 anton 15933: VolksForth was written by Klaus Schleisiek, Bernd Pennemann, Georg
15934: Rehfeld and Dietrich Weineck for the C64 (called UltraForth there) in
1.147 anton 15935: the mid-80s and ported to the Atari ST in 1986. It descends from fig-Forth.
1.1 anton 15936:
1.147 anton 15937: @c Henry Laxen and Mike Perry wrote F83 as a model implementation of the
15938: @c Forth-83 standard. !! Pedigree? When?
1.1 anton 15939:
15940: A team led by Bill Ragsdale implemented fig-Forth on many processors in
15941: 1979. Robert Selzer and Bill Ragsdale developed the original
15942: implementation of fig-Forth for the 6502 based on microForth.
15943:
15944: The principal architect of microForth was Dean Sanderson. microForth was
15945: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
15946: the 1802, and subsequently implemented on the 8080, the 6800 and the
15947: Z80.
15948:
15949: All earlier Forth systems were custom-made, usually by Charles Moore,
15950: who discovered (as he puts it) Forth during the late 60s. The first full
15951: Forth existed in 1971.
15952:
1.81 anton 15953: A part of the information in this section comes from
15954: @cite{@uref{http://www.forth.com/Content/History/History1.htm,The
15955: Evolution of Forth}} by Elizabeth D. Rather, Donald R. Colburn and
1.147 anton 15956: Charles H. Moore, presented at the HOPL-II conference and preprinted
15957: in SIGPLAN Notices 28(3), 1993. You can find more historical and
15958: genealogical information about Forth there. For a more general (and
15959: graphical) Forth family tree look see
15960: @cite{@uref{http://www.complang.tuwien.ac.at/forth/family-tree/},
15961: Forth Family Tree and Timeline}.
1.1 anton 15962:
1.81 anton 15963: @c ------------------------------------------------------------------
1.113 anton 15964: @node Forth-related information, Licenses, Origin, Top
1.21 crook 15965: @appendix Other Forth-related information
15966: @cindex Forth-related information
15967:
1.81 anton 15968: @c anton: I threw most of this stuff out, because it can be found through
15969: @c the FAQ and the FAQ is more likely to be up-to-date.
1.21 crook 15970:
15971: @cindex comp.lang.forth
15972: @cindex frequently asked questions
1.81 anton 15973: There is an active news group (comp.lang.forth) discussing Forth
15974: (including Gforth) and Forth-related issues. Its
15975: @uref{http://www.complang.tuwien.ac.at/forth/faq/faq-general-2.html,FAQs}
15976: (frequently asked questions and their answers) contains a lot of
15977: information on Forth. You should read it before posting to
15978: comp.lang.forth.
1.21 crook 15979:
1.81 anton 15980: The ANS Forth standard is most usable in its
15981: @uref{http://www.taygeta.com/forth/dpans.html, HTML form}.
1.21 crook 15982:
1.113 anton 15983: @c ---------------------------------------------------
15984: @node Licenses, Word Index, Forth-related information, Top
15985: @appendix Licenses
15986:
15987: @menu
15988: * GNU Free Documentation License:: License for copying this manual.
15989: * Copying:: GPL (for copying this software).
15990: @end menu
15991:
15992: @include fdl.texi
15993:
15994: @include gpl.texi
15995:
15996:
15997:
1.81 anton 15998: @c ------------------------------------------------------------------
1.113 anton 15999: @node Word Index, Concept Index, Licenses, Top
1.1 anton 16000: @unnumbered Word Index
16001:
1.26 crook 16002: This index is a list of Forth words that have ``glossary'' entries
16003: within this manual. Each word is listed with its stack effect and
16004: wordset.
1.1 anton 16005:
16006: @printindex fn
16007:
1.81 anton 16008: @c anton: the name index seems superfluous given the word and concept indices.
16009:
16010: @c @node Name Index, Concept Index, Word Index, Top
16011: @c @unnumbered Name Index
1.41 anton 16012:
1.81 anton 16013: @c This index is a list of Forth words that have ``glossary'' entries
16014: @c within this manual.
1.41 anton 16015:
1.81 anton 16016: @c @printindex ky
1.41 anton 16017:
1.113 anton 16018: @c -------------------------------------------------------
1.81 anton 16019: @node Concept Index, , Word Index, Top
1.1 anton 16020: @unnumbered Concept and Word Index
16021:
1.26 crook 16022: Not all entries listed in this index are present verbatim in the
16023: text. This index also duplicates, in abbreviated form, all of the words
16024: listed in the Word Index (only the names are listed for the words here).
1.1 anton 16025:
16026: @printindex cp
16027:
16028: @bye
1.81 anton 16029:
16030:
1.1 anton 16031:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>