1: \input texinfo @c -*-texinfo-*-
2: @comment The source is gforth.ds, from which gforth.texi is generated
3:
4: @comment TODO: nac29jan99 - a list of things to add in the next edit:
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.
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
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
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:
24: @c POSTPONE, COMPILE, [COMPILE], LITERAL should have their own section
25:
26: @comment %**start of header (This is for running Texinfo on a region.)
27: @setfilename gforth.info
28: @include version.texi
29: @settitle Gforth Manual
30: @c @syncodeindex pg cp
31:
32: @macro progstyle {}
33: Programming style note:
34: @end macro
35:
36: @macro assignment {}
37: @table @i
38: @item Assignment:
39: @end macro
40: @macro endassignment {}
41: @end table
42: @end macro
43:
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:
57: @comment %**end of header (This is for running Texinfo on a region.)
58: @copying
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.
63:
64: Copyright @copyright{} 1995, 1996, 1997, 1998, 2000, 2003, 2004,2005 Free Software Foundation, Inc.
65:
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
80:
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
91:
92: @titlepage
93: @title Gforth
94: @subtitle for version @value{VERSION}, @value{UPDATED}
95: @author Neal Crook
96: @author Anton Ertl
97: @author David Kuehling
98: @author Bernd Paysan
99: @author Jens Wilke
100: @page
101: @vskip 0pt plus 1filll
102: @insertcopying
103: @end titlepage
104:
105: @contents
106:
107: @ifnottex
108: @node Top, Goals, (dir), (dir)
109: @top Gforth
110:
111: @insertcopying
112: @end ifnottex
113:
114: @menu
115: * Goals:: About the Gforth Project
116: * Gforth Environment:: Starting (and exiting) Gforth
117: * Tutorial:: Hands-on Forth Tutorial
118: * Introduction:: An introduction to ANS Forth
119: * Words:: Forth words available in Gforth
120: * Error messages:: How to interpret them
121: * Tools:: Programming tools
122: * ANS conformance:: Implementation-defined options etc.
123: * Standard vs Extensions:: Should I use extensions?
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
129: * Cross Compiler:: The Cross Compiler
130: * Bugs:: How to report them
131: * Origin:: Authors and ancestors of Gforth
132: * Forth-related information:: Books and places to look on the WWW
133: * Licenses::
134: * Word Index:: An item for each Forth word
135: * Concept Index:: A menu covering many topics
136:
137: @detailmenu
138: --- The Detailed Node Listing ---
139:
140: Gforth Environment
141:
142: * Invoking Gforth:: Getting in
143: * Leaving Gforth:: Getting out
144: * Command-line editing::
145: * Environment variables:: that affect how Gforth starts up
146: * Gforth Files:: What gets installed and where
147: * Gforth in pipes::
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::
177: * Files Tutorial::
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::
188:
189: An Introduction to ANS Forth
190:
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::
199:
200: Forth Words
201:
202: * Notation::
203: * Case insensitivity::
204: * Comments::
205: * Boolean Flags::
206: * Arithmetic::
207: * Stack Manipulation::
208: * Memory::
209: * Control Structures::
210: * Defining Words::
211: * Interpretation and Compilation Semantics::
212: * Tokens for Words::
213: * Compiling words::
214: * The Text Interpreter::
215: * The Input Stream::
216: * Word Lists::
217: * Environmental Queries::
218: * Files::
219: * Blocks::
220: * Other I/O::
221: * OS command line arguments::
222: * Locals::
223: * Structures::
224: * Object-oriented Forth::
225: * Programming Tools::
226: * C Interface::
227: * Assembler and Code Words::
228: * Threading Words::
229: * Passing Commands to the OS::
230: * Keeping track of Time::
231: * Miscellaneous Words::
232:
233: Arithmetic
234:
235: * Single precision::
236: * Double precision:: Double-cell integer arithmetic
237: * Bitwise operations::
238: * Numeric comparison::
239: * Mixed precision:: Operations with single and double-cell integers
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:
252: * Memory model::
253: * Dictionary allocation::
254: * Heap Allocation::
255: * Memory Access::
256: * Address arithmetic::
257: * Memory Blocks::
258:
259: Control Structures
260:
261: * Selection:: IF ... ELSE ... ENDIF
262: * Simple Loops:: BEGIN ...
263: * Counted Loops:: DO
264: * Arbitrary control structures::
265: * Calls and returns::
266: * Exception Handling::
267:
268: Defining Words
269:
270: * CREATE::
271: * Variables:: Variables and user variables
272: * Constants::
273: * Values:: Initialised variables
274: * Colon Definitions::
275: * Anonymous Definitions:: Definitions without names
276: * Supplying names:: Passing definition names as strings
277: * User-defined Defining Words::
278: * Deferred words:: Allow forward references
279: * Aliases::
280:
281: User-defined Defining Words
282:
283: * CREATE..DOES> applications::
284: * CREATE..DOES> details::
285: * Advanced does> usage example::
286: * Const-does>::
287:
288: Interpretation and Compilation Semantics
289:
290: * Combined words::
291:
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:
298: Compiling words
299:
300: * Literals:: Compiling data values
301: * Macros:: Compiling words
302:
303: The Text Interpreter
304:
305: * Input Sources::
306: * Number Conversion::
307: * Interpret/Compile states::
308: * Interpreter Directives::
309:
310: Word Lists
311:
312: * Vocabularies::
313: * Why use word lists?::
314: * Word list example::
315:
316: Files
317:
318: * Forth source files::
319: * General files::
320: * Redirection::
321: * Search Paths::
322:
323: Search Paths
324:
325: * Source Search Paths::
326: * General Search Paths::
327:
328: Other I/O
329:
330: * Simple numeric output:: Predefined formats
331: * Formatted numeric output:: Formatted (pictured) output
332: * String Formats:: How Forth stores strings in memory
333: * Displaying characters and strings:: Other stuff
334: * Input:: Input
335: * Pipes:: How to create your own pipes
336: * Xchars and Unicode:: Non-ASCII characters
337:
338: Locals
339:
340: * Gforth locals::
341: * ANS Forth locals::
342:
343: Gforth locals
344:
345: * Where are locals visible by name?::
346: * How long do locals live?::
347: * Locals programming style::
348: * Locals implementation::
349:
350: Structures
351:
352: * Why explicit structure support?::
353: * Structure Usage::
354: * Structure Naming Convention::
355: * Structure Implementation::
356: * Structure Glossary::
357:
358: Object-oriented Forth
359:
360: * Why object-oriented programming?::
361: * Object-Oriented Terminology::
362: * Objects::
363: * OOF::
364: * Mini-OOF::
365: * Comparison with other object models::
366:
367: The @file{objects.fs} model
368:
369: * Properties of the Objects model::
370: * Basic Objects Usage::
371: * The Objects base class::
372: * Creating objects::
373: * Object-Oriented Programming Style::
374: * Class Binding::
375: * Method conveniences::
376: * Classes and Scoping::
377: * Dividing classes::
378: * Object Interfaces::
379: * Objects Implementation::
380: * Objects Glossary::
381:
382: The @file{oof.fs} model
383:
384: * Properties of the OOF model::
385: * Basic OOF Usage::
386: * The OOF base class::
387: * Class Declaration::
388: * Class Implementation::
389:
390: The @file{mini-oof.fs} model
391:
392: * Basic Mini-OOF Usage::
393: * Mini-OOF Example::
394: * Mini-OOF Implementation::
395:
396: Programming Tools
397:
398: * Examining:: Data and Code.
399: * Forgetting words:: Usually before reloading.
400: * Debugging:: Simple and quick.
401: * Assertions:: Making your programs self-checking.
402: * Singlestep Debugger:: Executing your program word by word.
403:
404: C Interface
405:
406: * Calling C Functions::
407: * Declaring C Functions::
408: * Callbacks::
409: * Low-Level C Interface Words::
410:
411: Assembler and Code Words
412:
413: * Code and ;code::
414: * Common Assembler:: Assembler Syntax
415: * Common Disassembler::
416: * 386 Assembler:: Deviations and special cases
417: * Alpha Assembler:: Deviations and special cases
418: * MIPS assembler:: Deviations and special cases
419: * PowerPC assembler:: Deviations and special cases
420: * Other assemblers:: How to write them
421:
422: Tools
423:
424: * ANS Report:: Report the words used, sorted by wordset.
425: * Stack depth changes:: Where does this stack item come from?
426:
427: ANS conformance
428:
429: * The Core Words::
430: * The optional Block word set::
431: * The optional Double Number word set::
432: * The optional Exception word set::
433: * The optional Facility word set::
434: * The optional File-Access word set::
435: * The optional Floating-Point word set::
436: * The optional Locals word set::
437: * The optional Memory-Allocation word set::
438: * The optional Programming-Tools word set::
439: * The optional Search-Order word set::
440:
441: The Core Words
442:
443: * core-idef:: Implementation Defined Options
444: * core-ambcond:: Ambiguous Conditions
445: * core-other:: Other System Documentation
446:
447: The optional Block word set
448:
449: * block-idef:: Implementation Defined Options
450: * block-ambcond:: Ambiguous Conditions
451: * block-other:: Other System Documentation
452:
453: The optional Double Number word set
454:
455: * double-ambcond:: Ambiguous Conditions
456:
457: The optional Exception word set
458:
459: * exception-idef:: Implementation Defined Options
460:
461: The optional Facility word set
462:
463: * facility-idef:: Implementation Defined Options
464: * facility-ambcond:: Ambiguous Conditions
465:
466: The optional File-Access word set
467:
468: * file-idef:: Implementation Defined Options
469: * file-ambcond:: Ambiguous Conditions
470:
471: The optional Floating-Point word set
472:
473: * floating-idef:: Implementation Defined Options
474: * floating-ambcond:: Ambiguous Conditions
475:
476: The optional Locals word set
477:
478: * locals-idef:: Implementation Defined Options
479: * locals-ambcond:: Ambiguous Conditions
480:
481: The optional Memory-Allocation word set
482:
483: * memory-idef:: Implementation Defined Options
484:
485: The optional Programming-Tools word set
486:
487: * programming-idef:: Implementation Defined Options
488: * programming-ambcond:: Ambiguous Conditions
489:
490: The optional Search-Order word set
491:
492: * search-idef:: Implementation Defined Options
493: * search-ambcond:: Ambiguous Conditions
494:
495: Emacs and Gforth
496:
497: * Installing gforth.el:: Making Emacs aware of Forth.
498: * Emacs Tags:: Viewing the source of a word in Emacs.
499: * Hilighting:: Making Forth code look prettier.
500: * Auto-Indentation:: Customizing auto-indentation.
501: * Blocks Files:: Reading and writing blocks files.
502:
503: Image Files
504:
505: * Image Licensing Issues:: Distribution terms for images.
506: * Image File Background:: Why have image files?
507: * Non-Relocatable Image Files:: don't always work.
508: * Data-Relocatable Image Files:: are better.
509: * Fully Relocatable Image Files:: better yet.
510: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
511: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
512: * Modifying the Startup Sequence:: and turnkey applications.
513:
514: Fully Relocatable Image Files
515:
516: * gforthmi:: The normal way
517: * cross.fs:: The hard way
518:
519: Engine
520:
521: * Portability::
522: * Threading::
523: * Primitives::
524: * Performance::
525:
526: Threading
527:
528: * Scheduling::
529: * Direct or Indirect Threaded?::
530: * Dynamic Superinstructions::
531: * DOES>::
532:
533: Primitives
534:
535: * Automatic Generation::
536: * TOS Optimization::
537: * Produced code::
538:
539: Cross Compiler
540:
541: * Using the Cross Compiler::
542: * How the Cross Compiler Works::
543:
544: Licenses
545:
546: * GNU Free Documentation License:: License for copying this manual.
547: * Copying:: GPL (for copying this software).
548:
549: @end detailmenu
550: @end menu
551:
552: @c ----------------------------------------------------------
553: @iftex
554: @unnumbered Preface
555: @cindex Preface
556: This manual documents Gforth. Some introductory material is provided for
557: readers who are unfamiliar with Forth or who are migrating to Gforth
558: from other Forth compilers. However, this manual is primarily a
559: reference manual.
560: @end iftex
561:
562: @comment TODO much more blurb here.
563:
564: @c ******************************************************************
565: @node Goals, Gforth Environment, Top, Top
566: @comment node-name, next, previous, up
567: @chapter Goals of Gforth
568: @cindex goals of the Gforth project
569: The goal of the Gforth Project is to develop a standard model for
570: ANS Forth. This can be split into several subgoals:
571:
572: @itemize @bullet
573: @item
574: Gforth should conform to the ANS Forth Standard.
575: @item
576: It should be a model, i.e. it should define all the
577: implementation-dependent things.
578: @item
579: It should become standard, i.e. widely accepted and used. This goal
580: is the most difficult one.
581: @end itemize
582:
583: To achieve these goals Gforth should be
584: @itemize @bullet
585: @item
586: Similar to previous models (fig-Forth, F83)
587: @item
588: Powerful. It should provide for all the things that are considered
589: necessary today and even some that are not yet considered necessary.
590: @item
591: Efficient. It should not get the reputation of being exceptionally
592: slow.
593: @item
594: Free.
595: @item
596: Available on many machines/easy to port.
597: @end itemize
598:
599: Have we achieved these goals? Gforth conforms to the ANS Forth
600: standard. It may be considered a model, but we have not yet documented
601: which parts of the model are stable and which parts we are likely to
602: change. It certainly has not yet become a de facto standard, but it
603: appears to be quite popular. It has some similarities to and some
604: differences from previous models. It has some powerful features, but not
605: yet everything that we envisioned. We certainly have achieved our
606: execution speed goals (@pxref{Performance})@footnote{However, in 1998
607: the bar was raised when the major commercial Forth vendors switched to
608: native code compilers.}. It is free and available on many machines.
609:
610: @c ******************************************************************
611: @node Gforth Environment, Tutorial, Goals, Top
612: @chapter Gforth Environment
613: @cindex Gforth environment
614:
615: Note: ultimately, the Gforth man page will be auto-generated from the
616: material in this chapter.
617:
618: @menu
619: * Invoking Gforth:: Getting in
620: * Leaving Gforth:: Getting out
621: * Command-line editing::
622: * Environment variables:: that affect how Gforth starts up
623: * Gforth Files:: What gets installed and where
624: * Gforth in pipes::
625: * Startup speed:: When 35ms is not fast enough ...
626: @end menu
627:
628: For related information about the creation of images see @ref{Image Files}.
629:
630: @comment ----------------------------------------------
631: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
632: @section Invoking Gforth
633: @cindex invoking Gforth
634: @cindex running Gforth
635: @cindex command-line options
636: @cindex options on the command line
637: @cindex flags on the command line
638:
639: Gforth is made up of two parts; an executable ``engine'' (named
640: @command{gforth} or @command{gforth-fast}) and an image file. To start it, you
641: will usually just say @code{gforth} -- this automatically loads the
642: default image file @file{gforth.fi}. In many other cases the default
643: Gforth image will be invoked like this:
644: @example
645: gforth [file | -e forth-code] ...
646: @end example
647: @noindent
648: This interprets the contents of the files and the Forth code in the order they
649: are given.
650:
651: In addition to the @command{gforth} engine, there is also an engine
652: called @command{gforth-fast}, which is faster, but gives less
653: informative error messages (@pxref{Error messages}) and may catch some
654: errors (in particular, stack underflows and integer division errors)
655: later or not at all. You should use it for debugged,
656: performance-critical programs.
657:
658: Moreover, there is an engine called @command{gforth-itc}, which is
659: useful in some backwards-compatibility situations (@pxref{Direct or
660: Indirect Threaded?}).
661:
662: In general, the command line looks like this:
663:
664: @example
665: gforth[-fast] [engine options] [image options]
666: @end example
667:
668: The engine options must come before the rest of the command
669: line. They are:
670:
671: @table @code
672: @cindex -i, command-line option
673: @cindex --image-file, command-line option
674: @item --image-file @i{file}
675: @itemx -i @i{file}
676: Loads the Forth image @i{file} instead of the default
677: @file{gforth.fi} (@pxref{Image Files}).
678:
679: @cindex --appl-image, command-line option
680: @item --appl-image @i{file}
681: Loads the image @i{file} and leaves all further command-line arguments
682: to the image (instead of processing them as engine options). This is
683: useful for building executable application images on Unix, built with
684: @code{gforthmi --application ...}.
685:
686: @cindex --path, command-line option
687: @cindex -p, command-line option
688: @item --path @i{path}
689: @itemx -p @i{path}
690: Uses @i{path} for searching the image file and Forth source code files
691: instead of the default in the environment variable @code{GFORTHPATH} or
692: the path specified at installation time (e.g.,
693: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
694: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
695:
696: @cindex --dictionary-size, command-line option
697: @cindex -m, command-line option
698: @cindex @i{size} parameters for command-line options
699: @cindex size of the dictionary and the stacks
700: @item --dictionary-size @i{size}
701: @itemx -m @i{size}
702: Allocate @i{size} space for the Forth dictionary space instead of
703: using the default specified in the image (typically 256K). The
704: @i{size} specification for this and subsequent options consists of
705: an integer and a unit (e.g.,
706: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
707: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
708: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
709: @code{e} is used.
710:
711: @cindex --data-stack-size, command-line option
712: @cindex -d, command-line option
713: @item --data-stack-size @i{size}
714: @itemx -d @i{size}
715: Allocate @i{size} space for the data stack instead of using the
716: default specified in the image (typically 16K).
717:
718: @cindex --return-stack-size, command-line option
719: @cindex -r, command-line option
720: @item --return-stack-size @i{size}
721: @itemx -r @i{size}
722: Allocate @i{size} space for the return stack instead of using the
723: default specified in the image (typically 15K).
724:
725: @cindex --fp-stack-size, command-line option
726: @cindex -f, command-line option
727: @item --fp-stack-size @i{size}
728: @itemx -f @i{size}
729: Allocate @i{size} space for the floating point stack instead of
730: using the default specified in the image (typically 15.5K). In this case
731: the unit specifier @code{e} refers to floating point numbers.
732:
733: @cindex --locals-stack-size, command-line option
734: @cindex -l, command-line option
735: @item --locals-stack-size @i{size}
736: @itemx -l @i{size}
737: Allocate @i{size} space for the locals stack instead of using the
738: default specified in the image (typically 14.5K).
739:
740: @cindex -h, command-line option
741: @cindex --help, command-line option
742: @item --help
743: @itemx -h
744: Print a message about the command-line options
745:
746: @cindex -v, command-line option
747: @cindex --version, command-line option
748: @item --version
749: @itemx -v
750: Print version and exit
751:
752: @cindex --debug, command-line option
753: @item --debug
754: Print some information useful for debugging on startup.
755:
756: @cindex --offset-image, command-line option
757: @item --offset-image
758: Start the dictionary at a slightly different position than would be used
759: otherwise (useful for creating data-relocatable images,
760: @pxref{Data-Relocatable Image Files}).
761:
762: @cindex --no-offset-im, command-line option
763: @item --no-offset-im
764: Start the dictionary at the normal position.
765:
766: @cindex --clear-dictionary, command-line option
767: @item --clear-dictionary
768: Initialize all bytes in the dictionary to 0 before loading the image
769: (@pxref{Data-Relocatable Image Files}).
770:
771: @cindex --die-on-signal, command-line-option
772: @item --die-on-signal
773: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
774: or the segmentation violation SIGSEGV) by translating it into a Forth
775: @code{THROW}. With this option, Gforth exits if it receives such a
776: signal. This option is useful when the engine and/or the image might be
777: severely broken (such that it causes another signal before recovering
778: from the first); this option avoids endless loops in such cases.
779:
780: @cindex --no-dynamic, command-line option
781: @cindex --dynamic, command-line option
782: @item --no-dynamic
783: @item --dynamic
784: Disable or enable dynamic superinstructions with replication
785: (@pxref{Dynamic Superinstructions}).
786:
787: @cindex --no-super, command-line option
788: @item --no-super
789: Disable dynamic superinstructions, use just dynamic replication; this is
790: useful if you want to patch threaded code (@pxref{Dynamic
791: Superinstructions}).
792:
793: @cindex --ss-number, command-line option
794: @item --ss-number=@var{N}
795: Use only the first @var{N} static superinstructions compiled into the
796: engine (default: use them all; note that only @code{gforth-fast} has
797: any). This option is useful for measuring the performance impact of
798: static superinstructions.
799:
800: @cindex --ss-min-..., command-line options
801: @item --ss-min-codesize
802: @item --ss-min-ls
803: @item --ss-min-lsu
804: @item --ss-min-nexts
805: Use specified metric for determining the cost of a primitive or static
806: superinstruction for static superinstruction selection. @code{Codesize}
807: is the native code size of the primive or static superinstruction,
808: @code{ls} is the number of loads and stores, @code{lsu} is the number of
809: loads, stores, and updates, and @code{nexts} is the number of dispatches
810: (not taking dynamic superinstructions into account), i.e. every
811: primitive or static superinstruction has cost 1. Default:
812: @code{codesize} if you use dynamic code generation, otherwise
813: @code{nexts}.
814:
815: @cindex --ss-greedy, command-line option
816: @item --ss-greedy
817: This option is useful for measuring the performance impact of static
818: superinstructions. By default, an optimal shortest-path algorithm is
819: used for selecting static superinstructions. With @option{--ss-greedy}
820: this algorithm is modified to assume that anything after the static
821: superinstruction currently under consideration is not combined into
822: static superinstructions. With @option{--ss-min-nexts} this produces
823: the same result as a greedy algorithm that always selects the longest
824: superinstruction available at the moment. E.g., if there are
825: superinstructions AB and BCD, then for the sequence A B C D the optimal
826: algorithm will select A BCD and the greedy algorithm will select AB C D.
827:
828: @cindex --print-metrics, command-line option
829: @item --print-metrics
830: Prints some metrics used during static superinstruction selection:
831: @code{code size} is the actual size of the dynamically generated code.
832: @code{Metric codesize} is the sum of the codesize metrics as seen by
833: static superinstruction selection; there is a difference from @code{code
834: size}, because not all primitives and static superinstructions are
835: compiled into dynamically generated code, and because of markers. The
836: other metrics correspond to the @option{ss-min-...} options. This
837: option is useful for evaluating the effects of the @option{--ss-...}
838: options.
839:
840: @end table
841:
842: @cindex loading files at startup
843: @cindex executing code on startup
844: @cindex batch processing with Gforth
845: As explained above, the image-specific command-line arguments for the
846: default image @file{gforth.fi} consist of a sequence of filenames and
847: @code{-e @var{forth-code}} options that are interpreted in the sequence
848: in which they are given. The @code{-e @var{forth-code}} or
849: @code{--evaluate @var{forth-code}} option evaluates the Forth code. This
850: option takes only one argument; if you want to evaluate more Forth
851: words, you have to quote them or use @code{-e} several times. To exit
852: after processing the command line (instead of entering interactive mode)
853: append @code{-e bye} to the command line. You can also process the
854: command-line arguments with a Forth program (@pxref{OS command line
855: arguments}).
856:
857: @cindex versions, invoking other versions of Gforth
858: If you have several versions of Gforth installed, @code{gforth} will
859: invoke the version that was installed last. @code{gforth-@i{version}}
860: invokes a specific version. If your environment contains the variable
861: @code{GFORTHPATH}, you may want to override it by using the
862: @code{--path} option.
863:
864: Not yet implemented:
865: On startup the system first executes the system initialization file
866: (unless the option @code{--no-init-file} is given; note that the system
867: resulting from using this option may not be ANS Forth conformant). Then
868: the user initialization file @file{.gforth.fs} is executed, unless the
869: option @code{--no-rc} is given; this file is searched for in @file{.},
870: then in @file{~}, then in the normal path (see above).
871:
872:
873:
874: @comment ----------------------------------------------
875: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
876: @section Leaving Gforth
877: @cindex Gforth - leaving
878: @cindex leaving Gforth
879:
880: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
881: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
882: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
883: data are discarded. For ways of saving the state of the system before
884: leaving Gforth see @ref{Image Files}.
885:
886: doc-bye
887:
888:
889: @comment ----------------------------------------------
890: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
891: @section Command-line editing
892: @cindex command-line editing
893:
894: Gforth maintains a history file that records every line that you type to
895: the text interpreter. This file is preserved between sessions, and is
896: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
897: repeatedly you can recall successively older commands from this (or
898: previous) session(s). The full list of command-line editing facilities is:
899:
900: @itemize @bullet
901: @item
902: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
903: commands from the history buffer.
904: @item
905: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
906: from the history buffer.
907: @item
908: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
909: @item
910: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
911: @item
912: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
913: closing up the line.
914: @item
915: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
916: @item
917: @kbd{Ctrl-a} to move the cursor to the start of the line.
918: @item
919: @kbd{Ctrl-e} to move the cursor to the end of the line.
920: @item
921: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
922: line.
923: @item
924: @key{TAB} to step through all possible full-word completions of the word
925: currently being typed.
926: @item
927: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
928: using @code{bye}).
929: @item
930: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
931: character under the cursor.
932: @end itemize
933:
934: When editing, displayable characters are inserted to the left of the
935: cursor position; the line is always in ``insert'' (as opposed to
936: ``overstrike'') mode.
937:
938: @cindex history file
939: @cindex @file{.gforth-history}
940: On Unix systems, the history file is @file{~/.gforth-history} by
941: default@footnote{i.e. it is stored in the user's home directory.}. You
942: can find out the name and location of your history file using:
943:
944: @example
945: history-file type \ Unix-class systems
946:
947: history-file type \ Other systems
948: history-dir type
949: @end example
950:
951: If you enter long definitions by hand, you can use a text editor to
952: paste them out of the history file into a Forth source file for reuse at
953: a later time.
954:
955: Gforth never trims the size of the history file, so you should do this
956: periodically, if necessary.
957:
958: @comment this is all defined in history.fs
959: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
960: @comment chosen?
961:
962:
963: @comment ----------------------------------------------
964: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
965: @section Environment variables
966: @cindex environment variables
967:
968: Gforth uses these environment variables:
969:
970: @itemize @bullet
971: @item
972: @cindex @code{GFORTHHIST} -- environment variable
973: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
974: open/create the history file, @file{.gforth-history}. Default:
975: @code{$HOME}.
976:
977: @item
978: @cindex @code{GFORTHPATH} -- environment variable
979: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
980: for Forth source-code files.
981:
982: @item
983: @cindex @code{LANG} -- environment variable
984: @code{LANG} -- see @code{LC_CTYPE}
985:
986: @item
987: @cindex @code{LC_ALL} -- environment variable
988: @code{LC_ALL} -- see @code{LC_CTYPE}
989:
990: @item
991: @cindex @code{LC_CTYPE} -- environment variable
992: @code{LC_CTYPE} -- If this variable contains ``UTF-8'' on Gforth
993: startup, Gforth uses the UTF-8 encoding for strings internally and
994: expects its input and produces its output in UTF-8 encoding, otherwise
995: the encoding is 8bit (see @pxref{Xchars and Unicode}). If this
996: environment variable is unset, Gforth looks in @code{LC_ALL}, and if
997: that is unset, in @code{LANG}.
998:
999: @item
1000: @cindex @code{GFORTHSYSTEMPREFIX} -- environment variable
1001:
1002: @code{GFORTHSYSTEMPREFIX} -- specifies what to prepend to the argument
1003: of @code{system} before passing it to C's @code{system()}. Default:
1004: @code{"./$COMSPEC /c "} on Windows, @code{""} on other OSs. The prefix
1005: and the command are directly concatenated, so if a space between them is
1006: necessary, append it to the prefix.
1007:
1008: @item
1009: @cindex @code{GFORTH} -- environment variable
1010: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1011:
1012: @item
1013: @cindex @code{GFORTHD} -- environment variable
1014: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1015:
1016: @item
1017: @cindex @code{TMP}, @code{TEMP} - environment variable
1018: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1019: location for the history file.
1020: @end itemize
1021:
1022: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1023: @comment mentioning these.
1024:
1025: All the Gforth environment variables default to sensible values if they
1026: are not set.
1027:
1028:
1029: @comment ----------------------------------------------
1030: @node Gforth Files, Gforth in pipes, Environment variables, Gforth Environment
1031: @section Gforth files
1032: @cindex Gforth files
1033:
1034: When you install Gforth on a Unix system, it installs files in these
1035: locations by default:
1036:
1037: @itemize @bullet
1038: @item
1039: @file{/usr/local/bin/gforth}
1040: @item
1041: @file{/usr/local/bin/gforthmi}
1042: @item
1043: @file{/usr/local/man/man1/gforth.1} - man page.
1044: @item
1045: @file{/usr/local/info} - the Info version of this manual.
1046: @item
1047: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1048: @item
1049: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1050: @item
1051: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1052: @item
1053: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1054: @end itemize
1055:
1056: You can select different places for installation by using
1057: @code{configure} options (listed with @code{configure --help}).
1058:
1059: @comment ----------------------------------------------
1060: @node Gforth in pipes, Startup speed, Gforth Files, Gforth Environment
1061: @section Gforth in pipes
1062: @cindex pipes, Gforth as part of
1063:
1064: Gforth can be used in pipes created elsewhere (described here). It can
1065: also create pipes on its own (@pxref{Pipes}).
1066:
1067: @cindex input from pipes
1068: If you pipe into Gforth, your program should read with @code{read-file}
1069: or @code{read-line} from @code{stdin} (@pxref{General files}).
1070: @code{Key} does not recognize the end of input. Words like
1071: @code{accept} echo the input and are therefore usually not useful for
1072: reading from a pipe. You have to invoke the Forth program with an OS
1073: command-line option, as you have no chance to use the Forth command line
1074: (the text interpreter would try to interpret the pipe input).
1075:
1076: @cindex output in pipes
1077: You can output to a pipe with @code{type}, @code{emit}, @code{cr} etc.
1078:
1079: @cindex silent exiting from Gforth
1080: When you write to a pipe that has been closed at the other end, Gforth
1081: receives a SIGPIPE signal (``pipe broken''). Gforth translates this
1082: into the exception @code{broken-pipe-error}. If your application does
1083: not catch that exception, the system catches it and exits, usually
1084: silently (unless you were working on the Forth command line; then it
1085: prints an error message and exits). This is usually the desired
1086: behaviour.
1087:
1088: If you do not like this behaviour, you have to catch the exception
1089: yourself, and react to it.
1090:
1091: Here's an example of an invocation of Gforth that is usable in a pipe:
1092:
1093: @example
1094: gforth -e ": foo begin pad dup 10 stdin read-file throw dup while \
1095: type repeat ; foo bye"
1096: @end example
1097:
1098: This example just copies the input verbatim to the output. A very
1099: simple pipe containing this example looks like this:
1100:
1101: @example
1102: cat startup.fs |
1103: gforth -e ": foo begin pad dup 80 stdin read-file throw dup while \
1104: type repeat ; foo bye"|
1105: head
1106: @end example
1107:
1108: @cindex stderr and pipes
1109: Pipes involving Gforth's @code{stderr} output do not work.
1110:
1111: @comment ----------------------------------------------
1112: @node Startup speed, , Gforth in pipes, Gforth Environment
1113: @section Startup speed
1114: @cindex Startup speed
1115: @cindex speed, startup
1116:
1117: If Gforth is used for CGI scripts or in shell scripts, its startup
1118: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1119: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1120: system time.
1121:
1122: If startup speed is a problem, you may consider the following ways to
1123: improve it; or you may consider ways to reduce the number of startups
1124: (for example, by using Fast-CGI).
1125:
1126: An easy step that influences Gforth startup speed is the use of the
1127: @option{--no-dynamic} option; this decreases image loading speed, but
1128: increases compile-time and run-time.
1129:
1130: Another step to improve startup speed is to statically link Gforth, by
1131: building it with @code{XLDFLAGS=-static}. This requires more memory for
1132: the code and will therefore slow down the first invocation, but
1133: subsequent invocations avoid the dynamic linking overhead. Another
1134: disadvantage is that Gforth won't profit from library upgrades. As a
1135: result, @code{gforth-static -e bye} takes about 17.1ms user and
1136: 8.2ms system time.
1137:
1138: The next step to improve startup speed is to use a non-relocatable image
1139: (@pxref{Non-Relocatable Image Files}). You can create this image with
1140: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1141: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1142: and a part of the copy-on-write overhead. The disadvantage is that the
1143: non-relocatable image does not work if the OS gives Gforth a different
1144: address for the dictionary, for whatever reason; so you better provide a
1145: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1146: bye} takes about 15.3ms user and 7.5ms system time.
1147:
1148: The final step is to disable dictionary hashing in Gforth. Gforth
1149: builds the hash table on startup, which takes much of the startup
1150: overhead. You can do this by commenting out the @code{include hash.fs}
1151: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1152: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1153: The disadvantages are that functionality like @code{table} and
1154: @code{ekey} is missing and that text interpretation (e.g., compiling)
1155: now takes much longer. So, you should only use this method if there is
1156: no significant text interpretation to perform (the script should be
1157: compiled into the image, amongst other things). @code{gforth-static -i
1158: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1159:
1160: @c ******************************************************************
1161: @node Tutorial, Introduction, Gforth Environment, Top
1162: @chapter Forth Tutorial
1163: @cindex Tutorial
1164: @cindex Forth Tutorial
1165:
1166: @c Topics from nac's Introduction that could be mentioned:
1167: @c press <ret> after each line
1168: @c Prompt
1169: @c numbers vs. words in dictionary on text interpretation
1170: @c what happens on redefinition
1171: @c parsing words (in particular, defining words)
1172:
1173: The difference of this chapter from the Introduction
1174: (@pxref{Introduction}) is that this tutorial is more fast-paced, should
1175: be used while sitting in front of a computer, and covers much more
1176: material, but does not explain how the Forth system works.
1177:
1178: This tutorial can be used with any ANS-compliant Forth; any
1179: Gforth-specific features are marked as such and you can skip them if you
1180: work with another Forth. This tutorial does not explain all features of
1181: Forth, just enough to get you started and give you some ideas about the
1182: facilities available in Forth. Read the rest of the manual and the
1183: standard when you are through this.
1184:
1185: The intended way to use this tutorial is that you work through it while
1186: sitting in front of the console, take a look at the examples and predict
1187: what they will do, then try them out; if the outcome is not as expected,
1188: find out why (e.g., by trying out variations of the example), so you
1189: understand what's going on. There are also some assignments that you
1190: should solve.
1191:
1192: This tutorial assumes that you have programmed before and know what,
1193: e.g., a loop is.
1194:
1195: @c !! explain compat library
1196:
1197: @menu
1198: * Starting Gforth Tutorial::
1199: * Syntax Tutorial::
1200: * Crash Course Tutorial::
1201: * Stack Tutorial::
1202: * Arithmetics Tutorial::
1203: * Stack Manipulation Tutorial::
1204: * Using files for Forth code Tutorial::
1205: * Comments Tutorial::
1206: * Colon Definitions Tutorial::
1207: * Decompilation Tutorial::
1208: * Stack-Effect Comments Tutorial::
1209: * Types Tutorial::
1210: * Factoring Tutorial::
1211: * Designing the stack effect Tutorial::
1212: * Local Variables Tutorial::
1213: * Conditional execution Tutorial::
1214: * Flags and Comparisons Tutorial::
1215: * General Loops Tutorial::
1216: * Counted loops Tutorial::
1217: * Recursion Tutorial::
1218: * Leaving definitions or loops Tutorial::
1219: * Return Stack Tutorial::
1220: * Memory Tutorial::
1221: * Characters and Strings Tutorial::
1222: * Alignment Tutorial::
1223: * Files Tutorial::
1224: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1225: * Execution Tokens Tutorial::
1226: * Exceptions Tutorial::
1227: * Defining Words Tutorial::
1228: * Arrays and Records Tutorial::
1229: * POSTPONE Tutorial::
1230: * Literal Tutorial::
1231: * Advanced macros Tutorial::
1232: * Compilation Tokens Tutorial::
1233: * Wordlists and Search Order Tutorial::
1234: @end menu
1235:
1236: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1237: @section Starting Gforth
1238: @cindex starting Gforth tutorial
1239: You can start Gforth by typing its name:
1240:
1241: @example
1242: gforth
1243: @end example
1244:
1245: That puts you into interactive mode; you can leave Gforth by typing
1246: @code{bye}. While in Gforth, you can edit the command line and access
1247: the command line history with cursor keys, similar to bash.
1248:
1249:
1250: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1251: @section Syntax
1252: @cindex syntax tutorial
1253:
1254: A @dfn{word} is a sequence of arbitrary characters (expcept white
1255: space). Words are separated by white space. E.g., each of the
1256: following lines contains exactly one word:
1257:
1258: @example
1259: word
1260: !@@#$%^&*()
1261: 1234567890
1262: 5!a
1263: @end example
1264:
1265: A frequent beginner's error is to leave away necessary white space,
1266: resulting in an error like @samp{Undefined word}; so if you see such an
1267: error, check if you have put spaces wherever necessary.
1268:
1269: @example
1270: ." hello, world" \ correct
1271: ."hello, world" \ gives an "Undefined word" error
1272: @end example
1273:
1274: Gforth and most other Forth systems ignore differences in case (they are
1275: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1276: your system is case-sensitive, you may have to type all the examples
1277: given here in upper case.
1278:
1279:
1280: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1281: @section Crash Course
1282:
1283: Type
1284:
1285: @example
1286: 0 0 !
1287: here execute
1288: ' catch >body 20 erase abort
1289: ' (quit) >body 20 erase
1290: @end example
1291:
1292: The last two examples are guaranteed to destroy parts of Gforth (and
1293: most other systems), so you better leave Gforth afterwards (if it has
1294: not finished by itself). On some systems you may have to kill gforth
1295: from outside (e.g., in Unix with @code{kill}).
1296:
1297: Now that you know how to produce crashes (and that there's not much to
1298: them), let's learn how to produce meaningful programs.
1299:
1300:
1301: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1302: @section Stack
1303: @cindex stack tutorial
1304:
1305: The most obvious feature of Forth is the stack. When you type in a
1306: number, it is pushed on the stack. You can display the content of the
1307: stack with @code{.s}.
1308:
1309: @example
1310: 1 2 .s
1311: 3 .s
1312: @end example
1313:
1314: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1315: appear in @code{.s} output as they appeared in the input.
1316:
1317: You can print the top of stack element with @code{.}.
1318:
1319: @example
1320: 1 2 3 . . .
1321: @end example
1322:
1323: In general, words consume their stack arguments (@code{.s} is an
1324: exception).
1325:
1326: @quotation Assignment
1327: What does the stack contain after @code{5 6 7 .}?
1328: @end quotation
1329:
1330:
1331: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1332: @section Arithmetics
1333: @cindex arithmetics tutorial
1334:
1335: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1336: operate on the top two stack items:
1337:
1338: @example
1339: 2 2 .s
1340: + .s
1341: .
1342: 2 1 - .
1343: 7 3 mod .
1344: @end example
1345:
1346: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1347: as in the corresponding infix expression (this is generally the case in
1348: Forth).
1349:
1350: Parentheses are superfluous (and not available), because the order of
1351: the words unambiguously determines the order of evaluation and the
1352: operands:
1353:
1354: @example
1355: 3 4 + 5 * .
1356: 3 4 5 * + .
1357: @end example
1358:
1359: @quotation Assignment
1360: What are the infix expressions corresponding to the Forth code above?
1361: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1362: known as Postfix or RPN (Reverse Polish Notation).}.
1363: @end quotation
1364:
1365: To change the sign, use @code{negate}:
1366:
1367: @example
1368: 2 negate .
1369: @end example
1370:
1371: @quotation Assignment
1372: Convert -(-3)*4-5 to Forth.
1373: @end quotation
1374:
1375: @code{/mod} performs both @code{/} and @code{mod}.
1376:
1377: @example
1378: 7 3 /mod . .
1379: @end example
1380:
1381: Reference: @ref{Arithmetic}.
1382:
1383:
1384: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1385: @section Stack Manipulation
1386: @cindex stack manipulation tutorial
1387:
1388: Stack manipulation words rearrange the data on the stack.
1389:
1390: @example
1391: 1 .s drop .s
1392: 1 .s dup .s drop drop .s
1393: 1 2 .s over .s drop drop drop
1394: 1 2 .s swap .s drop drop
1395: 1 2 3 .s rot .s drop drop drop
1396: @end example
1397:
1398: These are the most important stack manipulation words. There are also
1399: variants that manipulate twice as many stack items:
1400:
1401: @example
1402: 1 2 3 4 .s 2swap .s 2drop 2drop
1403: @end example
1404:
1405: Two more stack manipulation words are:
1406:
1407: @example
1408: 1 2 .s nip .s drop
1409: 1 2 .s tuck .s 2drop drop
1410: @end example
1411:
1412: @quotation Assignment
1413: Replace @code{nip} and @code{tuck} with combinations of other stack
1414: manipulation words.
1415:
1416: @example
1417: Given: How do you get:
1418: 1 2 3 3 2 1
1419: 1 2 3 1 2 3 2
1420: 1 2 3 1 2 3 3
1421: 1 2 3 1 3 3
1422: 1 2 3 2 1 3
1423: 1 2 3 4 4 3 2 1
1424: 1 2 3 1 2 3 1 2 3
1425: 1 2 3 4 1 2 3 4 1 2
1426: 1 2 3
1427: 1 2 3 1 2 3 4
1428: 1 2 3 1 3
1429: @end example
1430: @end quotation
1431:
1432: @example
1433: 5 dup * .
1434: @end example
1435:
1436: @quotation Assignment
1437: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1438: Write a piece of Forth code that expects two numbers on the stack
1439: (@var{a} and @var{b}, with @var{b} on top) and computes
1440: @code{(a-b)(a+1)}.
1441: @end quotation
1442:
1443: Reference: @ref{Stack Manipulation}.
1444:
1445:
1446: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1447: @section Using files for Forth code
1448: @cindex loading Forth code, tutorial
1449: @cindex files containing Forth code, tutorial
1450:
1451: While working at the Forth command line is convenient for one-line
1452: examples and short one-off code, you probably want to store your source
1453: code in files for convenient editing and persistence. You can use your
1454: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1455: Gforth}) to create @var{file.fs} and use
1456:
1457: @example
1458: s" @var{file.fs}" included
1459: @end example
1460:
1461: to load it into your Forth system. The file name extension I use for
1462: Forth files is @samp{.fs}.
1463:
1464: You can easily start Gforth with some files loaded like this:
1465:
1466: @example
1467: gforth @var{file1.fs} @var{file2.fs}
1468: @end example
1469:
1470: If an error occurs during loading these files, Gforth terminates,
1471: whereas an error during @code{INCLUDED} within Gforth usually gives you
1472: a Gforth command line. Starting the Forth system every time gives you a
1473: clean start every time, without interference from the results of earlier
1474: tries.
1475:
1476: I often put all the tests in a file, then load the code and run the
1477: tests with
1478:
1479: @example
1480: gforth @var{code.fs} @var{tests.fs} -e bye
1481: @end example
1482:
1483: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1484: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1485: restart this command without ado.
1486:
1487: The advantage of this approach is that the tests can be repeated easily
1488: every time the program ist changed, making it easy to catch bugs
1489: introduced by the change.
1490:
1491: Reference: @ref{Forth source files}.
1492:
1493:
1494: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1495: @section Comments
1496: @cindex comments tutorial
1497:
1498: @example
1499: \ That's a comment; it ends at the end of the line
1500: ( Another comment; it ends here: ) .s
1501: @end example
1502:
1503: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1504: separated with white space from the following text.
1505:
1506: @example
1507: \This gives an "Undefined word" error
1508: @end example
1509:
1510: The first @code{)} ends a comment started with @code{(}, so you cannot
1511: nest @code{(}-comments; and you cannot comment out text containing a
1512: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1513: avoid @code{)} in word names.}.
1514:
1515: I use @code{\}-comments for descriptive text and for commenting out code
1516: of one or more line; I use @code{(}-comments for describing the stack
1517: effect, the stack contents, or for commenting out sub-line pieces of
1518: code.
1519:
1520: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1521: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1522: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1523: with @kbd{M-q}.
1524:
1525: Reference: @ref{Comments}.
1526:
1527:
1528: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1529: @section Colon Definitions
1530: @cindex colon definitions, tutorial
1531: @cindex definitions, tutorial
1532: @cindex procedures, tutorial
1533: @cindex functions, tutorial
1534:
1535: are similar to procedures and functions in other programming languages.
1536:
1537: @example
1538: : squared ( n -- n^2 )
1539: dup * ;
1540: 5 squared .
1541: 7 squared .
1542: @end example
1543:
1544: @code{:} starts the colon definition; its name is @code{squared}. The
1545: following comment describes its stack effect. The words @code{dup *}
1546: are not executed, but compiled into the definition. @code{;} ends the
1547: colon definition.
1548:
1549: The newly-defined word can be used like any other word, including using
1550: it in other definitions:
1551:
1552: @example
1553: : cubed ( n -- n^3 )
1554: dup squared * ;
1555: -5 cubed .
1556: : fourth-power ( n -- n^4 )
1557: squared squared ;
1558: 3 fourth-power .
1559: @end example
1560:
1561: @quotation Assignment
1562: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1563: @code{/mod} in terms of other Forth words, and check if they work (hint:
1564: test your tests on the originals first). Don't let the
1565: @samp{redefined}-Messages spook you, they are just warnings.
1566: @end quotation
1567:
1568: Reference: @ref{Colon Definitions}.
1569:
1570:
1571: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1572: @section Decompilation
1573: @cindex decompilation tutorial
1574: @cindex see tutorial
1575:
1576: You can decompile colon definitions with @code{see}:
1577:
1578: @example
1579: see squared
1580: see cubed
1581: @end example
1582:
1583: In Gforth @code{see} shows you a reconstruction of the source code from
1584: the executable code. Informations that were present in the source, but
1585: not in the executable code, are lost (e.g., comments).
1586:
1587: You can also decompile the predefined words:
1588:
1589: @example
1590: see .
1591: see +
1592: @end example
1593:
1594:
1595: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1596: @section Stack-Effect Comments
1597: @cindex stack-effect comments, tutorial
1598: @cindex --, tutorial
1599: By convention the comment after the name of a definition describes the
1600: stack effect: The part in from of the @samp{--} describes the state of
1601: the stack before the execution of the definition, i.e., the parameters
1602: that are passed into the colon definition; the part behind the @samp{--}
1603: is the state of the stack after the execution of the definition, i.e.,
1604: the results of the definition. The stack comment only shows the top
1605: stack items that the definition accesses and/or changes.
1606:
1607: You should put a correct stack effect on every definition, even if it is
1608: just @code{( -- )}. You should also add some descriptive comment to
1609: more complicated words (I usually do this in the lines following
1610: @code{:}). If you don't do this, your code becomes unreadable (because
1611: you have to work through every definition before you can understand
1612: any).
1613:
1614: @quotation Assignment
1615: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1616: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1617: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1618: are done, you can compare your stack effects to those in this manual
1619: (@pxref{Word Index}).
1620: @end quotation
1621:
1622: Sometimes programmers put comments at various places in colon
1623: definitions that describe the contents of the stack at that place (stack
1624: comments); i.e., they are like the first part of a stack-effect
1625: comment. E.g.,
1626:
1627: @example
1628: : cubed ( n -- n^3 )
1629: dup squared ( n n^2 ) * ;
1630: @end example
1631:
1632: In this case the stack comment is pretty superfluous, because the word
1633: is simple enough. If you think it would be a good idea to add such a
1634: comment to increase readability, you should also consider factoring the
1635: word into several simpler words (@pxref{Factoring Tutorial,,
1636: Factoring}), which typically eliminates the need for the stack comment;
1637: however, if you decide not to refactor it, then having such a comment is
1638: better than not having it.
1639:
1640: The names of the stack items in stack-effect and stack comments in the
1641: standard, in this manual, and in many programs specify the type through
1642: a type prefix, similar to Fortran and Hungarian notation. The most
1643: frequent prefixes are:
1644:
1645: @table @code
1646: @item n
1647: signed integer
1648: @item u
1649: unsigned integer
1650: @item c
1651: character
1652: @item f
1653: Boolean flags, i.e. @code{false} or @code{true}.
1654: @item a-addr,a-
1655: Cell-aligned address
1656: @item c-addr,c-
1657: Char-aligned address (note that a Char may have two bytes in Windows NT)
1658: @item xt
1659: Execution token, same size as Cell
1660: @item w,x
1661: Cell, can contain an integer or an address. It usually takes 32, 64 or
1662: 16 bits (depending on your platform and Forth system). A cell is more
1663: commonly known as machine word, but the term @emph{word} already means
1664: something different in Forth.
1665: @item d
1666: signed double-cell integer
1667: @item ud
1668: unsigned double-cell integer
1669: @item r
1670: Float (on the FP stack)
1671: @end table
1672:
1673: You can find a more complete list in @ref{Notation}.
1674:
1675: @quotation Assignment
1676: Write stack-effect comments for all definitions you have written up to
1677: now.
1678: @end quotation
1679:
1680:
1681: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1682: @section Types
1683: @cindex types tutorial
1684:
1685: In Forth the names of the operations are not overloaded; so similar
1686: operations on different types need different names; e.g., @code{+} adds
1687: integers, and you have to use @code{f+} to add floating-point numbers.
1688: The following prefixes are often used for related operations on
1689: different types:
1690:
1691: @table @code
1692: @item (none)
1693: signed integer
1694: @item u
1695: unsigned integer
1696: @item c
1697: character
1698: @item d
1699: signed double-cell integer
1700: @item ud, du
1701: unsigned double-cell integer
1702: @item 2
1703: two cells (not-necessarily double-cell numbers)
1704: @item m, um
1705: mixed single-cell and double-cell operations
1706: @item f
1707: floating-point (note that in stack comments @samp{f} represents flags,
1708: and @samp{r} represents FP numbers).
1709: @end table
1710:
1711: If there are no differences between the signed and the unsigned variant
1712: (e.g., for @code{+}), there is only the prefix-less variant.
1713:
1714: Forth does not perform type checking, neither at compile time, nor at
1715: run time. If you use the wrong oeration, the data are interpreted
1716: incorrectly:
1717:
1718: @example
1719: -1 u.
1720: @end example
1721:
1722: If you have only experience with type-checked languages until now, and
1723: have heard how important type-checking is, don't panic! In my
1724: experience (and that of other Forthers), type errors in Forth code are
1725: usually easy to find (once you get used to it), the increased vigilance
1726: of the programmer tends to catch some harder errors in addition to most
1727: type errors, and you never have to work around the type system, so in
1728: most situations the lack of type-checking seems to be a win (projects to
1729: add type checking to Forth have not caught on).
1730:
1731:
1732: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1733: @section Factoring
1734: @cindex factoring tutorial
1735:
1736: If you try to write longer definitions, you will soon find it hard to
1737: keep track of the stack contents. Therefore, good Forth programmers
1738: tend to write only short definitions (e.g., three lines). The art of
1739: finding meaningful short definitions is known as factoring (as in
1740: factoring polynomials).
1741:
1742: Well-factored programs offer additional advantages: smaller, more
1743: general words, are easier to test and debug and can be reused more and
1744: better than larger, specialized words.
1745:
1746: So, if you run into difficulties with stack management, when writing
1747: code, try to define meaningful factors for the word, and define the word
1748: in terms of those. Even if a factor contains only two words, it is
1749: often helpful.
1750:
1751: Good factoring is not easy, and it takes some practice to get the knack
1752: for it; but even experienced Forth programmers often don't find the
1753: right solution right away, but only when rewriting the program. So, if
1754: you don't come up with a good solution immediately, keep trying, don't
1755: despair.
1756:
1757: @c example !!
1758:
1759:
1760: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
1761: @section Designing the stack effect
1762: @cindex Stack effect design, tutorial
1763: @cindex design of stack effects, tutorial
1764:
1765: In other languages you can use an arbitrary order of parameters for a
1766: function; and since there is only one result, you don't have to deal with
1767: the order of results, either.
1768:
1769: In Forth (and other stack-based languages, e.g., PostScript) the
1770: parameter and result order of a definition is important and should be
1771: designed well. The general guideline is to design the stack effect such
1772: that the word is simple to use in most cases, even if that complicates
1773: the implementation of the word. Some concrete rules are:
1774:
1775: @itemize @bullet
1776:
1777: @item
1778: Words consume all of their parameters (e.g., @code{.}).
1779:
1780: @item
1781: If there is a convention on the order of parameters (e.g., from
1782: mathematics or another programming language), stick with it (e.g.,
1783: @code{-}).
1784:
1785: @item
1786: If one parameter usually requires only a short computation (e.g., it is
1787: a constant), pass it on the top of the stack. Conversely, parameters
1788: that usually require a long sequence of code to compute should be passed
1789: as the bottom (i.e., first) parameter. This makes the code easier to
1790: read, because reader does not need to keep track of the bottom item
1791: through a long sequence of code (or, alternatively, through stack
1792: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
1793: address on top of the stack because it is usually simpler to compute
1794: than the stored value (often the address is just a variable).
1795:
1796: @item
1797: Similarly, results that are usually consumed quickly should be returned
1798: on the top of stack, whereas a result that is often used in long
1799: computations should be passed as bottom result. E.g., the file words
1800: like @code{open-file} return the error code on the top of stack, because
1801: it is usually consumed quickly by @code{throw}; moreover, the error code
1802: has to be checked before doing anything with the other results.
1803:
1804: @end itemize
1805:
1806: These rules are just general guidelines, don't lose sight of the overall
1807: goal to make the words easy to use. E.g., if the convention rule
1808: conflicts with the computation-length rule, you might decide in favour
1809: of the convention if the word will be used rarely, and in favour of the
1810: computation-length rule if the word will be used frequently (because
1811: with frequent use the cost of breaking the computation-length rule would
1812: be quite high, and frequent use makes it easier to remember an
1813: unconventional order).
1814:
1815: @c example !! structure package
1816:
1817:
1818: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
1819: @section Local Variables
1820: @cindex local variables, tutorial
1821:
1822: You can define local variables (@emph{locals}) in a colon definition:
1823:
1824: @example
1825: : swap @{ a b -- b a @}
1826: b a ;
1827: 1 2 swap .s 2drop
1828: @end example
1829:
1830: (If your Forth system does not support this syntax, include
1831: @file{compat/anslocals.fs} first).
1832:
1833: In this example @code{@{ a b -- b a @}} is the locals definition; it
1834: takes two cells from the stack, puts the top of stack in @code{b} and
1835: the next stack element in @code{a}. @code{--} starts a comment ending
1836: with @code{@}}. After the locals definition, using the name of the
1837: local will push its value on the stack. You can leave the comment
1838: part (@code{-- b a}) away:
1839:
1840: @example
1841: : swap ( x1 x2 -- x2 x1 )
1842: @{ a b @} b a ;
1843: @end example
1844:
1845: In Gforth you can have several locals definitions, anywhere in a colon
1846: definition; in contrast, in a standard program you can have only one
1847: locals definition per colon definition, and that locals definition must
1848: be outside any control structure.
1849:
1850: With locals you can write slightly longer definitions without running
1851: into stack trouble. However, I recommend trying to write colon
1852: definitions without locals for exercise purposes to help you gain the
1853: essential factoring skills.
1854:
1855: @quotation Assignment
1856: Rewrite your definitions until now with locals
1857: @end quotation
1858:
1859: Reference: @ref{Locals}.
1860:
1861:
1862: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
1863: @section Conditional execution
1864: @cindex conditionals, tutorial
1865: @cindex if, tutorial
1866:
1867: In Forth you can use control structures only inside colon definitions.
1868: An @code{if}-structure looks like this:
1869:
1870: @example
1871: : abs ( n1 -- +n2 )
1872: dup 0 < if
1873: negate
1874: endif ;
1875: 5 abs .
1876: -5 abs .
1877: @end example
1878:
1879: @code{if} takes a flag from the stack. If the flag is non-zero (true),
1880: the following code is performed, otherwise execution continues after the
1881: @code{endif} (or @code{else}). @code{<} compares the top two stack
1882: elements and prioduces a flag:
1883:
1884: @example
1885: 1 2 < .
1886: 2 1 < .
1887: 1 1 < .
1888: @end example
1889:
1890: Actually the standard name for @code{endif} is @code{then}. This
1891: tutorial presents the examples using @code{endif}, because this is often
1892: less confusing for people familiar with other programming languages
1893: where @code{then} has a different meaning. If your system does not have
1894: @code{endif}, define it with
1895:
1896: @example
1897: : endif postpone then ; immediate
1898: @end example
1899:
1900: You can optionally use an @code{else}-part:
1901:
1902: @example
1903: : min ( n1 n2 -- n )
1904: 2dup < if
1905: drop
1906: else
1907: nip
1908: endif ;
1909: 2 3 min .
1910: 3 2 min .
1911: @end example
1912:
1913: @quotation Assignment
1914: Write @code{min} without @code{else}-part (hint: what's the definition
1915: of @code{nip}?).
1916: @end quotation
1917:
1918: Reference: @ref{Selection}.
1919:
1920:
1921: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
1922: @section Flags and Comparisons
1923: @cindex flags tutorial
1924: @cindex comparison tutorial
1925:
1926: In a false-flag all bits are clear (0 when interpreted as integer). In
1927: a canonical true-flag all bits are set (-1 as a twos-complement signed
1928: integer); in many contexts (e.g., @code{if}) any non-zero value is
1929: treated as true flag.
1930:
1931: @example
1932: false .
1933: true .
1934: true hex u. decimal
1935: @end example
1936:
1937: Comparison words produce canonical flags:
1938:
1939: @example
1940: 1 1 = .
1941: 1 0= .
1942: 0 1 < .
1943: 0 0 < .
1944: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
1945: -1 1 < .
1946: @end example
1947:
1948: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
1949: (or none) and the comparisons @code{= <> < > <= >=}. Only a part of
1950: these combinations are standard (for details see the standard,
1951: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
1952:
1953: You can use @code{and or xor invert} can be used as operations on
1954: canonical flags. Actually they are bitwise operations:
1955:
1956: @example
1957: 1 2 and .
1958: 1 2 or .
1959: 1 3 xor .
1960: 1 invert .
1961: @end example
1962:
1963: You can convert a zero/non-zero flag into a canonical flag with
1964: @code{0<>} (and complement it on the way with @code{0=}).
1965:
1966: @example
1967: 1 0= .
1968: 1 0<> .
1969: @end example
1970:
1971: You can use the all-bits-set feature of canonical flags and the bitwise
1972: operation of the Boolean operations to avoid @code{if}s:
1973:
1974: @example
1975: : foo ( n1 -- n2 )
1976: 0= if
1977: 14
1978: else
1979: 0
1980: endif ;
1981: 0 foo .
1982: 1 foo .
1983:
1984: : foo ( n1 -- n2 )
1985: 0= 14 and ;
1986: 0 foo .
1987: 1 foo .
1988: @end example
1989:
1990: @quotation Assignment
1991: Write @code{min} without @code{if}.
1992: @end quotation
1993:
1994: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
1995: @ref{Bitwise operations}.
1996:
1997:
1998: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
1999: @section General Loops
2000: @cindex loops, indefinite, tutorial
2001:
2002: The endless loop is the most simple one:
2003:
2004: @example
2005: : endless ( -- )
2006: 0 begin
2007: dup . 1+
2008: again ;
2009: endless
2010: @end example
2011:
2012: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2013: does nothing at run-time, @code{again} jumps back to @code{begin}.
2014:
2015: A loop with one exit at any place looks like this:
2016:
2017: @example
2018: : log2 ( +n1 -- n2 )
2019: \ logarithmus dualis of n1>0, rounded down to the next integer
2020: assert( dup 0> )
2021: 2/ 0 begin
2022: over 0> while
2023: 1+ swap 2/ swap
2024: repeat
2025: nip ;
2026: 7 log2 .
2027: 8 log2 .
2028: @end example
2029:
2030: At run-time @code{while} consumes a flag; if it is 0, execution
2031: continues behind the @code{repeat}; if the flag is non-zero, execution
2032: continues behind the @code{while}. @code{Repeat} jumps back to
2033: @code{begin}, just like @code{again}.
2034:
2035: In Forth there are many combinations/abbreviations, like @code{1+}.
2036: However, @code{2/} is not one of them; it shifts its argument right by
2037: one bit (arithmetic shift right):
2038:
2039: @example
2040: -5 2 / .
2041: -5 2/ .
2042: @end example
2043:
2044: @code{assert(} is no standard word, but you can get it on systems other
2045: then Gforth by including @file{compat/assert.fs}. You can see what it
2046: does by trying
2047:
2048: @example
2049: 0 log2 .
2050: @end example
2051:
2052: Here's a loop with an exit at the end:
2053:
2054: @example
2055: : log2 ( +n1 -- n2 )
2056: \ logarithmus dualis of n1>0, rounded down to the next integer
2057: assert( dup 0 > )
2058: -1 begin
2059: 1+ swap 2/ swap
2060: over 0 <=
2061: until
2062: nip ;
2063: @end example
2064:
2065: @code{Until} consumes a flag; if it is non-zero, execution continues at
2066: the @code{begin}, otherwise after the @code{until}.
2067:
2068: @quotation Assignment
2069: Write a definition for computing the greatest common divisor.
2070: @end quotation
2071:
2072: Reference: @ref{Simple Loops}.
2073:
2074:
2075: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2076: @section Counted loops
2077: @cindex loops, counted, tutorial
2078:
2079: @example
2080: : ^ ( n1 u -- n )
2081: \ n = the uth power of u1
2082: 1 swap 0 u+do
2083: over *
2084: loop
2085: nip ;
2086: 3 2 ^ .
2087: 4 3 ^ .
2088: @end example
2089:
2090: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2091: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2092: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2093: times (or not at all, if @code{u3-u4<0}).
2094:
2095: You can see the stack effect design rules at work in the stack effect of
2096: the loop start words: Since the start value of the loop is more
2097: frequently constant than the end value, the start value is passed on
2098: the top-of-stack.
2099:
2100: You can access the counter of a counted loop with @code{i}:
2101:
2102: @example
2103: : fac ( u -- u! )
2104: 1 swap 1+ 1 u+do
2105: i *
2106: loop ;
2107: 5 fac .
2108: 7 fac .
2109: @end example
2110:
2111: There is also @code{+do}, which expects signed numbers (important for
2112: deciding whether to enter the loop).
2113:
2114: @quotation Assignment
2115: Write a definition for computing the nth Fibonacci number.
2116: @end quotation
2117:
2118: You can also use increments other than 1:
2119:
2120: @example
2121: : up2 ( n1 n2 -- )
2122: +do
2123: i .
2124: 2 +loop ;
2125: 10 0 up2
2126:
2127: : down2 ( n1 n2 -- )
2128: -do
2129: i .
2130: 2 -loop ;
2131: 0 10 down2
2132: @end example
2133:
2134: Reference: @ref{Counted Loops}.
2135:
2136:
2137: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2138: @section Recursion
2139: @cindex recursion tutorial
2140:
2141: Usually the name of a definition is not visible in the definition; but
2142: earlier definitions are usually visible:
2143:
2144: @example
2145: 1 0 / . \ "Floating-point unidentified fault" in Gforth on some platforms
2146: : / ( n1 n2 -- n )
2147: dup 0= if
2148: -10 throw \ report division by zero
2149: endif
2150: / \ old version
2151: ;
2152: 1 0 /
2153: @end example
2154:
2155: For recursive definitions you can use @code{recursive} (non-standard) or
2156: @code{recurse}:
2157:
2158: @example
2159: : fac1 ( n -- n! ) recursive
2160: dup 0> if
2161: dup 1- fac1 *
2162: else
2163: drop 1
2164: endif ;
2165: 7 fac1 .
2166:
2167: : fac2 ( n -- n! )
2168: dup 0> if
2169: dup 1- recurse *
2170: else
2171: drop 1
2172: endif ;
2173: 8 fac2 .
2174: @end example
2175:
2176: @quotation Assignment
2177: Write a recursive definition for computing the nth Fibonacci number.
2178: @end quotation
2179:
2180: Reference (including indirect recursion): @xref{Calls and returns}.
2181:
2182:
2183: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2184: @section Leaving definitions or loops
2185: @cindex leaving definitions, tutorial
2186: @cindex leaving loops, tutorial
2187:
2188: @code{EXIT} exits the current definition right away. For every counted
2189: loop that is left in this way, an @code{UNLOOP} has to be performed
2190: before the @code{EXIT}:
2191:
2192: @c !! real examples
2193: @example
2194: : ...
2195: ... u+do
2196: ... if
2197: ... unloop exit
2198: endif
2199: ...
2200: loop
2201: ... ;
2202: @end example
2203:
2204: @code{LEAVE} leaves the innermost counted loop right away:
2205:
2206: @example
2207: : ...
2208: ... u+do
2209: ... if
2210: ... leave
2211: endif
2212: ...
2213: loop
2214: ... ;
2215: @end example
2216:
2217: @c !! example
2218:
2219: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
2220:
2221:
2222: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2223: @section Return Stack
2224: @cindex return stack tutorial
2225:
2226: In addition to the data stack Forth also has a second stack, the return
2227: stack; most Forth systems store the return addresses of procedure calls
2228: there (thus its name). Programmers can also use this stack:
2229:
2230: @example
2231: : foo ( n1 n2 -- )
2232: .s
2233: >r .s
2234: r@@ .
2235: >r .s
2236: r@@ .
2237: r> .
2238: r@@ .
2239: r> . ;
2240: 1 2 foo
2241: @end example
2242:
2243: @code{>r} takes an element from the data stack and pushes it onto the
2244: return stack; conversely, @code{r>} moves an elementm from the return to
2245: the data stack; @code{r@@} pushes a copy of the top of the return stack
2246: on the data stack.
2247:
2248: Forth programmers usually use the return stack for storing data
2249: temporarily, if using the data stack alone would be too complex, and
2250: factoring and locals are not an option:
2251:
2252: @example
2253: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2254: rot >r rot r> ;
2255: @end example
2256:
2257: The return address of the definition and the loop control parameters of
2258: counted loops usually reside on the return stack, so you have to take
2259: all items, that you have pushed on the return stack in a colon
2260: definition or counted loop, from the return stack before the definition
2261: or loop ends. You cannot access items that you pushed on the return
2262: stack outside some definition or loop within the definition of loop.
2263:
2264: If you miscount the return stack items, this usually ends in a crash:
2265:
2266: @example
2267: : crash ( n -- )
2268: >r ;
2269: 5 crash
2270: @end example
2271:
2272: You cannot mix using locals and using the return stack (according to the
2273: standard; Gforth has no problem). However, they solve the same
2274: problems, so this shouldn't be an issue.
2275:
2276: @quotation Assignment
2277: Can you rewrite any of the definitions you wrote until now in a better
2278: way using the return stack?
2279: @end quotation
2280:
2281: Reference: @ref{Return stack}.
2282:
2283:
2284: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2285: @section Memory
2286: @cindex memory access/allocation tutorial
2287:
2288: You can create a global variable @code{v} with
2289:
2290: @example
2291: variable v ( -- addr )
2292: @end example
2293:
2294: @code{v} pushes the address of a cell in memory on the stack. This cell
2295: was reserved by @code{variable}. You can use @code{!} (store) to store
2296: values into this cell and @code{@@} (fetch) to load the value from the
2297: stack into memory:
2298:
2299: @example
2300: v .
2301: 5 v ! .s
2302: v @@ .
2303: @end example
2304:
2305: You can see a raw dump of memory with @code{dump}:
2306:
2307: @example
2308: v 1 cells .s dump
2309: @end example
2310:
2311: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2312: generally, address units (aus)) that @code{n1 cells} occupy. You can
2313: also reserve more memory:
2314:
2315: @example
2316: create v2 20 cells allot
2317: v2 20 cells dump
2318: @end example
2319:
2320: creates a word @code{v2} and reserves 20 uninitialized cells; the
2321: address pushed by @code{v2} points to the start of these 20 cells. You
2322: can use address arithmetic to access these cells:
2323:
2324: @example
2325: 3 v2 5 cells + !
2326: v2 20 cells dump
2327: @end example
2328:
2329: You can reserve and initialize memory with @code{,}:
2330:
2331: @example
2332: create v3
2333: 5 , 4 , 3 , 2 , 1 ,
2334: v3 @@ .
2335: v3 cell+ @@ .
2336: v3 2 cells + @@ .
2337: v3 5 cells dump
2338: @end example
2339:
2340: @quotation Assignment
2341: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2342: @code{u} cells, with the first of these cells at @code{addr}, the next
2343: one at @code{addr cell+} etc.
2344: @end quotation
2345:
2346: You can also reserve memory without creating a new word:
2347:
2348: @example
2349: here 10 cells allot .
2350: here .
2351: @end example
2352:
2353: @code{Here} pushes the start address of the memory area. You should
2354: store it somewhere, or you will have a hard time finding the memory area
2355: again.
2356:
2357: @code{Allot} manages dictionary memory. The dictionary memory contains
2358: the system's data structures for words etc. on Gforth and most other
2359: Forth systems. It is managed like a stack: You can free the memory that
2360: you have just @code{allot}ed with
2361:
2362: @example
2363: -10 cells allot
2364: here .
2365: @end example
2366:
2367: Note that you cannot do this if you have created a new word in the
2368: meantime (because then your @code{allot}ed memory is no longer on the
2369: top of the dictionary ``stack'').
2370:
2371: Alternatively, you can use @code{allocate} and @code{free} which allow
2372: freeing memory in any order:
2373:
2374: @example
2375: 10 cells allocate throw .s
2376: 20 cells allocate throw .s
2377: swap
2378: free throw
2379: free throw
2380: @end example
2381:
2382: The @code{throw}s deal with errors (e.g., out of memory).
2383:
2384: And there is also a
2385: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2386: garbage collector}, which eliminates the need to @code{free} memory
2387: explicitly.
2388:
2389: Reference: @ref{Memory}.
2390:
2391:
2392: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2393: @section Characters and Strings
2394: @cindex strings tutorial
2395: @cindex characters tutorial
2396:
2397: On the stack characters take up a cell, like numbers. In memory they
2398: have their own size (one 8-bit byte on most systems), and therefore
2399: require their own words for memory access:
2400:
2401: @example
2402: create v4
2403: 104 c, 97 c, 108 c, 108 c, 111 c,
2404: v4 4 chars + c@@ .
2405: v4 5 chars dump
2406: @end example
2407:
2408: The preferred representation of strings on the stack is @code{addr
2409: u-count}, where @code{addr} is the address of the first character and
2410: @code{u-count} is the number of characters in the string.
2411:
2412: @example
2413: v4 5 type
2414: @end example
2415:
2416: You get a string constant with
2417:
2418: @example
2419: s" hello, world" .s
2420: type
2421: @end example
2422:
2423: Make sure you have a space between @code{s"} and the string; @code{s"}
2424: is a normal Forth word and must be delimited with white space (try what
2425: happens when you remove the space).
2426:
2427: However, this interpretive use of @code{s"} is quite restricted: the
2428: string exists only until the next call of @code{s"} (some Forth systems
2429: keep more than one of these strings, but usually they still have a
2430: limited lifetime).
2431:
2432: @example
2433: s" hello," s" world" .s
2434: type
2435: type
2436: @end example
2437:
2438: You can also use @code{s"} in a definition, and the resulting
2439: strings then live forever (well, for as long as the definition):
2440:
2441: @example
2442: : foo s" hello," s" world" ;
2443: foo .s
2444: type
2445: type
2446: @end example
2447:
2448: @quotation Assignment
2449: @code{Emit ( c -- )} types @code{c} as character (not a number).
2450: Implement @code{type ( addr u -- )}.
2451: @end quotation
2452:
2453: Reference: @ref{Memory Blocks}.
2454:
2455:
2456: @node Alignment Tutorial, Files Tutorial, Characters and Strings Tutorial, Tutorial
2457: @section Alignment
2458: @cindex alignment tutorial
2459: @cindex memory alignment tutorial
2460:
2461: On many processors cells have to be aligned in memory, if you want to
2462: access them with @code{@@} and @code{!} (and even if the processor does
2463: not require alignment, access to aligned cells is faster).
2464:
2465: @code{Create} aligns @code{here} (i.e., the place where the next
2466: allocation will occur, and that the @code{create}d word points to).
2467: Likewise, the memory produced by @code{allocate} starts at an aligned
2468: address. Adding a number of @code{cells} to an aligned address produces
2469: another aligned address.
2470:
2471: However, address arithmetic involving @code{char+} and @code{chars} can
2472: create an address that is not cell-aligned. @code{Aligned ( addr --
2473: a-addr )} produces the next aligned address:
2474:
2475: @example
2476: v3 char+ aligned .s @@ .
2477: v3 char+ .s @@ .
2478: @end example
2479:
2480: Similarly, @code{align} advances @code{here} to the next aligned
2481: address:
2482:
2483: @example
2484: create v5 97 c,
2485: here .
2486: align here .
2487: 1000 ,
2488: @end example
2489:
2490: Note that you should use aligned addresses even if your processor does
2491: not require them, if you want your program to be portable.
2492:
2493: Reference: @ref{Address arithmetic}.
2494:
2495:
2496: @node Files Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Alignment Tutorial, Tutorial
2497: @section Files
2498: @cindex files tutorial
2499:
2500: This section gives a short introduction into how to use files inside
2501: Forth. It's broken up into five easy steps:
2502:
2503: @enumerate 1
2504: @item Opened an ASCII text file for input
2505: @item Opened a file for output
2506: @item Read input file until string matched (or some other condition matched)
2507: @item Wrote some lines from input ( modified or not) to output
2508: @item Closed the files.
2509: @end enumerate
2510:
2511: Reference: @ref{General files}.
2512:
2513: @subsection Open file for input
2514:
2515: @example
2516: s" foo.in" r/o open-file throw Value fd-in
2517: @end example
2518:
2519: @subsection Create file for output
2520:
2521: @example
2522: s" foo.out" w/o create-file throw Value fd-out
2523: @end example
2524:
2525: The available file modes are r/o for read-only access, r/w for
2526: read-write access, and w/o for write-only access. You could open both
2527: files with r/w, too, if you like. All file words return error codes; for
2528: most applications, it's best to pass there error codes with @code{throw}
2529: to the outer error handler.
2530:
2531: If you want words for opening and assigning, define them as follows:
2532:
2533: @example
2534: 0 Value fd-in
2535: 0 Value fd-out
2536: : open-input ( addr u -- ) r/o open-file throw to fd-in ;
2537: : open-output ( addr u -- ) w/o create-file throw to fd-out ;
2538: @end example
2539:
2540: Usage example:
2541:
2542: @example
2543: s" foo.in" open-input
2544: s" foo.out" open-output
2545: @end example
2546:
2547: @subsection Scan file for a particular line
2548:
2549: @example
2550: 256 Constant max-line
2551: Create line-buffer max-line 2 + allot
2552:
2553: : scan-file ( addr u -- )
2554: begin
2555: line-buffer max-line fd-in read-line throw
2556: while
2557: >r 2dup line-buffer r> compare 0=
2558: until
2559: else
2560: drop
2561: then
2562: 2drop ;
2563: @end example
2564:
2565: @code{read-line ( addr u1 fd -- u2 flag ior )} reads up to u1 bytes into
2566: the buffer at addr, and returns the number of bytes read, a flag that is
2567: false when the end of file is reached, and an error code.
2568:
2569: @code{compare ( addr1 u1 addr2 u2 -- n )} compares two strings and
2570: returns zero if both strings are equal. It returns a positive number if
2571: the first string is lexically greater, a negative if the second string
2572: is lexically greater.
2573:
2574: We haven't seen this loop here; it has two exits. Since the @code{while}
2575: exits with the number of bytes read on the stack, we have to clean up
2576: that separately; that's after the @code{else}.
2577:
2578: Usage example:
2579:
2580: @example
2581: s" The text I search is here" scan-file
2582: @end example
2583:
2584: @subsection Copy input to output
2585:
2586: @example
2587: : copy-file ( -- )
2588: begin
2589: line-buffer max-line fd-in read-line throw
2590: while
2591: line-buffer swap fd-out write-file throw
2592: repeat ;
2593: @end example
2594:
2595: @subsection Close files
2596:
2597: @example
2598: fd-in close-file throw
2599: fd-out close-file throw
2600: @end example
2601:
2602: Likewise, you can put that into definitions, too:
2603:
2604: @example
2605: : close-input ( -- ) fd-in close-file throw ;
2606: : close-output ( -- ) fd-out close-file throw ;
2607: @end example
2608:
2609: @quotation Assignment
2610: How could you modify @code{copy-file} so that it copies until a second line is
2611: matched? Can you write a program that extracts a section of a text file,
2612: given the line that starts and the line that terminates that section?
2613: @end quotation
2614:
2615: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Files Tutorial, Tutorial
2616: @section Interpretation and Compilation Semantics and Immediacy
2617: @cindex semantics tutorial
2618: @cindex interpretation semantics tutorial
2619: @cindex compilation semantics tutorial
2620: @cindex immediate, tutorial
2621:
2622: When a word is compiled, it behaves differently from being interpreted.
2623: E.g., consider @code{+}:
2624:
2625: @example
2626: 1 2 + .
2627: : foo + ;
2628: @end example
2629:
2630: These two behaviours are known as compilation and interpretation
2631: semantics. For normal words (e.g., @code{+}), the compilation semantics
2632: is to append the interpretation semantics to the currently defined word
2633: (@code{foo} in the example above). I.e., when @code{foo} is executed
2634: later, the interpretation semantics of @code{+} (i.e., adding two
2635: numbers) will be performed.
2636:
2637: However, there are words with non-default compilation semantics, e.g.,
2638: the control-flow words like @code{if}. You can use @code{immediate} to
2639: change the compilation semantics of the last defined word to be equal to
2640: the interpretation semantics:
2641:
2642: @example
2643: : [FOO] ( -- )
2644: 5 . ; immediate
2645:
2646: [FOO]
2647: : bar ( -- )
2648: [FOO] ;
2649: bar
2650: see bar
2651: @end example
2652:
2653: Two conventions to mark words with non-default compilation semnatics are
2654: names with brackets (more frequently used) and to write them all in
2655: upper case (less frequently used).
2656:
2657: In Gforth (and many other systems) you can also remove the
2658: interpretation semantics with @code{compile-only} (the compilation
2659: semantics is derived from the original interpretation semantics):
2660:
2661: @example
2662: : flip ( -- )
2663: 6 . ; compile-only \ but not immediate
2664: flip
2665:
2666: : flop ( -- )
2667: flip ;
2668: flop
2669: @end example
2670:
2671: In this example the interpretation semantics of @code{flop} is equal to
2672: the original interpretation semantics of @code{flip}.
2673:
2674: The text interpreter has two states: in interpret state, it performs the
2675: interpretation semantics of words it encounters; in compile state, it
2676: performs the compilation semantics of these words.
2677:
2678: Among other things, @code{:} switches into compile state, and @code{;}
2679: switches back to interpret state. They contain the factors @code{]}
2680: (switch to compile state) and @code{[} (switch to interpret state), that
2681: do nothing but switch the state.
2682:
2683: @example
2684: : xxx ( -- )
2685: [ 5 . ]
2686: ;
2687:
2688: xxx
2689: see xxx
2690: @end example
2691:
2692: These brackets are also the source of the naming convention mentioned
2693: above.
2694:
2695: Reference: @ref{Interpretation and Compilation Semantics}.
2696:
2697:
2698: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2699: @section Execution Tokens
2700: @cindex execution tokens tutorial
2701: @cindex XT tutorial
2702:
2703: @code{' word} gives you the execution token (XT) of a word. The XT is a
2704: cell representing the interpretation semantics of a word. You can
2705: execute this semantics with @code{execute}:
2706:
2707: @example
2708: ' + .s
2709: 1 2 rot execute .
2710: @end example
2711:
2712: The XT is similar to a function pointer in C. However, parameter
2713: passing through the stack makes it a little more flexible:
2714:
2715: @example
2716: : map-array ( ... addr u xt -- ... )
2717: \ executes xt ( ... x -- ... ) for every element of the array starting
2718: \ at addr and containing u elements
2719: @{ xt @}
2720: cells over + swap ?do
2721: i @@ xt execute
2722: 1 cells +loop ;
2723:
2724: create a 3 , 4 , 2 , -1 , 4 ,
2725: a 5 ' . map-array .s
2726: 0 a 5 ' + map-array .
2727: s" max-n" environment? drop .s
2728: a 5 ' min map-array .
2729: @end example
2730:
2731: You can use map-array with the XTs of words that consume one element
2732: more than they produce. In theory you can also use it with other XTs,
2733: but the stack effect then depends on the size of the array, which is
2734: hard to understand.
2735:
2736: Since XTs are cell-sized, you can store them in memory and manipulate
2737: them on the stack like other cells. You can also compile the XT into a
2738: word with @code{compile,}:
2739:
2740: @example
2741: : foo1 ( n1 n2 -- n )
2742: [ ' + compile, ] ;
2743: see foo
2744: @end example
2745:
2746: This is non-standard, because @code{compile,} has no compilation
2747: semantics in the standard, but it works in good Forth systems. For the
2748: broken ones, use
2749:
2750: @example
2751: : [compile,] compile, ; immediate
2752:
2753: : foo1 ( n1 n2 -- n )
2754: [ ' + ] [compile,] ;
2755: see foo
2756: @end example
2757:
2758: @code{'} is a word with default compilation semantics; it parses the
2759: next word when its interpretation semantics are executed, not during
2760: compilation:
2761:
2762: @example
2763: : foo ( -- xt )
2764: ' ;
2765: see foo
2766: : bar ( ... "word" -- ... )
2767: ' execute ;
2768: see bar
2769: 1 2 bar + .
2770: @end example
2771:
2772: You often want to parse a word during compilation and compile its XT so
2773: it will be pushed on the stack at run-time. @code{[']} does this:
2774:
2775: @example
2776: : xt-+ ( -- xt )
2777: ['] + ;
2778: see xt-+
2779: 1 2 xt-+ execute .
2780: @end example
2781:
2782: Many programmers tend to see @code{'} and the word it parses as one
2783: unit, and expect it to behave like @code{[']} when compiled, and are
2784: confused by the actual behaviour. If you are, just remember that the
2785: Forth system just takes @code{'} as one unit and has no idea that it is
2786: a parsing word (attempts to convenience programmers in this issue have
2787: usually resulted in even worse pitfalls, see
2788: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
2789: @code{State}-smartness---Why it is evil and How to Exorcise it}).
2790:
2791: Note that the state of the interpreter does not come into play when
2792: creating and executing XTs. I.e., even when you execute @code{'} in
2793: compile state, it still gives you the interpretation semantics. And
2794: whatever that state is, @code{execute} performs the semantics
2795: represented by the XT (i.e., for XTs produced with @code{'} the
2796: interpretation semantics).
2797:
2798: Reference: @ref{Tokens for Words}.
2799:
2800:
2801: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
2802: @section Exceptions
2803: @cindex exceptions tutorial
2804:
2805: @code{throw ( n -- )} causes an exception unless n is zero.
2806:
2807: @example
2808: 100 throw .s
2809: 0 throw .s
2810: @end example
2811:
2812: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
2813: it catches exceptions and pushes the number of the exception on the
2814: stack (or 0, if the xt executed without exception). If there was an
2815: exception, the stacks have the same depth as when entering @code{catch}:
2816:
2817: @example
2818: .s
2819: 3 0 ' / catch .s
2820: 3 2 ' / catch .s
2821: @end example
2822:
2823: @quotation Assignment
2824: Try the same with @code{execute} instead of @code{catch}.
2825: @end quotation
2826:
2827: @code{Throw} always jumps to the dynamically next enclosing
2828: @code{catch}, even if it has to leave several call levels to achieve
2829: this:
2830:
2831: @example
2832: : foo 100 throw ;
2833: : foo1 foo ." after foo" ;
2834: : bar ['] foo1 catch ;
2835: bar .
2836: @end example
2837:
2838: It is often important to restore a value upon leaving a definition, even
2839: if the definition is left through an exception. You can ensure this
2840: like this:
2841:
2842: @example
2843: : ...
2844: save-x
2845: ['] word-changing-x catch ( ... n )
2846: restore-x
2847: ( ... n ) throw ;
2848: @end example
2849:
2850: Gforth provides an alternative syntax in addition to @code{catch}:
2851: @code{try ... recover ... endtry}. If the code between @code{try} and
2852: @code{recover} has an exception, the stack depths are restored, the
2853: exception number is pushed on the stack, and the code between
2854: @code{recover} and @code{endtry} is performed. E.g., the definition for
2855: @code{catch} is
2856:
2857: @example
2858: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
2859: try
2860: execute 0
2861: recover
2862: nip
2863: endtry ;
2864: @end example
2865:
2866: The equivalent to the restoration code above is
2867:
2868: @example
2869: : ...
2870: save-x
2871: try
2872: word-changing-x 0
2873: recover endtry
2874: restore-x
2875: throw ;
2876: @end example
2877:
2878: This works if @code{word-changing-x} does not change the stack depth,
2879: otherwise you should add some code between @code{recover} and
2880: @code{endtry} to balance the stack.
2881:
2882: Reference: @ref{Exception Handling}.
2883:
2884:
2885: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
2886: @section Defining Words
2887: @cindex defining words tutorial
2888: @cindex does> tutorial
2889: @cindex create...does> tutorial
2890:
2891: @c before semantics?
2892:
2893: @code{:}, @code{create}, and @code{variable} are definition words: They
2894: define other words. @code{Constant} is another definition word:
2895:
2896: @example
2897: 5 constant foo
2898: foo .
2899: @end example
2900:
2901: You can also use the prefixes @code{2} (double-cell) and @code{f}
2902: (floating point) with @code{variable} and @code{constant}.
2903:
2904: You can also define your own defining words. E.g.:
2905:
2906: @example
2907: : variable ( "name" -- )
2908: create 0 , ;
2909: @end example
2910:
2911: You can also define defining words that create words that do something
2912: other than just producing their address:
2913:
2914: @example
2915: : constant ( n "name" -- )
2916: create ,
2917: does> ( -- n )
2918: ( addr ) @@ ;
2919:
2920: 5 constant foo
2921: foo .
2922: @end example
2923:
2924: The definition of @code{constant} above ends at the @code{does>}; i.e.,
2925: @code{does>} replaces @code{;}, but it also does something else: It
2926: changes the last defined word such that it pushes the address of the
2927: body of the word and then performs the code after the @code{does>}
2928: whenever it is called.
2929:
2930: In the example above, @code{constant} uses @code{,} to store 5 into the
2931: body of @code{foo}. When @code{foo} executes, it pushes the address of
2932: the body onto the stack, then (in the code after the @code{does>})
2933: fetches the 5 from there.
2934:
2935: The stack comment near the @code{does>} reflects the stack effect of the
2936: defined word, not the stack effect of the code after the @code{does>}
2937: (the difference is that the code expects the address of the body that
2938: the stack comment does not show).
2939:
2940: You can use these definition words to do factoring in cases that involve
2941: (other) definition words. E.g., a field offset is always added to an
2942: address. Instead of defining
2943:
2944: @example
2945: 2 cells constant offset-field1
2946: @end example
2947:
2948: and using this like
2949:
2950: @example
2951: ( addr ) offset-field1 +
2952: @end example
2953:
2954: you can define a definition word
2955:
2956: @example
2957: : simple-field ( n "name" -- )
2958: create ,
2959: does> ( n1 -- n1+n )
2960: ( addr ) @@ + ;
2961: @end example
2962:
2963: Definition and use of field offsets now look like this:
2964:
2965: @example
2966: 2 cells simple-field field1
2967: create mystruct 4 cells allot
2968: mystruct .s field1 .s drop
2969: @end example
2970:
2971: If you want to do something with the word without performing the code
2972: after the @code{does>}, you can access the body of a @code{create}d word
2973: with @code{>body ( xt -- addr )}:
2974:
2975: @example
2976: : value ( n "name" -- )
2977: create ,
2978: does> ( -- n1 )
2979: @@ ;
2980: : to ( n "name" -- )
2981: ' >body ! ;
2982:
2983: 5 value foo
2984: foo .
2985: 7 to foo
2986: foo .
2987: @end example
2988:
2989: @quotation Assignment
2990: Define @code{defer ( "name" -- )}, which creates a word that stores an
2991: XT (at the start the XT of @code{abort}), and upon execution
2992: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
2993: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
2994: recursion is one application of @code{defer}.
2995: @end quotation
2996:
2997: Reference: @ref{User-defined Defining Words}.
2998:
2999:
3000: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3001: @section Arrays and Records
3002: @cindex arrays tutorial
3003: @cindex records tutorial
3004: @cindex structs tutorial
3005:
3006: Forth has no standard words for defining data structures such as arrays
3007: and records (structs in C terminology), but you can build them yourself
3008: based on address arithmetic. You can also define words for defining
3009: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
3010:
3011: One of the first projects a Forth newcomer sets out upon when learning
3012: about defining words is an array defining word (possibly for
3013: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3014: learn something from it. However, don't be disappointed when you later
3015: learn that you have little use for these words (inappropriate use would
3016: be even worse). I have not yet found a set of useful array words yet;
3017: the needs are just too diverse, and named, global arrays (the result of
3018: naive use of defining words) are often not flexible enough (e.g.,
3019: consider how to pass them as parameters). Another such project is a set
3020: of words to help dealing with strings.
3021:
3022: On the other hand, there is a useful set of record words, and it has
3023: been defined in @file{compat/struct.fs}; these words are predefined in
3024: Gforth. They are explained in depth elsewhere in this manual (see
3025: @pxref{Structures}). The @code{simple-field} example above is
3026: simplified variant of fields in this package.
3027:
3028:
3029: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3030: @section @code{POSTPONE}
3031: @cindex postpone tutorial
3032:
3033: You can compile the compilation semantics (instead of compiling the
3034: interpretation semantics) of a word with @code{POSTPONE}:
3035:
3036: @example
3037: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3038: POSTPONE + ; immediate
3039: : foo ( n1 n2 -- n )
3040: MY-+ ;
3041: 1 2 foo .
3042: see foo
3043: @end example
3044:
3045: During the definition of @code{foo} the text interpreter performs the
3046: compilation semantics of @code{MY-+}, which performs the compilation
3047: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3048:
3049: This example also displays separate stack comments for the compilation
3050: semantics and for the stack effect of the compiled code. For words with
3051: default compilation semantics these stack effects are usually not
3052: displayed; the stack effect of the compilation semantics is always
3053: @code{( -- )} for these words, the stack effect for the compiled code is
3054: the stack effect of the interpretation semantics.
3055:
3056: Note that the state of the interpreter does not come into play when
3057: performing the compilation semantics in this way. You can also perform
3058: it interpretively, e.g.:
3059:
3060: @example
3061: : foo2 ( n1 n2 -- n )
3062: [ MY-+ ] ;
3063: 1 2 foo .
3064: see foo
3065: @end example
3066:
3067: However, there are some broken Forth systems where this does not always
3068: work, and therefore this practice was been declared non-standard in
3069: 1999.
3070: @c !! repair.fs
3071:
3072: Here is another example for using @code{POSTPONE}:
3073:
3074: @example
3075: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3076: POSTPONE negate POSTPONE + ; immediate compile-only
3077: : bar ( n1 n2 -- n )
3078: MY-- ;
3079: 2 1 bar .
3080: see bar
3081: @end example
3082:
3083: You can define @code{ENDIF} in this way:
3084:
3085: @example
3086: : ENDIF ( Compilation: orig -- )
3087: POSTPONE then ; immediate
3088: @end example
3089:
3090: @quotation Assignment
3091: Write @code{MY-2DUP} that has compilation semantics equivalent to
3092: @code{2dup}, but compiles @code{over over}.
3093: @end quotation
3094:
3095: @c !! @xref{Macros} for reference
3096:
3097:
3098: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3099: @section @code{Literal}
3100: @cindex literal tutorial
3101:
3102: You cannot @code{POSTPONE} numbers:
3103:
3104: @example
3105: : [FOO] POSTPONE 500 ; immediate
3106: @end example
3107:
3108: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
3109:
3110: @example
3111: : [FOO] ( compilation: --; run-time: -- n )
3112: 500 POSTPONE literal ; immediate
3113:
3114: : flip [FOO] ;
3115: flip .
3116: see flip
3117: @end example
3118:
3119: @code{LITERAL} consumes a number at compile-time (when it's compilation
3120: semantics are executed) and pushes it at run-time (when the code it
3121: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3122: number computed at compile time into the current word:
3123:
3124: @example
3125: : bar ( -- n )
3126: [ 2 2 + ] literal ;
3127: see bar
3128: @end example
3129:
3130: @quotation Assignment
3131: Write @code{]L} which allows writing the example above as @code{: bar (
3132: -- n ) [ 2 2 + ]L ;}
3133: @end quotation
3134:
3135: @c !! @xref{Macros} for reference
3136:
3137:
3138: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3139: @section Advanced macros
3140: @cindex macros, advanced tutorial
3141: @cindex run-time code generation, tutorial
3142:
3143: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
3144: Execution Tokens}. It frequently performs @code{execute}, a relatively
3145: expensive operation in some Forth implementations. You can use
3146: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3147: and produce a word that contains the word to be performed directly:
3148:
3149: @c use ]] ... [[
3150: @example
3151: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3152: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3153: \ array beginning at addr and containing u elements
3154: @{ xt @}
3155: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
3156: POSTPONE i POSTPONE @@ xt compile,
3157: 1 cells POSTPONE literal POSTPONE +loop ;
3158:
3159: : sum-array ( addr u -- n )
3160: 0 rot rot [ ' + compile-map-array ] ;
3161: see sum-array
3162: a 5 sum-array .
3163: @end example
3164:
3165: You can use the full power of Forth for generating the code; here's an
3166: example where the code is generated in a loop:
3167:
3168: @example
3169: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3170: \ n2=n1+(addr1)*n, addr2=addr1+cell
3171: POSTPONE tuck POSTPONE @@
3172: POSTPONE literal POSTPONE * POSTPONE +
3173: POSTPONE swap POSTPONE cell+ ;
3174:
3175: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
3176: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
3177: 0 postpone literal postpone swap
3178: [ ' compile-vmul-step compile-map-array ]
3179: postpone drop ;
3180: see compile-vmul
3181:
3182: : a-vmul ( addr -- n )
3183: \ n=a*v, where v is a vector that's as long as a and starts at addr
3184: [ a 5 compile-vmul ] ;
3185: see a-vmul
3186: a a-vmul .
3187: @end example
3188:
3189: This example uses @code{compile-map-array} to show off, but you could
3190: also use @code{map-array} instead (try it now!).
3191:
3192: You can use this technique for efficient multiplication of large
3193: matrices. In matrix multiplication, you multiply every line of one
3194: matrix with every column of the other matrix. You can generate the code
3195: for one line once, and use it for every column. The only downside of
3196: this technique is that it is cumbersome to recover the memory consumed
3197: by the generated code when you are done (and in more complicated cases
3198: it is not possible portably).
3199:
3200: @c !! @xref{Macros} for reference
3201:
3202:
3203: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3204: @section Compilation Tokens
3205: @cindex compilation tokens, tutorial
3206: @cindex CT, tutorial
3207:
3208: This section is Gforth-specific. You can skip it.
3209:
3210: @code{' word compile,} compiles the interpretation semantics. For words
3211: with default compilation semantics this is the same as performing the
3212: compilation semantics. To represent the compilation semantics of other
3213: words (e.g., words like @code{if} that have no interpretation
3214: semantics), Gforth has the concept of a compilation token (CT,
3215: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3216: You can perform the compilation semantics represented by a CT with
3217: @code{execute}:
3218:
3219: @example
3220: : foo2 ( n1 n2 -- n )
3221: [ comp' + execute ] ;
3222: see foo
3223: @end example
3224:
3225: You can compile the compilation semantics represented by a CT with
3226: @code{postpone,}:
3227:
3228: @example
3229: : foo3 ( -- )
3230: [ comp' + postpone, ] ;
3231: see foo3
3232: @end example
3233:
3234: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
3235: @code{comp'} is particularly useful for words that have no
3236: interpretation semantics:
3237:
3238: @example
3239: ' if
3240: comp' if .s 2drop
3241: @end example
3242:
3243: Reference: @ref{Tokens for Words}.
3244:
3245:
3246: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3247: @section Wordlists and Search Order
3248: @cindex wordlists tutorial
3249: @cindex search order, tutorial
3250:
3251: The dictionary is not just a memory area that allows you to allocate
3252: memory with @code{allot}, it also contains the Forth words, arranged in
3253: several wordlists. When searching for a word in a wordlist,
3254: conceptually you start searching at the youngest and proceed towards
3255: older words (in reality most systems nowadays use hash-tables); i.e., if
3256: you define a word with the same name as an older word, the new word
3257: shadows the older word.
3258:
3259: Which wordlists are searched in which order is determined by the search
3260: order. You can display the search order with @code{order}. It displays
3261: first the search order, starting with the wordlist searched first, then
3262: it displays the wordlist that will contain newly defined words.
3263:
3264: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
3265:
3266: @example
3267: wordlist constant mywords
3268: @end example
3269:
3270: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3271: defined words (the @emph{current} wordlist):
3272:
3273: @example
3274: mywords set-current
3275: order
3276: @end example
3277:
3278: Gforth does not display a name for the wordlist in @code{mywords}
3279: because this wordlist was created anonymously with @code{wordlist}.
3280:
3281: You can get the current wordlist with @code{get-current ( -- wid)}. If
3282: you want to put something into a specific wordlist without overall
3283: effect on the current wordlist, this typically looks like this:
3284:
3285: @example
3286: get-current mywords set-current ( wid )
3287: create someword
3288: ( wid ) set-current
3289: @end example
3290:
3291: You can write the search order with @code{set-order ( wid1 .. widn n --
3292: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3293: searched wordlist is topmost.
3294:
3295: @example
3296: get-order mywords swap 1+ set-order
3297: order
3298: @end example
3299:
3300: Yes, the order of wordlists in the output of @code{order} is reversed
3301: from stack comments and the output of @code{.s} and thus unintuitive.
3302:
3303: @quotation Assignment
3304: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3305: wordlist to the search order. Define @code{previous ( -- )}, which
3306: removes the first searched wordlist from the search order. Experiment
3307: with boundary conditions (you will see some crashes or situations that
3308: are hard or impossible to leave).
3309: @end quotation
3310:
3311: The search order is a powerful foundation for providing features similar
3312: to Modula-2 modules and C++ namespaces. However, trying to modularize
3313: programs in this way has disadvantages for debugging and reuse/factoring
3314: that overcome the advantages in my experience (I don't do huge projects,
3315: though). These disadvantages are not so clear in other
3316: languages/programming environments, because these languages are not so
3317: strong in debugging and reuse.
3318:
3319: @c !! example
3320:
3321: Reference: @ref{Word Lists}.
3322:
3323: @c ******************************************************************
3324: @node Introduction, Words, Tutorial, Top
3325: @comment node-name, next, previous, up
3326: @chapter An Introduction to ANS Forth
3327: @cindex Forth - an introduction
3328:
3329: The difference of this chapter from the Tutorial (@pxref{Tutorial}) is
3330: that it is slower-paced in its examples, but uses them to dive deep into
3331: explaining Forth internals (not covered by the Tutorial). Apart from
3332: that, this chapter covers far less material. It is suitable for reading
3333: without using a computer.
3334:
3335: The primary purpose of this manual is to document Gforth. However, since
3336: Forth is not a widely-known language and there is a lack of up-to-date
3337: teaching material, it seems worthwhile to provide some introductory
3338: material. For other sources of Forth-related
3339: information, see @ref{Forth-related information}.
3340:
3341: The examples in this section should work on any ANS Forth; the
3342: output shown was produced using Gforth. Each example attempts to
3343: reproduce the exact output that Gforth produces. If you try out the
3344: examples (and you should), what you should type is shown @kbd{like this}
3345: and Gforth's response is shown @code{like this}. The single exception is
3346: that, where the example shows @key{RET} it means that you should
3347: press the ``carriage return'' key. Unfortunately, some output formats for
3348: this manual cannot show the difference between @kbd{this} and
3349: @code{this} which will make trying out the examples harder (but not
3350: impossible).
3351:
3352: Forth is an unusual language. It provides an interactive development
3353: environment which includes both an interpreter and compiler. Forth
3354: programming style encourages you to break a problem down into many
3355: @cindex factoring
3356: small fragments (@dfn{factoring}), and then to develop and test each
3357: fragment interactively. Forth advocates assert that breaking the
3358: edit-compile-test cycle used by conventional programming languages can
3359: lead to great productivity improvements.
3360:
3361: @menu
3362: * Introducing the Text Interpreter::
3363: * Stacks and Postfix notation::
3364: * Your first definition::
3365: * How does that work?::
3366: * Forth is written in Forth::
3367: * Review - elements of a Forth system::
3368: * Where to go next::
3369: * Exercises::
3370: @end menu
3371:
3372: @comment ----------------------------------------------
3373: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3374: @section Introducing the Text Interpreter
3375: @cindex text interpreter
3376: @cindex outer interpreter
3377:
3378: @c IMO this is too detailed and the pace is too slow for
3379: @c an introduction. If you know German, take a look at
3380: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3381: @c to see how I do it - anton
3382:
3383: @c nac-> Where I have accepted your comments 100% and modified the text
3384: @c accordingly, I have deleted your comments. Elsewhere I have added a
3385: @c response like this to attempt to rationalise what I have done. Of
3386: @c course, this is a very clumsy mechanism for something that would be
3387: @c done far more efficiently over a beer. Please delete any dialogue
3388: @c you consider closed.
3389:
3390: When you invoke the Forth image, you will see a startup banner printed
3391: and nothing else (if you have Gforth installed on your system, try
3392: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
3393: its command line interpreter, which is called the @dfn{Text Interpreter}
3394: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
3395: about the text interpreter as you read through this chapter, for more
3396: detail @pxref{The Text Interpreter}).
3397:
3398: Although it's not obvious, Forth is actually waiting for your
3399: input. Type a number and press the @key{RET} key:
3400:
3401: @example
3402: @kbd{45@key{RET}} ok
3403: @end example
3404:
3405: Rather than give you a prompt to invite you to input something, the text
3406: interpreter prints a status message @i{after} it has processed a line
3407: of input. The status message in this case (``@code{ ok}'' followed by
3408: carriage-return) indicates that the text interpreter was able to process
3409: all of your input successfully. Now type something illegal:
3410:
3411: @example
3412: @kbd{qwer341@key{RET}}
3413: *the terminal*:2: Undefined word
3414: >>>qwer341<<<
3415: Backtrace:
3416: $2A95B42A20 throw
3417: $2A95B57FB8 no.extensions
3418: @end example
3419:
3420: The exact text, other than the ``Undefined word'' may differ slightly
3421: on your system, but the effect is the same; when the text interpreter
3422: detects an error, it discards any remaining text on a line, resets
3423: certain internal state and prints an error message. For a detailed
3424: description of error messages see @ref{Error messages}.
3425:
3426: The text interpreter waits for you to press carriage-return, and then
3427: processes your input line. Starting at the beginning of the line, it
3428: breaks the line into groups of characters separated by spaces. For each
3429: group of characters in turn, it makes two attempts to do something:
3430:
3431: @itemize @bullet
3432: @item
3433: @cindex name dictionary
3434: It tries to treat it as a command. It does this by searching a @dfn{name
3435: dictionary}. If the group of characters matches an entry in the name
3436: dictionary, the name dictionary provides the text interpreter with
3437: information that allows the text interpreter perform some actions. In
3438: Forth jargon, we say that the group
3439: @cindex word
3440: @cindex definition
3441: @cindex execution token
3442: @cindex xt
3443: of characters names a @dfn{word}, that the dictionary search returns an
3444: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3445: word, and that the text interpreter executes the xt. Often, the terms
3446: @dfn{word} and @dfn{definition} are used interchangeably.
3447: @item
3448: If the text interpreter fails to find a match in the name dictionary, it
3449: tries to treat the group of characters as a number in the current number
3450: base (when you start up Forth, the current number base is base 10). If
3451: the group of characters legitimately represents a number, the text
3452: interpreter pushes the number onto a stack (we'll learn more about that
3453: in the next section).
3454: @end itemize
3455:
3456: If the text interpreter is unable to do either of these things with any
3457: group of characters, it discards the group of characters and the rest of
3458: the line, then prints an error message. If the text interpreter reaches
3459: the end of the line without error, it prints the status message ``@code{ ok}''
3460: followed by carriage-return.
3461:
3462: This is the simplest command we can give to the text interpreter:
3463:
3464: @example
3465: @key{RET} ok
3466: @end example
3467:
3468: The text interpreter did everything we asked it to do (nothing) without
3469: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3470: command:
3471:
3472: @example
3473: @kbd{12 dup fred dup@key{RET}}
3474: *the terminal*:3: Undefined word
3475: 12 dup >>>fred<<< dup
3476: Backtrace:
3477: $2A95B42A20 throw
3478: $2A95B57FB8 no.extensions
3479: @end example
3480:
3481: When you press the carriage-return key, the text interpreter starts to
3482: work its way along the line:
3483:
3484: @itemize @bullet
3485: @item
3486: When it gets to the space after the @code{2}, it takes the group of
3487: characters @code{12} and looks them up in the name
3488: dictionary@footnote{We can't tell if it found them or not, but assume
3489: for now that it did not}. There is no match for this group of characters
3490: in the name dictionary, so it tries to treat them as a number. It is
3491: able to do this successfully, so it puts the number, 12, ``on the stack''
3492: (whatever that means).
3493: @item
3494: The text interpreter resumes scanning the line and gets the next group
3495: of characters, @code{dup}. It looks it up in the name dictionary and
3496: (you'll have to take my word for this) finds it, and executes the word
3497: @code{dup} (whatever that means).
3498: @item
3499: Once again, the text interpreter resumes scanning the line and gets the
3500: group of characters @code{fred}. It looks them up in the name
3501: dictionary, but can't find them. It tries to treat them as a number, but
3502: they don't represent any legal number.
3503: @end itemize
3504:
3505: At this point, the text interpreter gives up and prints an error
3506: message. The error message shows exactly how far the text interpreter
3507: got in processing the line. In particular, it shows that the text
3508: interpreter made no attempt to do anything with the final character
3509: group, @code{dup}, even though we have good reason to believe that the
3510: text interpreter would have no problem looking that word up and
3511: executing it a second time.
3512:
3513:
3514: @comment ----------------------------------------------
3515: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3516: @section Stacks, postfix notation and parameter passing
3517: @cindex text interpreter
3518: @cindex outer interpreter
3519:
3520: In procedural programming languages (like C and Pascal), the
3521: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3522: functions or procedures are called with @dfn{explicit parameters}. For
3523: example, in C we might write:
3524:
3525: @example
3526: total = total + new_volume(length,height,depth);
3527: @end example
3528:
3529: @noindent
3530: where new_volume is a function-call to another piece of code, and total,
3531: length, height and depth are all variables. length, height and depth are
3532: parameters to the function-call.
3533:
3534: In Forth, the equivalent of the function or procedure is the
3535: @dfn{definition} and parameters are implicitly passed between
3536: definitions using a shared stack that is visible to the
3537: programmer. Although Forth does support variables, the existence of the
3538: stack means that they are used far less often than in most other
3539: programming languages. When the text interpreter encounters a number, it
3540: will place (@dfn{push}) it on the stack. There are several stacks (the
3541: actual number is implementation-dependent ...) and the particular stack
3542: used for any operation is implied unambiguously by the operation being
3543: performed. The stack used for all integer operations is called the @dfn{data
3544: stack} and, since this is the stack used most commonly, references to
3545: ``the data stack'' are often abbreviated to ``the stack''.
3546:
3547: The stacks have a last-in, first-out (LIFO) organisation. If you type:
3548:
3549: @example
3550: @kbd{1 2 3@key{RET}} ok
3551: @end example
3552:
3553: Then this instructs the text interpreter to placed three numbers on the
3554: (data) stack. An analogy for the behaviour of the stack is to take a
3555: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3556: the table. The 3 was the last card onto the pile (``last-in'') and if
3557: you take a card off the pile then, unless you're prepared to fiddle a
3558: bit, the card that you take off will be the 3 (``first-out''). The
3559: number that will be first-out of the stack is called the @dfn{top of
3560: stack}, which
3561: @cindex TOS definition
3562: is often abbreviated to @dfn{TOS}.
3563:
3564: To understand how parameters are passed in Forth, consider the
3565: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3566: be surprised to learn that this definition performs addition. More
3567: precisely, it adds two number together and produces a result. Where does
3568: it get the two numbers from? It takes the top two numbers off the
3569: stack. Where does it place the result? On the stack. You can act-out the
3570: behaviour of @code{+} with your playing cards like this:
3571:
3572: @itemize @bullet
3573: @item
3574: Pick up two cards from the stack on the table
3575: @item
3576: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3577: numbers''
3578: @item
3579: Decide that the answer is 5
3580: @item
3581: Shuffle the two cards back into the pack and find a 5
3582: @item
3583: Put a 5 on the remaining ace that's on the table.
3584: @end itemize
3585:
3586: If you don't have a pack of cards handy but you do have Forth running,
3587: you can use the definition @code{.s} to show the current state of the stack,
3588: without affecting the stack. Type:
3589:
3590: @example
3591: @kbd{clearstacks 1 2 3@key{RET}} ok
3592: @kbd{.s@key{RET}} <3> 1 2 3 ok
3593: @end example
3594:
3595: The text interpreter looks up the word @code{clearstacks} and executes
3596: it; it tidies up the stacks and removes any entries that may have been
3597: left on it by earlier examples. The text interpreter pushes each of the
3598: three numbers in turn onto the stack. Finally, the text interpreter
3599: looks up the word @code{.s} and executes it. The effect of executing
3600: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3601: followed by a list of all the items on the stack; the item on the far
3602: right-hand side is the TOS.
3603:
3604: You can now type:
3605:
3606: @example
3607: @kbd{+ .s@key{RET}} <2> 1 5 ok
3608: @end example
3609:
3610: @noindent
3611: which is correct; there are now 2 items on the stack and the result of
3612: the addition is 5.
3613:
3614: If you're playing with cards, try doing a second addition: pick up the
3615: two cards, work out that their sum is 6, shuffle them into the pack,
3616: look for a 6 and place that on the table. You now have just one item on
3617: the stack. What happens if you try to do a third addition? Pick up the
3618: first card, pick up the second card -- ah! There is no second card. This
3619: is called a @dfn{stack underflow} and consitutes an error. If you try to
3620: do the same thing with Forth it often reports an error (probably a Stack
3621: Underflow or an Invalid Memory Address error).
3622:
3623: The opposite situation to a stack underflow is a @dfn{stack overflow},
3624: which simply accepts that there is a finite amount of storage space
3625: reserved for the stack. To stretch the playing card analogy, if you had
3626: enough packs of cards and you piled the cards up on the table, you would
3627: eventually be unable to add another card; you'd hit the ceiling. Gforth
3628: allows you to set the maximum size of the stacks. In general, the only
3629: time that you will get a stack overflow is because a definition has a
3630: bug in it and is generating data on the stack uncontrollably.
3631:
3632: There's one final use for the playing card analogy. If you model your
3633: stack using a pack of playing cards, the maximum number of items on
3634: your stack will be 52 (I assume you didn't use the Joker). The maximum
3635: @i{value} of any item on the stack is 13 (the King). In fact, the only
3636: possible numbers are positive integer numbers 1 through 13; you can't
3637: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3638: think about some of the cards, you can accommodate different
3639: numbers. For example, you could think of the Jack as representing 0,
3640: the Queen as representing -1 and the King as representing -2. Your
3641: @i{range} remains unchanged (you can still only represent a total of 13
3642: numbers) but the numbers that you can represent are -2 through 10.
3643:
3644: In that analogy, the limit was the amount of information that a single
3645: stack entry could hold, and Forth has a similar limit. In Forth, the
3646: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3647: implementation dependent and affects the maximum value that a stack
3648: entry can hold. A Standard Forth provides a cell size of at least
3649: 16-bits, and most desktop systems use a cell size of 32-bits.
3650:
3651: Forth does not do any type checking for you, so you are free to
3652: manipulate and combine stack items in any way you wish. A convenient way
3653: of treating stack items is as 2's complement signed integers, and that
3654: is what Standard words like @code{+} do. Therefore you can type:
3655:
3656: @example
3657: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
3658: @end example
3659:
3660: If you use numbers and definitions like @code{+} in order to turn Forth
3661: into a great big pocket calculator, you will realise that it's rather
3662: different from a normal calculator. Rather than typing 2 + 3 = you had
3663: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3664: result). The terminology used to describe this difference is to say that
3665: your calculator uses @dfn{Infix Notation} (parameters and operators are
3666: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3667: operators are separate), also called @dfn{Reverse Polish Notation}.
3668:
3669: Whilst postfix notation might look confusing to begin with, it has
3670: several important advantages:
3671:
3672: @itemize @bullet
3673: @item
3674: it is unambiguous
3675: @item
3676: it is more concise
3677: @item
3678: it fits naturally with a stack-based system
3679: @end itemize
3680:
3681: To examine these claims in more detail, consider these sums:
3682:
3683: @example
3684: 6 + 5 * 4 =
3685: 4 * 5 + 6 =
3686: @end example
3687:
3688: If you're just learning maths or your maths is very rusty, you will
3689: probably come up with the answer 44 for the first and 26 for the
3690: second. If you are a bit of a whizz at maths you will remember the
3691: @i{convention} that multiplication takes precendence over addition, and
3692: you'd come up with the answer 26 both times. To explain the answer 26
3693: to someone who got the answer 44, you'd probably rewrite the first sum
3694: like this:
3695:
3696: @example
3697: 6 + (5 * 4) =
3698: @end example
3699:
3700: If what you really wanted was to perform the addition before the
3701: multiplication, you would have to use parentheses to force it.
3702:
3703: If you did the first two sums on a pocket calculator you would probably
3704: get the right answers, unless you were very cautious and entered them using
3705: these keystroke sequences:
3706:
3707: 6 + 5 = * 4 =
3708: 4 * 5 = + 6 =
3709:
3710: Postfix notation is unambiguous because the order that the operators
3711: are applied is always explicit; that also means that parentheses are
3712: never required. The operators are @i{active} (the act of quoting the
3713: operator makes the operation occur) which removes the need for ``=''.
3714:
3715: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3716: equivalent ways:
3717:
3718: @example
3719: 6 5 4 * + or:
3720: 5 4 * 6 +
3721: @end example
3722:
3723: An important thing that you should notice about this notation is that
3724: the @i{order} of the numbers does not change; if you want to subtract
3725: 2 from 10 you type @code{10 2 -}.
3726:
3727: The reason that Forth uses postfix notation is very simple to explain: it
3728: makes the implementation extremely simple, and it follows naturally from
3729: using the stack as a mechanism for passing parameters. Another way of
3730: thinking about this is to realise that all Forth definitions are
3731: @i{active}; they execute as they are encountered by the text
3732: interpreter. The result of this is that the syntax of Forth is trivially
3733: simple.
3734:
3735:
3736:
3737: @comment ----------------------------------------------
3738: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3739: @section Your first Forth definition
3740: @cindex first definition
3741:
3742: Until now, the examples we've seen have been trivial; we've just been
3743: using Forth as a bigger-than-pocket calculator. Also, each calculation
3744: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3745: again@footnote{That's not quite true. If you press the up-arrow key on
3746: your keyboard you should be able to scroll back to any earlier command,
3747: edit it and re-enter it.} In this section we'll see how to add new
3748: words to Forth's vocabulary.
3749:
3750: The easiest way to create a new word is to use a @dfn{colon
3751: definition}. We'll define a few and try them out before worrying too
3752: much about how they work. Try typing in these examples; be careful to
3753: copy the spaces accurately:
3754:
3755: @example
3756: : add-two 2 + . ;
3757: : greet ." Hello and welcome" ;
3758: : demo 5 add-two ;
3759: @end example
3760:
3761: @noindent
3762: Now try them out:
3763:
3764: @example
3765: @kbd{greet@key{RET}} Hello and welcome ok
3766: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
3767: @kbd{4 add-two@key{RET}} 6 ok
3768: @kbd{demo@key{RET}} 7 ok
3769: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
3770: @end example
3771:
3772: The first new thing that we've introduced here is the pair of words
3773: @code{:} and @code{;}. These are used to start and terminate a new
3774: definition, respectively. The first word after the @code{:} is the name
3775: for the new definition.
3776:
3777: As you can see from the examples, a definition is built up of words that
3778: have already been defined; Forth makes no distinction between
3779: definitions that existed when you started the system up, and those that
3780: you define yourself.
3781:
3782: The examples also introduce the words @code{.} (dot), @code{."}
3783: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
3784: the stack and displays it. It's like @code{.s} except that it only
3785: displays the top item of the stack and it is destructive; after it has
3786: executed, the number is no longer on the stack. There is always one
3787: space printed after the number, and no spaces before it. Dot-quote
3788: defines a string (a sequence of characters) that will be printed when
3789: the word is executed. The string can contain any printable characters
3790: except @code{"}. A @code{"} has a special function; it is not a Forth
3791: word but it acts as a delimiter (the way that delimiters work is
3792: described in the next section). Finally, @code{dup} duplicates the value
3793: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
3794:
3795: We already know that the text interpreter searches through the
3796: dictionary to locate names. If you've followed the examples earlier, you
3797: will already have a definition called @code{add-two}. Lets try modifying
3798: it by typing in a new definition:
3799:
3800: @example
3801: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
3802: @end example
3803:
3804: Forth recognised that we were defining a word that already exists, and
3805: printed a message to warn us of that fact. Let's try out the new
3806: definition:
3807:
3808: @example
3809: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
3810: @end example
3811:
3812: @noindent
3813: All that we've actually done here, though, is to create a new
3814: definition, with a particular name. The fact that there was already a
3815: definition with the same name did not make any difference to the way
3816: that the new definition was created (except that Forth printed a warning
3817: message). The old definition of add-two still exists (try @code{demo}
3818: again to see that this is true). Any new definition will use the new
3819: definition of @code{add-two}, but old definitions continue to use the
3820: version that already existed at the time that they were @code{compiled}.
3821:
3822: Before you go on to the next section, try defining and redefining some
3823: words of your own.
3824:
3825: @comment ----------------------------------------------
3826: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
3827: @section How does that work?
3828: @cindex parsing words
3829:
3830: @c That's pretty deep (IMO way too deep) for an introduction. - anton
3831:
3832: @c Is it a good idea to talk about the interpretation semantics of a
3833: @c number? We don't have an xt to go along with it. - anton
3834:
3835: @c Now that I have eliminated execution semantics, I wonder if it would not
3836: @c be better to keep them (or add run-time semantics), to make it easier to
3837: @c explain what compilation semantics usually does. - anton
3838:
3839: @c nac-> I removed the term ``default compilation sematics'' from the
3840: @c introductory chapter. Removing ``execution semantics'' was making
3841: @c everything simpler to explain, then I think the use of this term made
3842: @c everything more complex again. I replaced it with ``default
3843: @c semantics'' (which is used elsewhere in the manual) by which I mean
3844: @c ``a definition that has neither the immediate nor the compile-only
3845: @c flag set''.
3846:
3847: @c anton: I have eliminated default semantics (except in one place where it
3848: @c means "default interpretation and compilation semantics"), because it
3849: @c makes no sense in the presence of combined words. I reverted to
3850: @c "execution semantics" where necessary.
3851:
3852: @c nac-> I reworded big chunks of the ``how does that work''
3853: @c section (and, unusually for me, I think I even made it shorter!). See
3854: @c what you think -- I know I have not addressed your primary concern
3855: @c that it is too heavy-going for an introduction. From what I understood
3856: @c of your course notes it looks as though they might be a good framework.
3857: @c Things that I've tried to capture here are some things that came as a
3858: @c great revelation here when I first understood them. Also, I like the
3859: @c fact that a very simple code example shows up almost all of the issues
3860: @c that you need to understand to see how Forth works. That's unique and
3861: @c worthwhile to emphasise.
3862:
3863: @c anton: I think it's a good idea to present the details, especially those
3864: @c that you found to be a revelation, and probably the tutorial tries to be
3865: @c too superficial and does not get some of the things across that make
3866: @c Forth special. I do believe that most of the time these things should
3867: @c be discussed at the end of a section or in separate sections instead of
3868: @c in the middle of a section (e.g., the stuff you added in "User-defined
3869: @c defining words" leads in a completely different direction from the rest
3870: @c of the section).
3871:
3872: Now we're going to take another look at the definition of @code{add-two}
3873: from the previous section. From our knowledge of the way that the text
3874: interpreter works, we would have expected this result when we tried to
3875: define @code{add-two}:
3876:
3877: @example
3878: @kbd{: add-two 2 + . ;@key{RET}}
3879: *the terminal*:4: Undefined word
3880: : >>>add-two<<< 2 + . ;
3881: @end example
3882:
3883: The reason that this didn't happen is bound up in the way that @code{:}
3884: works. The word @code{:} does two special things. The first special
3885: thing that it does prevents the text interpreter from ever seeing the
3886: characters @code{add-two}. The text interpreter uses a variable called
3887: @cindex modifying >IN
3888: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
3889: input line. When it encounters the word @code{:} it behaves in exactly
3890: the same way as it does for any other word; it looks it up in the name
3891: dictionary, finds its xt and executes it. When @code{:} executes, it
3892: looks at the input buffer, finds the word @code{add-two} and advances the
3893: value of @code{>IN} to point past it. It then does some other stuff
3894: associated with creating the new definition (including creating an entry
3895: for @code{add-two} in the name dictionary). When the execution of @code{:}
3896: completes, control returns to the text interpreter, which is oblivious
3897: to the fact that it has been tricked into ignoring part of the input
3898: line.
3899:
3900: @cindex parsing words
3901: Words like @code{:} -- words that advance the value of @code{>IN} and so
3902: prevent the text interpreter from acting on the whole of the input line
3903: -- are called @dfn{parsing words}.
3904:
3905: @cindex @code{state} - effect on the text interpreter
3906: @cindex text interpreter - effect of state
3907: The second special thing that @code{:} does is change the value of a
3908: variable called @code{state}, which affects the way that the text
3909: interpreter behaves. When Gforth starts up, @code{state} has the value
3910: 0, and the text interpreter is said to be @dfn{interpreting}. During a
3911: colon definition (started with @code{:}), @code{state} is set to -1 and
3912: the text interpreter is said to be @dfn{compiling}.
3913:
3914: In this example, the text interpreter is compiling when it processes the
3915: string ``@code{2 + . ;}''. It still breaks the string down into
3916: character sequences in the same way. However, instead of pushing the
3917: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
3918: into the definition of @code{add-two} that will make the number @code{2} get
3919: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
3920: the behaviours of @code{+} and @code{.} are also compiled into the
3921: definition.
3922:
3923: One category of words don't get compiled. These so-called @dfn{immediate
3924: words} get executed (performed @i{now}) regardless of whether the text
3925: interpreter is interpreting or compiling. The word @code{;} is an
3926: immediate word. Rather than being compiled into the definition, it
3927: executes. Its effect is to terminate the current definition, which
3928: includes changing the value of @code{state} back to 0.
3929:
3930: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
3931: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
3932: definition.
3933:
3934: In Forth, every word or number can be described in terms of two
3935: properties:
3936:
3937: @itemize @bullet
3938: @item
3939: @cindex interpretation semantics
3940: Its @dfn{interpretation semantics} describe how it will behave when the
3941: text interpreter encounters it in @dfn{interpret} state. The
3942: interpretation semantics of a word are represented by an @dfn{execution
3943: token}.
3944: @item
3945: @cindex compilation semantics
3946: Its @dfn{compilation semantics} describe how it will behave when the
3947: text interpreter encounters it in @dfn{compile} state. The compilation
3948: semantics of a word are represented in an implementation-dependent way;
3949: Gforth uses a @dfn{compilation token}.
3950: @end itemize
3951:
3952: @noindent
3953: Numbers are always treated in a fixed way:
3954:
3955: @itemize @bullet
3956: @item
3957: When the number is @dfn{interpreted}, its behaviour is to push the
3958: number onto the stack.
3959: @item
3960: When the number is @dfn{compiled}, a piece of code is appended to the
3961: current definition that pushes the number when it runs. (In other words,
3962: the compilation semantics of a number are to postpone its interpretation
3963: semantics until the run-time of the definition that it is being compiled
3964: into.)
3965: @end itemize
3966:
3967: Words don't behave in such a regular way, but most have @i{default
3968: semantics} which means that they behave like this:
3969:
3970: @itemize @bullet
3971: @item
3972: The @dfn{interpretation semantics} of the word are to do something useful.
3973: @item
3974: The @dfn{compilation semantics} of the word are to append its
3975: @dfn{interpretation semantics} to the current definition (so that its
3976: run-time behaviour is to do something useful).
3977: @end itemize
3978:
3979: @cindex immediate words
3980: The actual behaviour of any particular word can be controlled by using
3981: the words @code{immediate} and @code{compile-only} when the word is
3982: defined. These words set flags in the name dictionary entry of the most
3983: recently defined word, and these flags are retrieved by the text
3984: interpreter when it finds the word in the name dictionary.
3985:
3986: A word that is marked as @dfn{immediate} has compilation semantics that
3987: are identical to its interpretation semantics. In other words, it
3988: behaves like this:
3989:
3990: @itemize @bullet
3991: @item
3992: The @dfn{interpretation semantics} of the word are to do something useful.
3993: @item
3994: The @dfn{compilation semantics} of the word are to do something useful
3995: (and actually the same thing); i.e., it is executed during compilation.
3996: @end itemize
3997:
3998: Marking a word as @dfn{compile-only} prohibits the text interpreter from
3999: performing the interpretation semantics of the word directly; an attempt
4000: to do so will generate an error. It is never necessary to use
4001: @code{compile-only} (and it is not even part of ANS Forth, though it is
4002: provided by many implementations) but it is good etiquette to apply it
4003: to a word that will not behave correctly (and might have unexpected
4004: side-effects) in interpret state. For example, it is only legal to use
4005: the conditional word @code{IF} within a definition. If you forget this
4006: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4007: @code{compile-only} allows the text interpreter to generate a helpful
4008: error message rather than subjecting you to the consequences of your
4009: folly.
4010:
4011: This example shows the difference between an immediate and a
4012: non-immediate word:
4013:
4014: @example
4015: : show-state state @@ . ;
4016: : show-state-now show-state ; immediate
4017: : word1 show-state ;
4018: : word2 show-state-now ;
4019: @end example
4020:
4021: The word @code{immediate} after the definition of @code{show-state-now}
4022: makes that word an immediate word. These definitions introduce a new
4023: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4024: variable, and leaves it on the stack. Therefore, the behaviour of
4025: @code{show-state} is to print a number that represents the current value
4026: of @code{state}.
4027:
4028: When you execute @code{word1}, it prints the number 0, indicating that
4029: the system is interpreting. When the text interpreter compiled the
4030: definition of @code{word1}, it encountered @code{show-state} whose
4031: compilation semantics are to append its interpretation semantics to the
4032: current definition. When you execute @code{word1}, it performs the
4033: interpretation semantics of @code{show-state}. At the time that @code{word1}
4034: (and therefore @code{show-state}) are executed, the system is
4035: interpreting.
4036:
4037: When you pressed @key{RET} after entering the definition of @code{word2},
4038: you should have seen the number -1 printed, followed by ``@code{
4039: ok}''. When the text interpreter compiled the definition of
4040: @code{word2}, it encountered @code{show-state-now}, an immediate word,
4041: whose compilation semantics are therefore to perform its interpretation
4042: semantics. It is executed straight away (even before the text
4043: interpreter has moved on to process another group of characters; the
4044: @code{;} in this example). The effect of executing it are to display the
4045: value of @code{state} @i{at the time that the definition of}
4046: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4047: system is compiling at this time. If you execute @code{word2} it does
4048: nothing at all.
4049:
4050: @cindex @code{."}, how it works
4051: Before leaving the subject of immediate words, consider the behaviour of
4052: @code{."} in the definition of @code{greet}, in the previous
4053: section. This word is both a parsing word and an immediate word. Notice
4054: that there is a space between @code{."} and the start of the text
4055: @code{Hello and welcome}, but that there is no space between the last
4056: letter of @code{welcome} and the @code{"} character. The reason for this
4057: is that @code{."} is a Forth word; it must have a space after it so that
4058: the text interpreter can identify it. The @code{"} is not a Forth word;
4059: it is a @dfn{delimiter}. The examples earlier show that, when the string
4060: is displayed, there is neither a space before the @code{H} nor after the
4061: @code{e}. Since @code{."} is an immediate word, it executes at the time
4062: that @code{greet} is defined. When it executes, its behaviour is to
4063: search forward in the input line looking for the delimiter. When it
4064: finds the delimiter, it updates @code{>IN} to point past the
4065: delimiter. It also compiles some magic code into the definition of
4066: @code{greet}; the xt of a run-time routine that prints a text string. It
4067: compiles the string @code{Hello and welcome} into memory so that it is
4068: available to be printed later. When the text interpreter gains control,
4069: the next word it finds in the input stream is @code{;} and so it
4070: terminates the definition of @code{greet}.
4071:
4072:
4073: @comment ----------------------------------------------
4074: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4075: @section Forth is written in Forth
4076: @cindex structure of Forth programs
4077:
4078: When you start up a Forth compiler, a large number of definitions
4079: already exist. In Forth, you develop a new application using bottom-up
4080: programming techniques to create new definitions that are defined in
4081: terms of existing definitions. As you create each definition you can
4082: test and debug it interactively.
4083:
4084: If you have tried out the examples in this section, you will probably
4085: have typed them in by hand; when you leave Gforth, your definitions will
4086: be lost. You can avoid this by using a text editor to enter Forth source
4087: code into a file, and then loading code from the file using
4088: @code{include} (@pxref{Forth source files}). A Forth source file is
4089: processed by the text interpreter, just as though you had typed it in by
4090: hand@footnote{Actually, there are some subtle differences -- see
4091: @ref{The Text Interpreter}.}.
4092:
4093: Gforth also supports the traditional Forth alternative to using text
4094: files for program entry (@pxref{Blocks}).
4095:
4096: In common with many, if not most, Forth compilers, most of Gforth is
4097: actually written in Forth. All of the @file{.fs} files in the
4098: installation directory@footnote{For example,
4099: @file{/usr/local/share/gforth...}} are Forth source files, which you can
4100: study to see examples of Forth programming.
4101:
4102: Gforth maintains a history file that records every line that you type to
4103: the text interpreter. This file is preserved between sessions, and is
4104: used to provide a command-line recall facility. If you enter long
4105: definitions by hand, you can use a text editor to paste them out of the
4106: history file into a Forth source file for reuse at a later time
4107: (for more information @pxref{Command-line editing}).
4108:
4109:
4110: @comment ----------------------------------------------
4111: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4112: @section Review - elements of a Forth system
4113: @cindex elements of a Forth system
4114:
4115: To summarise this chapter:
4116:
4117: @itemize @bullet
4118: @item
4119: Forth programs use @dfn{factoring} to break a problem down into small
4120: fragments called @dfn{words} or @dfn{definitions}.
4121: @item
4122: Forth program development is an interactive process.
4123: @item
4124: The main command loop that accepts input, and controls both
4125: interpretation and compilation, is called the @dfn{text interpreter}
4126: (also known as the @dfn{outer interpreter}).
4127: @item
4128: Forth has a very simple syntax, consisting of words and numbers
4129: separated by spaces or carriage-return characters. Any additional syntax
4130: is imposed by @dfn{parsing words}.
4131: @item
4132: Forth uses a stack to pass parameters between words. As a result, it
4133: uses postfix notation.
4134: @item
4135: To use a word that has previously been defined, the text interpreter
4136: searches for the word in the @dfn{name dictionary}.
4137: @item
4138: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
4139: @item
4140: The text interpreter uses the value of @code{state} to select between
4141: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4142: semantics} of a word that it encounters.
4143: @item
4144: The relationship between the @dfn{interpretation semantics} and
4145: @dfn{compilation semantics} for a word
4146: depend upon the way in which the word was defined (for example, whether
4147: it is an @dfn{immediate} word).
4148: @item
4149: Forth definitions can be implemented in Forth (called @dfn{high-level
4150: definitions}) or in some other way (usually a lower-level language and
4151: as a result often called @dfn{low-level definitions}, @dfn{code
4152: definitions} or @dfn{primitives}).
4153: @item
4154: Many Forth systems are implemented mainly in Forth.
4155: @end itemize
4156:
4157:
4158: @comment ----------------------------------------------
4159: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
4160: @section Where To Go Next
4161: @cindex where to go next
4162:
4163: Amazing as it may seem, if you have read (and understood) this far, you
4164: know almost all the fundamentals about the inner workings of a Forth
4165: system. You certainly know enough to be able to read and understand the
4166: rest of this manual and the ANS Forth document, to learn more about the
4167: facilities that Forth in general and Gforth in particular provide. Even
4168: scarier, you know almost enough to implement your own Forth system.
4169: However, that's not a good idea just yet... better to try writing some
4170: programs in Gforth.
4171:
4172: Forth has such a rich vocabulary that it can be hard to know where to
4173: start in learning it. This section suggests a few sets of words that are
4174: enough to write small but useful programs. Use the word index in this
4175: document to learn more about each word, then try it out and try to write
4176: small definitions using it. Start by experimenting with these words:
4177:
4178: @itemize @bullet
4179: @item
4180: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4181: @item
4182: Comparison: @code{MIN MAX =}
4183: @item
4184: Logic: @code{AND OR XOR NOT}
4185: @item
4186: Stack manipulation: @code{DUP DROP SWAP OVER}
4187: @item
4188: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
4189: @item
4190: Input/Output: @code{. ." EMIT CR KEY}
4191: @item
4192: Defining words: @code{: ; CREATE}
4193: @item
4194: Memory allocation words: @code{ALLOT ,}
4195: @item
4196: Tools: @code{SEE WORDS .S MARKER}
4197: @end itemize
4198:
4199: When you have mastered those, go on to:
4200:
4201: @itemize @bullet
4202: @item
4203: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
4204: @item
4205: Memory access: @code{@@ !}
4206: @end itemize
4207:
4208: When you have mastered these, there's nothing for it but to read through
4209: the whole of this manual and find out what you've missed.
4210:
4211: @comment ----------------------------------------------
4212: @node Exercises, , Where to go next, Introduction
4213: @section Exercises
4214: @cindex exercises
4215:
4216: TODO: provide a set of programming excercises linked into the stuff done
4217: already and into other sections of the manual. Provide solutions to all
4218: the exercises in a .fs file in the distribution.
4219:
4220: @c Get some inspiration from Starting Forth and Kelly&Spies.
4221:
4222: @c excercises:
4223: @c 1. take inches and convert to feet and inches.
4224: @c 2. take temperature and convert from fahrenheight to celcius;
4225: @c may need to care about symmetric vs floored??
4226: @c 3. take input line and do character substitution
4227: @c to encipher or decipher
4228: @c 4. as above but work on a file for in and out
4229: @c 5. take input line and convert to pig-latin
4230: @c
4231: @c thing of sets of things to exercise then come up with
4232: @c problems that need those things.
4233:
4234:
4235: @c ******************************************************************
4236: @node Words, Error messages, Introduction, Top
4237: @chapter Forth Words
4238: @cindex words
4239:
4240: @menu
4241: * Notation::
4242: * Case insensitivity::
4243: * Comments::
4244: * Boolean Flags::
4245: * Arithmetic::
4246: * Stack Manipulation::
4247: * Memory::
4248: * Control Structures::
4249: * Defining Words::
4250: * Interpretation and Compilation Semantics::
4251: * Tokens for Words::
4252: * Compiling words::
4253: * The Text Interpreter::
4254: * The Input Stream::
4255: * Word Lists::
4256: * Environmental Queries::
4257: * Files::
4258: * Blocks::
4259: * Other I/O::
4260: * OS command line arguments::
4261: * Locals::
4262: * Structures::
4263: * Object-oriented Forth::
4264: * Programming Tools::
4265: * C Interface::
4266: * Assembler and Code Words::
4267: * Threading Words::
4268: * Passing Commands to the OS::
4269: * Keeping track of Time::
4270: * Miscellaneous Words::
4271: @end menu
4272:
4273: @node Notation, Case insensitivity, Words, Words
4274: @section Notation
4275: @cindex notation of glossary entries
4276: @cindex format of glossary entries
4277: @cindex glossary notation format
4278: @cindex word glossary entry format
4279:
4280: The Forth words are described in this section in the glossary notation
4281: that has become a de-facto standard for Forth texts:
4282:
4283: @format
4284: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
4285: @end format
4286: @i{Description}
4287:
4288: @table @var
4289: @item word
4290: The name of the word.
4291:
4292: @item Stack effect
4293: @cindex stack effect
4294: The stack effect is written in the notation @code{@i{before} --
4295: @i{after}}, where @i{before} and @i{after} describe the top of
4296: stack entries before and after the execution of the word. The rest of
4297: the stack is not touched by the word. The top of stack is rightmost,
4298: i.e., a stack sequence is written as it is typed in. Note that Gforth
4299: uses a separate floating point stack, but a unified stack
4300: notation. Also, return stack effects are not shown in @i{stack
4301: effect}, but in @i{Description}. The name of a stack item describes
4302: the type and/or the function of the item. See below for a discussion of
4303: the types.
4304:
4305: All words have two stack effects: A compile-time stack effect and a
4306: run-time stack effect. The compile-time stack-effect of most words is
4307: @i{ -- }. If the compile-time stack-effect of a word deviates from
4308: this standard behaviour, or the word does other unusual things at
4309: compile time, both stack effects are shown; otherwise only the run-time
4310: stack effect is shown.
4311:
4312: @cindex pronounciation of words
4313: @item pronunciation
4314: How the word is pronounced.
4315:
4316: @cindex wordset
4317: @cindex environment wordset
4318: @item wordset
4319: The ANS Forth standard is divided into several word sets. A standard
4320: system need not support all of them. Therefore, in theory, the fewer
4321: word sets your program uses the more portable it will be. However, we
4322: suspect that most ANS Forth systems on personal machines will feature
4323: all word sets. Words that are not defined in ANS Forth have
4324: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
4325: describes words that will work in future releases of Gforth;
4326: @code{gforth-internal} words are more volatile. Environmental query
4327: strings are also displayed like words; you can recognize them by the
4328: @code{environment} in the word set field.
4329:
4330: @item Description
4331: A description of the behaviour of the word.
4332: @end table
4333:
4334: @cindex types of stack items
4335: @cindex stack item types
4336: The type of a stack item is specified by the character(s) the name
4337: starts with:
4338:
4339: @table @code
4340: @item f
4341: @cindex @code{f}, stack item type
4342: Boolean flags, i.e. @code{false} or @code{true}.
4343: @item c
4344: @cindex @code{c}, stack item type
4345: Char
4346: @item w
4347: @cindex @code{w}, stack item type
4348: Cell, can contain an integer or an address
4349: @item n
4350: @cindex @code{n}, stack item type
4351: signed integer
4352: @item u
4353: @cindex @code{u}, stack item type
4354: unsigned integer
4355: @item d
4356: @cindex @code{d}, stack item type
4357: double sized signed integer
4358: @item ud
4359: @cindex @code{ud}, stack item type
4360: double sized unsigned integer
4361: @item r
4362: @cindex @code{r}, stack item type
4363: Float (on the FP stack)
4364: @item a-
4365: @cindex @code{a_}, stack item type
4366: Cell-aligned address
4367: @item c-
4368: @cindex @code{c_}, stack item type
4369: Char-aligned address (note that a Char may have two bytes in Windows NT)
4370: @item f-
4371: @cindex @code{f_}, stack item type
4372: Float-aligned address
4373: @item df-
4374: @cindex @code{df_}, stack item type
4375: Address aligned for IEEE double precision float
4376: @item sf-
4377: @cindex @code{sf_}, stack item type
4378: Address aligned for IEEE single precision float
4379: @item xt
4380: @cindex @code{xt}, stack item type
4381: Execution token, same size as Cell
4382: @item wid
4383: @cindex @code{wid}, stack item type
4384: Word list ID, same size as Cell
4385: @item ior, wior
4386: @cindex ior type description
4387: @cindex wior type description
4388: I/O result code, cell-sized. In Gforth, you can @code{throw} iors.
4389: @item f83name
4390: @cindex @code{f83name}, stack item type
4391: Pointer to a name structure
4392: @item "
4393: @cindex @code{"}, stack item type
4394: string in the input stream (not on the stack). The terminating character
4395: is a blank by default. If it is not a blank, it is shown in @code{<>}
4396: quotes.
4397: @end table
4398:
4399: @comment ----------------------------------------------
4400: @node Case insensitivity, Comments, Notation, Words
4401: @section Case insensitivity
4402: @cindex case sensitivity
4403: @cindex upper and lower case
4404:
4405: Gforth is case-insensitive; you can enter definitions and invoke
4406: Standard words using upper, lower or mixed case (however,
4407: @pxref{core-idef, Implementation-defined options, Implementation-defined
4408: options}).
4409:
4410: ANS Forth only @i{requires} implementations to recognise Standard words
4411: when they are typed entirely in upper case. Therefore, a Standard
4412: program must use upper case for all Standard words. You can use whatever
4413: case you like for words that you define, but in a Standard program you
4414: have to use the words in the same case that you defined them.
4415:
4416: Gforth supports case sensitivity through @code{table}s (case-sensitive
4417: wordlists, @pxref{Word Lists}).
4418:
4419: Two people have asked how to convert Gforth to be case-sensitive; while
4420: we think this is a bad idea, you can change all wordlists into tables
4421: like this:
4422:
4423: @example
4424: ' table-find forth-wordlist wordlist-map @ !
4425: @end example
4426:
4427: Note that you now have to type the predefined words in the same case
4428: that we defined them, which are varying. You may want to convert them
4429: to your favourite case before doing this operation (I won't explain how,
4430: because if you are even contemplating doing this, you'd better have
4431: enough knowledge of Forth systems to know this already).
4432:
4433: @node Comments, Boolean Flags, Case insensitivity, Words
4434: @section Comments
4435: @cindex comments
4436:
4437: Forth supports two styles of comment; the traditional @i{in-line} comment,
4438: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
4439:
4440:
4441: doc-(
4442: doc-\
4443: doc-\G
4444:
4445:
4446: @node Boolean Flags, Arithmetic, Comments, Words
4447: @section Boolean Flags
4448: @cindex Boolean flags
4449:
4450: A Boolean flag is cell-sized. A cell with all bits clear represents the
4451: flag @code{false} and a flag with all bits set represents the flag
4452: @code{true}. Words that check a flag (for example, @code{IF}) will treat
4453: a cell that has @i{any} bit set as @code{true}.
4454: @c on and off to Memory?
4455: @c true and false to "Bitwise operations" or "Numeric comparison"?
4456:
4457: doc-true
4458: doc-false
4459: doc-on
4460: doc-off
4461:
4462:
4463: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
4464: @section Arithmetic
4465: @cindex arithmetic words
4466:
4467: @cindex division with potentially negative operands
4468: Forth arithmetic is not checked, i.e., you will not hear about integer
4469: overflow on addition or multiplication, you may hear about division by
4470: zero if you are lucky. The operator is written after the operands, but
4471: the operands are still in the original order. I.e., the infix @code{2-1}
4472: corresponds to @code{2 1 -}. Forth offers a variety of division
4473: operators. If you perform division with potentially negative operands,
4474: you do not want to use @code{/} or @code{/mod} with its undefined
4475: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4476: former, @pxref{Mixed precision}).
4477: @comment TODO discuss the different division forms and the std approach
4478:
4479: @menu
4480: * Single precision::
4481: * Double precision:: Double-cell integer arithmetic
4482: * Bitwise operations::
4483: * Numeric comparison::
4484: * Mixed precision:: Operations with single and double-cell integers
4485: * Floating Point::
4486: @end menu
4487:
4488: @node Single precision, Double precision, Arithmetic, Arithmetic
4489: @subsection Single precision
4490: @cindex single precision arithmetic words
4491:
4492: @c !! cell undefined
4493:
4494: By default, numbers in Forth are single-precision integers that are one
4495: cell in size. They can be signed or unsigned, depending upon how you
4496: treat them. For the rules used by the text interpreter for recognising
4497: single-precision integers see @ref{Number Conversion}.
4498:
4499: These words are all defined for signed operands, but some of them also
4500: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
4501: @code{*}.
4502:
4503: doc-+
4504: doc-1+
4505: doc-under+
4506: doc--
4507: doc-1-
4508: doc-*
4509: doc-/
4510: doc-mod
4511: doc-/mod
4512: doc-negate
4513: doc-abs
4514: doc-min
4515: doc-max
4516: doc-floored
4517:
4518:
4519: @node Double precision, Bitwise operations, Single precision, Arithmetic
4520: @subsection Double precision
4521: @cindex double precision arithmetic words
4522:
4523: For the rules used by the text interpreter for
4524: recognising double-precision integers, see @ref{Number Conversion}.
4525:
4526: A double precision number is represented by a cell pair, with the most
4527: significant cell at the TOS. It is trivial to convert an unsigned single
4528: to a double: simply push a @code{0} onto the TOS. Since numbers are
4529: represented by Gforth using 2's complement arithmetic, converting a
4530: signed single to a (signed) double requires sign-extension across the
4531: most significant cell. This can be achieved using @code{s>d}. The moral
4532: of the story is that you cannot convert a number without knowing whether
4533: it represents an unsigned or a signed number.
4534:
4535: These words are all defined for signed operands, but some of them also
4536: work for unsigned numbers: @code{d+}, @code{d-}.
4537:
4538: doc-s>d
4539: doc-d>s
4540: doc-d+
4541: doc-d-
4542: doc-dnegate
4543: doc-dabs
4544: doc-dmin
4545: doc-dmax
4546:
4547:
4548: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
4549: @subsection Bitwise operations
4550: @cindex bitwise operation words
4551:
4552:
4553: doc-and
4554: doc-or
4555: doc-xor
4556: doc-invert
4557: doc-lshift
4558: doc-rshift
4559: doc-2*
4560: doc-d2*
4561: doc-2/
4562: doc-d2/
4563:
4564:
4565: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
4566: @subsection Numeric comparison
4567: @cindex numeric comparison words
4568:
4569: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
4570: d0= d0<>}) work for for both signed and unsigned numbers.
4571:
4572: doc-<
4573: doc-<=
4574: doc-<>
4575: doc-=
4576: doc->
4577: doc->=
4578:
4579: doc-0<
4580: doc-0<=
4581: doc-0<>
4582: doc-0=
4583: doc-0>
4584: doc-0>=
4585:
4586: doc-u<
4587: doc-u<=
4588: @c u<> and u= exist but are the same as <> and =
4589: @c doc-u<>
4590: @c doc-u=
4591: doc-u>
4592: doc-u>=
4593:
4594: doc-within
4595:
4596: doc-d<
4597: doc-d<=
4598: doc-d<>
4599: doc-d=
4600: doc-d>
4601: doc-d>=
4602:
4603: doc-d0<
4604: doc-d0<=
4605: doc-d0<>
4606: doc-d0=
4607: doc-d0>
4608: doc-d0>=
4609:
4610: doc-du<
4611: doc-du<=
4612: @c du<> and du= exist but are the same as d<> and d=
4613: @c doc-du<>
4614: @c doc-du=
4615: doc-du>
4616: doc-du>=
4617:
4618:
4619: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
4620: @subsection Mixed precision
4621: @cindex mixed precision arithmetic words
4622:
4623:
4624: doc-m+
4625: doc-*/
4626: doc-*/mod
4627: doc-m*
4628: doc-um*
4629: doc-m*/
4630: doc-um/mod
4631: doc-fm/mod
4632: doc-sm/rem
4633:
4634:
4635: @node Floating Point, , Mixed precision, Arithmetic
4636: @subsection Floating Point
4637: @cindex floating point arithmetic words
4638:
4639: For the rules used by the text interpreter for
4640: recognising floating-point numbers see @ref{Number Conversion}.
4641:
4642: Gforth has a separate floating point stack, but the documentation uses
4643: the unified notation.@footnote{It's easy to generate the separate
4644: notation from that by just separating the floating-point numbers out:
4645: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
4646: r3 )}.}
4647:
4648: @cindex floating-point arithmetic, pitfalls
4649: Floating point numbers have a number of unpleasant surprises for the
4650: unwary (e.g., floating point addition is not associative) and even a few
4651: for the wary. You should not use them unless you know what you are doing
4652: or you don't care that the results you get are totally bogus. If you
4653: want to learn about the problems of floating point numbers (and how to
4654: avoid them), you might start with @cite{David Goldberg,
4655: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
4656: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
4657: Surveys 23(1):5@minus{}48, March 1991}.
4658:
4659:
4660: doc-d>f
4661: doc-f>d
4662: doc-f+
4663: doc-f-
4664: doc-f*
4665: doc-f/
4666: doc-fnegate
4667: doc-fabs
4668: doc-fmax
4669: doc-fmin
4670: doc-floor
4671: doc-fround
4672: doc-f**
4673: doc-fsqrt
4674: doc-fexp
4675: doc-fexpm1
4676: doc-fln
4677: doc-flnp1
4678: doc-flog
4679: doc-falog
4680: doc-f2*
4681: doc-f2/
4682: doc-1/f
4683: doc-precision
4684: doc-set-precision
4685:
4686: @cindex angles in trigonometric operations
4687: @cindex trigonometric operations
4688: Angles in floating point operations are given in radians (a full circle
4689: has 2 pi radians).
4690:
4691: doc-fsin
4692: doc-fcos
4693: doc-fsincos
4694: doc-ftan
4695: doc-fasin
4696: doc-facos
4697: doc-fatan
4698: doc-fatan2
4699: doc-fsinh
4700: doc-fcosh
4701: doc-ftanh
4702: doc-fasinh
4703: doc-facosh
4704: doc-fatanh
4705: doc-pi
4706:
4707: @cindex equality of floats
4708: @cindex floating-point comparisons
4709: One particular problem with floating-point arithmetic is that comparison
4710: for equality often fails when you would expect it to succeed. For this
4711: reason approximate equality is often preferred (but you still have to
4712: know what you are doing). Also note that IEEE NaNs may compare
4713: differently from what you might expect. The comparison words are:
4714:
4715: doc-f~rel
4716: doc-f~abs
4717: doc-f~
4718: doc-f=
4719: doc-f<>
4720:
4721: doc-f<
4722: doc-f<=
4723: doc-f>
4724: doc-f>=
4725:
4726: doc-f0<
4727: doc-f0<=
4728: doc-f0<>
4729: doc-f0=
4730: doc-f0>
4731: doc-f0>=
4732:
4733:
4734: @node Stack Manipulation, Memory, Arithmetic, Words
4735: @section Stack Manipulation
4736: @cindex stack manipulation words
4737:
4738: @cindex floating-point stack in the standard
4739: Gforth maintains a number of separate stacks:
4740:
4741: @cindex data stack
4742: @cindex parameter stack
4743: @itemize @bullet
4744: @item
4745: A data stack (also known as the @dfn{parameter stack}) -- for
4746: characters, cells, addresses, and double cells.
4747:
4748: @cindex floating-point stack
4749: @item
4750: A floating point stack -- for holding floating point (FP) numbers.
4751:
4752: @cindex return stack
4753: @item
4754: A return stack -- for holding the return addresses of colon
4755: definitions and other (non-FP) data.
4756:
4757: @cindex locals stack
4758: @item
4759: A locals stack -- for holding local variables.
4760: @end itemize
4761:
4762: @menu
4763: * Data stack::
4764: * Floating point stack::
4765: * Return stack::
4766: * Locals stack::
4767: * Stack pointer manipulation::
4768: @end menu
4769:
4770: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
4771: @subsection Data stack
4772: @cindex data stack manipulation words
4773: @cindex stack manipulations words, data stack
4774:
4775:
4776: doc-drop
4777: doc-nip
4778: doc-dup
4779: doc-over
4780: doc-tuck
4781: doc-swap
4782: doc-pick
4783: doc-rot
4784: doc--rot
4785: doc-?dup
4786: doc-roll
4787: doc-2drop
4788: doc-2nip
4789: doc-2dup
4790: doc-2over
4791: doc-2tuck
4792: doc-2swap
4793: doc-2rot
4794:
4795:
4796: @node Floating point stack, Return stack, Data stack, Stack Manipulation
4797: @subsection Floating point stack
4798: @cindex floating-point stack manipulation words
4799: @cindex stack manipulation words, floating-point stack
4800:
4801: Whilst every sane Forth has a separate floating-point stack, it is not
4802: strictly required; an ANS Forth system could theoretically keep
4803: floating-point numbers on the data stack. As an additional difficulty,
4804: you don't know how many cells a floating-point number takes. It is
4805: reportedly possible to write words in a way that they work also for a
4806: unified stack model, but we do not recommend trying it. Instead, just
4807: say that your program has an environmental dependency on a separate
4808: floating-point stack.
4809:
4810: doc-floating-stack
4811:
4812: doc-fdrop
4813: doc-fnip
4814: doc-fdup
4815: doc-fover
4816: doc-ftuck
4817: doc-fswap
4818: doc-fpick
4819: doc-frot
4820:
4821:
4822: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
4823: @subsection Return stack
4824: @cindex return stack manipulation words
4825: @cindex stack manipulation words, return stack
4826:
4827: @cindex return stack and locals
4828: @cindex locals and return stack
4829: A Forth system is allowed to keep local variables on the
4830: return stack. This is reasonable, as local variables usually eliminate
4831: the need to use the return stack explicitly. So, if you want to produce
4832: a standard compliant program and you are using local variables in a
4833: word, forget about return stack manipulations in that word (refer to the
4834: standard document for the exact rules).
4835:
4836: doc->r
4837: doc-r>
4838: doc-r@
4839: doc-rdrop
4840: doc-2>r
4841: doc-2r>
4842: doc-2r@
4843: doc-2rdrop
4844:
4845:
4846: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
4847: @subsection Locals stack
4848:
4849: Gforth uses an extra locals stack. It is described, along with the
4850: reasons for its existence, in @ref{Locals implementation}.
4851:
4852: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
4853: @subsection Stack pointer manipulation
4854: @cindex stack pointer manipulation words
4855:
4856: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
4857: doc-sp0
4858: doc-sp@
4859: doc-sp!
4860: doc-fp0
4861: doc-fp@
4862: doc-fp!
4863: doc-rp0
4864: doc-rp@
4865: doc-rp!
4866: doc-lp0
4867: doc-lp@
4868: doc-lp!
4869:
4870:
4871: @node Memory, Control Structures, Stack Manipulation, Words
4872: @section Memory
4873: @cindex memory words
4874:
4875: @menu
4876: * Memory model::
4877: * Dictionary allocation::
4878: * Heap Allocation::
4879: * Memory Access::
4880: * Address arithmetic::
4881: * Memory Blocks::
4882: @end menu
4883:
4884: In addition to the standard Forth memory allocation words, there is also
4885: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
4886: garbage collector}.
4887:
4888: @node Memory model, Dictionary allocation, Memory, Memory
4889: @subsection ANS Forth and Gforth memory models
4890:
4891: @c The ANS Forth description is a mess (e.g., is the heap part of
4892: @c the dictionary?), so let's not stick to closely with it.
4893:
4894: ANS Forth considers a Forth system as consisting of several address
4895: spaces, of which only @dfn{data space} is managed and accessible with
4896: the memory words. Memory not necessarily in data space includes the
4897: stacks, the code (called code space) and the headers (called name
4898: space). In Gforth everything is in data space, but the code for the
4899: primitives is usually read-only.
4900:
4901: Data space is divided into a number of areas: The (data space portion of
4902: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
4903: refer to the search data structure embodied in word lists and headers,
4904: because it is used for looking up names, just as you would in a
4905: conventional dictionary.}, the heap, and a number of system-allocated
4906: buffers.
4907:
4908: @cindex address arithmetic restrictions, ANS vs. Gforth
4909: @cindex contiguous regions, ANS vs. Gforth
4910: In ANS Forth data space is also divided into contiguous regions. You
4911: can only use address arithmetic within a contiguous region, not between
4912: them. Usually each allocation gives you one contiguous region, but the
4913: dictionary allocation words have additional rules (@pxref{Dictionary
4914: allocation}).
4915:
4916: Gforth provides one big address space, and address arithmetic can be
4917: performed between any addresses. However, in the dictionary headers or
4918: code are interleaved with data, so almost the only contiguous data space
4919: regions there are those described by ANS Forth as contiguous; but you
4920: can be sure that the dictionary is allocated towards increasing
4921: addresses even between contiguous regions. The memory order of
4922: allocations in the heap is platform-dependent (and possibly different
4923: from one run to the next).
4924:
4925:
4926: @node Dictionary allocation, Heap Allocation, Memory model, Memory
4927: @subsection Dictionary allocation
4928: @cindex reserving data space
4929: @cindex data space - reserving some
4930:
4931: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
4932: you want to deallocate X, you also deallocate everything
4933: allocated after X.
4934:
4935: @cindex contiguous regions in dictionary allocation
4936: The allocations using the words below are contiguous and grow the region
4937: towards increasing addresses. Other words that allocate dictionary
4938: memory of any kind (i.e., defining words including @code{:noname}) end
4939: the contiguous region and start a new one.
4940:
4941: In ANS Forth only @code{create}d words are guaranteed to produce an
4942: address that is the start of the following contiguous region. In
4943: particular, the cell allocated by @code{variable} is not guaranteed to
4944: be contiguous with following @code{allot}ed memory.
4945:
4946: You can deallocate memory by using @code{allot} with a negative argument
4947: (with some restrictions, see @code{allot}). For larger deallocations use
4948: @code{marker}.
4949:
4950:
4951: doc-here
4952: doc-unused
4953: doc-allot
4954: doc-c,
4955: doc-f,
4956: doc-,
4957: doc-2,
4958:
4959: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
4960: course you should allocate memory in an aligned way, too. I.e., before
4961: allocating allocating a cell, @code{here} must be cell-aligned, etc.
4962: The words below align @code{here} if it is not already. Basically it is
4963: only already aligned for a type, if the last allocation was a multiple
4964: of the size of this type and if @code{here} was aligned for this type
4965: before.
4966:
4967: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
4968: ANS Forth (@code{maxalign}ed in Gforth).
4969:
4970: doc-align
4971: doc-falign
4972: doc-sfalign
4973: doc-dfalign
4974: doc-maxalign
4975: doc-cfalign
4976:
4977:
4978: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
4979: @subsection Heap allocation
4980: @cindex heap allocation
4981: @cindex dynamic allocation of memory
4982: @cindex memory-allocation word set
4983:
4984: @cindex contiguous regions and heap allocation
4985: Heap allocation supports deallocation of allocated memory in any
4986: order. Dictionary allocation is not affected by it (i.e., it does not
4987: end a contiguous region). In Gforth, these words are implemented using
4988: the standard C library calls malloc(), free() and resize().
4989:
4990: The memory region produced by one invocation of @code{allocate} or
4991: @code{resize} is internally contiguous. There is no contiguity between
4992: such a region and any other region (including others allocated from the
4993: heap).
4994:
4995: doc-allocate
4996: doc-free
4997: doc-resize
4998:
4999:
5000: @node Memory Access, Address arithmetic, Heap Allocation, Memory
5001: @subsection Memory Access
5002: @cindex memory access words
5003:
5004: doc-@
5005: doc-!
5006: doc-+!
5007: doc-c@
5008: doc-c!
5009: doc-2@
5010: doc-2!
5011: doc-f@
5012: doc-f!
5013: doc-sf@
5014: doc-sf!
5015: doc-df@
5016: doc-df!
5017: doc-sw@
5018: doc-uw@
5019: doc-w!
5020: doc-sl@
5021: doc-ul@
5022: doc-l!
5023:
5024: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5025: @subsection Address arithmetic
5026: @cindex address arithmetic words
5027:
5028: Address arithmetic is the foundation on which you can build data
5029: structures like arrays, records (@pxref{Structures}) and objects
5030: (@pxref{Object-oriented Forth}).
5031:
5032: @cindex address unit
5033: @cindex au (address unit)
5034: ANS Forth does not specify the sizes of the data types. Instead, it
5035: offers a number of words for computing sizes and doing address
5036: arithmetic. Address arithmetic is performed in terms of address units
5037: (aus); on most systems the address unit is one byte. Note that a
5038: character may have more than one au, so @code{chars} is no noop (on
5039: platforms where it is a noop, it compiles to nothing).
5040:
5041: The basic address arithmetic words are @code{+} and @code{-}. E.g., if
5042: you have the address of a cell, perform @code{1 cells +}, and you will
5043: have the address of the next cell.
5044:
5045: @cindex contiguous regions and address arithmetic
5046: In ANS Forth you can perform address arithmetic only within a contiguous
5047: region, i.e., if you have an address into one region, you can only add
5048: and subtract such that the result is still within the region; you can
5049: only subtract or compare addresses from within the same contiguous
5050: region. Reasons: several contiguous regions can be arranged in memory
5051: in any way; on segmented systems addresses may have unusual
5052: representations, such that address arithmetic only works within a
5053: region. Gforth provides a few more guarantees (linear address space,
5054: dictionary grows upwards), but in general I have found it easy to stay
5055: within contiguous regions (exception: computing and comparing to the
5056: address just beyond the end of an array).
5057:
5058: @cindex alignment of addresses for types
5059: ANS Forth also defines words for aligning addresses for specific
5060: types. Many computers require that accesses to specific data types
5061: must only occur at specific addresses; e.g., that cells may only be
5062: accessed at addresses divisible by 4. Even if a machine allows unaligned
5063: accesses, it can usually perform aligned accesses faster.
5064:
5065: For the performance-conscious: alignment operations are usually only
5066: necessary during the definition of a data structure, not during the
5067: (more frequent) accesses to it.
5068:
5069: ANS Forth defines no words for character-aligning addresses. This is not
5070: an oversight, but reflects the fact that addresses that are not
5071: char-aligned have no use in the standard and therefore will not be
5072: created.
5073:
5074: @cindex @code{CREATE} and alignment
5075: ANS Forth guarantees that addresses returned by @code{CREATE}d words
5076: are cell-aligned; in addition, Gforth guarantees that these addresses
5077: are aligned for all purposes.
5078:
5079: Note that the ANS Forth word @code{char} has nothing to do with address
5080: arithmetic.
5081:
5082:
5083: doc-chars
5084: doc-char+
5085: doc-cells
5086: doc-cell+
5087: doc-cell
5088: doc-aligned
5089: doc-floats
5090: doc-float+
5091: doc-float
5092: doc-faligned
5093: doc-sfloats
5094: doc-sfloat+
5095: doc-sfaligned
5096: doc-dfloats
5097: doc-dfloat+
5098: doc-dfaligned
5099: doc-maxaligned
5100: doc-cfaligned
5101: doc-address-unit-bits
5102: doc-/w
5103: doc-/l
5104:
5105: @node Memory Blocks, , Address arithmetic, Memory
5106: @subsection Memory Blocks
5107: @cindex memory block words
5108: @cindex character strings - moving and copying
5109:
5110: Memory blocks often represent character strings; For ways of storing
5111: character strings in memory see @ref{String Formats}. For other
5112: string-processing words see @ref{Displaying characters and strings}.
5113:
5114: A few of these words work on address unit blocks. In that case, you
5115: usually have to insert @code{CHARS} before the word when working on
5116: character strings. Most words work on character blocks, and expect a
5117: char-aligned address.
5118:
5119: When copying characters between overlapping memory regions, use
5120: @code{chars move} or choose carefully between @code{cmove} and
5121: @code{cmove>}.
5122:
5123: doc-move
5124: doc-erase
5125: doc-cmove
5126: doc-cmove>
5127: doc-fill
5128: doc-blank
5129: doc-compare
5130: doc-str=
5131: doc-str<
5132: doc-string-prefix?
5133: doc-search
5134: doc--trailing
5135: doc-/string
5136: doc-bounds
5137: doc-pad
5138:
5139: @comment TODO examples
5140:
5141:
5142: @node Control Structures, Defining Words, Memory, Words
5143: @section Control Structures
5144: @cindex control structures
5145:
5146: Control structures in Forth cannot be used interpretively, only in a
5147: colon definition@footnote{To be precise, they have no interpretation
5148: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5149: not like this limitation, but have not seen a satisfying way around it
5150: yet, although many schemes have been proposed.
5151:
5152: @menu
5153: * Selection:: IF ... ELSE ... ENDIF
5154: * Simple Loops:: BEGIN ...
5155: * Counted Loops:: DO
5156: * Arbitrary control structures::
5157: * Calls and returns::
5158: * Exception Handling::
5159: @end menu
5160:
5161: @node Selection, Simple Loops, Control Structures, Control Structures
5162: @subsection Selection
5163: @cindex selection control structures
5164: @cindex control structures for selection
5165:
5166: @cindex @code{IF} control structure
5167: @example
5168: @i{flag}
5169: IF
5170: @i{code}
5171: ENDIF
5172: @end example
5173: @noindent
5174:
5175: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5176: with any bit set represents truth) @i{code} is executed.
5177:
5178: @example
5179: @i{flag}
5180: IF
5181: @i{code1}
5182: ELSE
5183: @i{code2}
5184: ENDIF
5185: @end example
5186:
5187: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5188: executed.
5189:
5190: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5191: standard, and @code{ENDIF} is not, although it is quite popular. We
5192: recommend using @code{ENDIF}, because it is less confusing for people
5193: who also know other languages (and is not prone to reinforcing negative
5194: prejudices against Forth in these people). Adding @code{ENDIF} to a
5195: system that only supplies @code{THEN} is simple:
5196: @example
5197: : ENDIF POSTPONE then ; immediate
5198: @end example
5199:
5200: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5201: (adv.)} has the following meanings:
5202: @quotation
5203: ... 2b: following next after in order ... 3d: as a necessary consequence
5204: (if you were there, then you saw them).
5205: @end quotation
5206: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5207: and many other programming languages has the meaning 3d.]
5208:
5209: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
5210: you can avoid using @code{?dup}. Using these alternatives is also more
5211: efficient than using @code{?dup}. Definitions in ANS Forth
5212: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5213: @file{compat/control.fs}.
5214:
5215: @cindex @code{CASE} control structure
5216: @example
5217: @i{n}
5218: CASE
5219: @i{n1} OF @i{code1} ENDOF
5220: @i{n2} OF @i{code2} ENDOF
5221: @dots{}
5222: ( n ) @i{default-code} ( n )
5223: ENDCASE ( )
5224: @end example
5225:
5226: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}. If
5227: no @i{ni} matches, the optional @i{default-code} is executed. The
5228: optional default case can be added by simply writing the code after
5229: the last @code{ENDOF}. It may use @i{n}, which is on top of the stack,
5230: but must not consume it. The value @i{n} is consumed by this
5231: construction (either by a OF that matches, or by the ENDCASE, if no OF
5232: matches).
5233:
5234: @progstyle
5235: To keep the code understandable, you should ensure that you change the
5236: stack in the same way (wrt. number and types of stack items consumed
5237: and pushed) on all paths through a selection construct.
5238:
5239: @node Simple Loops, Counted Loops, Selection, Control Structures
5240: @subsection Simple Loops
5241: @cindex simple loops
5242: @cindex loops without count
5243:
5244: @cindex @code{WHILE} loop
5245: @example
5246: BEGIN
5247: @i{code1}
5248: @i{flag}
5249: WHILE
5250: @i{code2}
5251: REPEAT
5252: @end example
5253:
5254: @i{code1} is executed and @i{flag} is computed. If it is true,
5255: @i{code2} is executed and the loop is restarted; If @i{flag} is
5256: false, execution continues after the @code{REPEAT}.
5257:
5258: @cindex @code{UNTIL} loop
5259: @example
5260: BEGIN
5261: @i{code}
5262: @i{flag}
5263: UNTIL
5264: @end example
5265:
5266: @i{code} is executed. The loop is restarted if @code{flag} is false.
5267:
5268: @progstyle
5269: To keep the code understandable, a complete iteration of the loop should
5270: not change the number and types of the items on the stacks.
5271:
5272: @cindex endless loop
5273: @cindex loops, endless
5274: @example
5275: BEGIN
5276: @i{code}
5277: AGAIN
5278: @end example
5279:
5280: This is an endless loop.
5281:
5282: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5283: @subsection Counted Loops
5284: @cindex counted loops
5285: @cindex loops, counted
5286: @cindex @code{DO} loops
5287:
5288: The basic counted loop is:
5289: @example
5290: @i{limit} @i{start}
5291: ?DO
5292: @i{body}
5293: LOOP
5294: @end example
5295:
5296: This performs one iteration for every integer, starting from @i{start}
5297: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
5298: accessed with @code{i}. For example, the loop:
5299: @example
5300: 10 0 ?DO
5301: i .
5302: LOOP
5303: @end example
5304: @noindent
5305: prints @code{0 1 2 3 4 5 6 7 8 9}
5306:
5307: The index of the innermost loop can be accessed with @code{i}, the index
5308: of the next loop with @code{j}, and the index of the third loop with
5309: @code{k}.
5310:
5311:
5312: doc-i
5313: doc-j
5314: doc-k
5315:
5316:
5317: The loop control data are kept on the return stack, so there are some
5318: restrictions on mixing return stack accesses and counted loop words. In
5319: particuler, if you put values on the return stack outside the loop, you
5320: cannot read them inside the loop@footnote{well, not in a way that is
5321: portable.}. If you put values on the return stack within a loop, you
5322: have to remove them before the end of the loop and before accessing the
5323: index of the loop.
5324:
5325: There are several variations on the counted loop:
5326:
5327: @itemize @bullet
5328: @item
5329: @code{LEAVE} leaves the innermost counted loop immediately; execution
5330: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5331:
5332: @example
5333: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5334: @end example
5335: prints @code{0 1 2 3}
5336:
5337:
5338: @item
5339: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5340: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5341: return stack so @code{EXIT} can get to its return address. For example:
5342:
5343: @example
5344: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5345: @end example
5346: prints @code{0 1 2 3}
5347:
5348:
5349: @item
5350: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
5351: (and @code{LOOP} iterates until they become equal by wrap-around
5352: arithmetic). This behaviour is usually not what you want. Therefore,
5353: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
5354: @code{?DO}), which do not enter the loop if @i{start} is greater than
5355: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
5356: unsigned loop parameters.
5357:
5358: @item
5359: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5360: the loop, independent of the loop parameters. Do not use @code{DO}, even
5361: if you know that the loop is entered in any case. Such knowledge tends
5362: to become invalid during maintenance of a program, and then the
5363: @code{DO} will make trouble.
5364:
5365: @item
5366: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5367: index by @i{n} instead of by 1. The loop is terminated when the border
5368: between @i{limit-1} and @i{limit} is crossed. E.g.:
5369:
5370: @example
5371: 4 0 +DO i . 2 +LOOP
5372: @end example
5373: @noindent
5374: prints @code{0 2}
5375:
5376: @example
5377: 4 1 +DO i . 2 +LOOP
5378: @end example
5379: @noindent
5380: prints @code{1 3}
5381:
5382: @item
5383: @cindex negative increment for counted loops
5384: @cindex counted loops with negative increment
5385: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
5386:
5387: @example
5388: -1 0 ?DO i . -1 +LOOP
5389: @end example
5390: @noindent
5391: prints @code{0 -1}
5392:
5393: @example
5394: 0 0 ?DO i . -1 +LOOP
5395: @end example
5396: prints nothing.
5397:
5398: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5399: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5400: index by @i{u} each iteration. The loop is terminated when the border
5401: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
5402: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5403:
5404: @example
5405: -2 0 -DO i . 1 -LOOP
5406: @end example
5407: @noindent
5408: prints @code{0 -1}
5409:
5410: @example
5411: -1 0 -DO i . 1 -LOOP
5412: @end example
5413: @noindent
5414: prints @code{0}
5415:
5416: @example
5417: 0 0 -DO i . 1 -LOOP
5418: @end example
5419: @noindent
5420: prints nothing.
5421:
5422: @end itemize
5423:
5424: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
5425: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5426: for these words that uses only standard words is provided in
5427: @file{compat/loops.fs}.
5428:
5429:
5430: @cindex @code{FOR} loops
5431: Another counted loop is:
5432: @example
5433: @i{n}
5434: FOR
5435: @i{body}
5436: NEXT
5437: @end example
5438: This is the preferred loop of native code compiler writers who are too
5439: lazy to optimize @code{?DO} loops properly. This loop structure is not
5440: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5441: @code{i} produces values starting with @i{n} and ending with 0. Other
5442: Forth systems may behave differently, even if they support @code{FOR}
5443: loops. To avoid problems, don't use @code{FOR} loops.
5444:
5445: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5446: @subsection Arbitrary control structures
5447: @cindex control structures, user-defined
5448:
5449: @cindex control-flow stack
5450: ANS Forth permits and supports using control structures in a non-nested
5451: way. Information about incomplete control structures is stored on the
5452: control-flow stack. This stack may be implemented on the Forth data
5453: stack, and this is what we have done in Gforth.
5454:
5455: @cindex @code{orig}, control-flow stack item
5456: @cindex @code{dest}, control-flow stack item
5457: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5458: entry represents a backward branch target. A few words are the basis for
5459: building any control structure possible (except control structures that
5460: need storage, like calls, coroutines, and backtracking).
5461:
5462:
5463: doc-if
5464: doc-ahead
5465: doc-then
5466: doc-begin
5467: doc-until
5468: doc-again
5469: doc-cs-pick
5470: doc-cs-roll
5471:
5472:
5473: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5474: manipulate the control-flow stack in a portable way. Without them, you
5475: would need to know how many stack items are occupied by a control-flow
5476: entry (many systems use one cell. In Gforth they currently take three,
5477: but this may change in the future).
5478:
5479: Some standard control structure words are built from these words:
5480:
5481:
5482: doc-else
5483: doc-while
5484: doc-repeat
5485:
5486:
5487: @noindent
5488: Gforth adds some more control-structure words:
5489:
5490:
5491: doc-endif
5492: doc-?dup-if
5493: doc-?dup-0=-if
5494:
5495:
5496: @noindent
5497: Counted loop words constitute a separate group of words:
5498:
5499:
5500: doc-?do
5501: doc-+do
5502: doc-u+do
5503: doc--do
5504: doc-u-do
5505: doc-do
5506: doc-for
5507: doc-loop
5508: doc-+loop
5509: doc--loop
5510: doc-next
5511: doc-leave
5512: doc-?leave
5513: doc-unloop
5514: doc-done
5515:
5516:
5517: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5518: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
5519: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5520: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5521: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5522: resolved (by using one of the loop-ending words or @code{DONE}).
5523:
5524: @noindent
5525: Another group of control structure words are:
5526:
5527:
5528: doc-case
5529: doc-endcase
5530: doc-of
5531: doc-endof
5532:
5533:
5534: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5535: @code{CS-ROLL}.
5536:
5537: @subsubsection Programming Style
5538: @cindex control structures programming style
5539: @cindex programming style, arbitrary control structures
5540:
5541: In order to ensure readability we recommend that you do not create
5542: arbitrary control structures directly, but define new control structure
5543: words for the control structure you want and use these words in your
5544: program. For example, instead of writing:
5545:
5546: @example
5547: BEGIN
5548: ...
5549: IF [ 1 CS-ROLL ]
5550: ...
5551: AGAIN THEN
5552: @end example
5553:
5554: @noindent
5555: we recommend defining control structure words, e.g.,
5556:
5557: @example
5558: : WHILE ( DEST -- ORIG DEST )
5559: POSTPONE IF
5560: 1 CS-ROLL ; immediate
5561:
5562: : REPEAT ( orig dest -- )
5563: POSTPONE AGAIN
5564: POSTPONE THEN ; immediate
5565: @end example
5566:
5567: @noindent
5568: and then using these to create the control structure:
5569:
5570: @example
5571: BEGIN
5572: ...
5573: WHILE
5574: ...
5575: REPEAT
5576: @end example
5577:
5578: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5579: @code{WHILE} are predefined, so in this example it would not be
5580: necessary to define them.
5581:
5582: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5583: @subsection Calls and returns
5584: @cindex calling a definition
5585: @cindex returning from a definition
5586:
5587: @cindex recursive definitions
5588: A definition can be called simply be writing the name of the definition
5589: to be called. Normally a definition is invisible during its own
5590: definition. If you want to write a directly recursive definition, you
5591: can use @code{recursive} to make the current definition visible, or
5592: @code{recurse} to call the current definition directly.
5593:
5594:
5595: doc-recursive
5596: doc-recurse
5597:
5598:
5599: @comment TODO add example of the two recursion methods
5600: @quotation
5601: @progstyle
5602: I prefer using @code{recursive} to @code{recurse}, because calling the
5603: definition by name is more descriptive (if the name is well-chosen) than
5604: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5605: implementation, it is much better to read (and think) ``now sort the
5606: partitions'' than to read ``now do a recursive call''.
5607: @end quotation
5608:
5609: For mutual recursion, use @code{Defer}red words, like this:
5610:
5611: @example
5612: Defer foo
5613:
5614: : bar ( ... -- ... )
5615: ... foo ... ;
5616:
5617: :noname ( ... -- ... )
5618: ... bar ... ;
5619: IS foo
5620: @end example
5621:
5622: Deferred words are discussed in more detail in @ref{Deferred words}.
5623:
5624: The current definition returns control to the calling definition when
5625: the end of the definition is reached or @code{EXIT} is encountered.
5626:
5627: doc-exit
5628: doc-;s
5629:
5630:
5631: @node Exception Handling, , Calls and returns, Control Structures
5632: @subsection Exception Handling
5633: @cindex exceptions
5634:
5635: @c quit is a very bad idea for error handling,
5636: @c because it does not translate into a THROW
5637: @c it also does not belong into this chapter
5638:
5639: If a word detects an error condition that it cannot handle, it can
5640: @code{throw} an exception. In the simplest case, this will terminate
5641: your program, and report an appropriate error.
5642:
5643: doc-throw
5644:
5645: @code{Throw} consumes a cell-sized error number on the stack. There are
5646: some predefined error numbers in ANS Forth (see @file{errors.fs}). In
5647: Gforth (and most other systems) you can use the iors produced by various
5648: words as error numbers (e.g., a typical use of @code{allocate} is
5649: @code{allocate throw}). Gforth also provides the word @code{exception}
5650: to define your own error numbers (with decent error reporting); an ANS
5651: Forth version of this word (but without the error messages) is available
5652: in @code{compat/except.fs}. And finally, you can use your own error
5653: numbers (anything outside the range -4095..0), but won't get nice error
5654: messages, only numbers. For example, try:
5655:
5656: @example
5657: -10 throw \ ANS defined
5658: -267 throw \ system defined
5659: s" my error" exception throw \ user defined
5660: 7 throw \ arbitrary number
5661: @end example
5662:
5663: doc---exception-exception
5664:
5665: A common idiom to @code{THROW} a specific error if a flag is true is
5666: this:
5667:
5668: @example
5669: @code{( flag ) 0<> @i{errno} and throw}
5670: @end example
5671:
5672: Your program can provide exception handlers to catch exceptions. An
5673: exception handler can be used to correct the problem, or to clean up
5674: some data structures and just throw the exception to the next exception
5675: handler. Note that @code{throw} jumps to the dynamically innermost
5676: exception handler. The system's exception handler is outermost, and just
5677: prints an error and restarts command-line interpretation (or, in batch
5678: mode (i.e., while processing the shell command line), leaves Gforth).
5679:
5680: The ANS Forth way to catch exceptions is @code{catch}:
5681:
5682: doc-catch
5683: doc-nothrow
5684:
5685: The most common use of exception handlers is to clean up the state when
5686: an error happens. E.g.,
5687:
5688: @example
5689: base @ >r hex \ actually the hex should be inside foo, or we h
5690: ['] foo catch ( nerror|0 )
5691: r> base !
5692: ( nerror|0 ) throw \ pass it on
5693: @end example
5694:
5695: A use of @code{catch} for handling the error @code{myerror} might look
5696: like this:
5697:
5698: @example
5699: ['] foo catch
5700: CASE
5701: myerror OF ... ( do something about it ) nothrow ENDOF
5702: dup throw \ default: pass other errors on, do nothing on non-errors
5703: ENDCASE
5704: @end example
5705:
5706: Having to wrap the code into a separate word is often cumbersome,
5707: therefore Gforth provides an alternative syntax:
5708:
5709: @example
5710: TRY
5711: @i{code1}
5712: RECOVER
5713: @i{code2} \ optional
5714: ENDTRY
5715: @end example
5716:
5717: This performs @i{Code1}. If @i{code1} completes normally, execution
5718: continues after the @code{endtry}. If @i{Code1} throws, the stacks are
5719: reset to the state during @code{try}, the throw value is pushed on the
5720: data stack, and execution constinues at @i{code2}, and finally falls
5721: through the @code{endtry} into the following code.
5722:
5723: doc-try
5724: doc-recover
5725: doc-endtry
5726:
5727: The cleanup example from above in this syntax:
5728:
5729: @example
5730: base @ >r TRY
5731: hex foo \ now the hex is placed correctly
5732: 0 \ value for throw
5733: RECOVER ENDTRY
5734: r> base ! throw
5735: @end example
5736:
5737: And here's the error handling example:
5738:
5739: @example
5740: TRY
5741: foo
5742: RECOVER
5743: CASE
5744: myerror OF ... ( do something about it ) nothrow ENDOF
5745: throw \ pass other errors on
5746: ENDCASE
5747: ENDTRY
5748: @end example
5749:
5750: @progstyle
5751: As usual, you should ensure that the stack depth is statically known at
5752: the end: either after the @code{throw} for passing on errors, or after
5753: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
5754: selection construct for handling the error).
5755:
5756: There are two alternatives to @code{throw}: @code{Abort"} is conditional
5757: and you can provide an error message. @code{Abort} just produces an
5758: ``Aborted'' error.
5759:
5760: The problem with these words is that exception handlers cannot
5761: differentiate between different @code{abort"}s; they just look like
5762: @code{-2 throw} to them (the error message cannot be accessed by
5763: standard programs). Similar @code{abort} looks like @code{-1 throw} to
5764: exception handlers.
5765:
5766: doc-abort"
5767: doc-abort
5768:
5769:
5770:
5771: @c -------------------------------------------------------------
5772: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
5773: @section Defining Words
5774: @cindex defining words
5775:
5776: Defining words are used to extend Forth by creating new entries in the dictionary.
5777:
5778: @menu
5779: * CREATE::
5780: * Variables:: Variables and user variables
5781: * Constants::
5782: * Values:: Initialised variables
5783: * Colon Definitions::
5784: * Anonymous Definitions:: Definitions without names
5785: * Supplying names:: Passing definition names as strings
5786: * User-defined Defining Words::
5787: * Deferred words:: Allow forward references
5788: * Aliases::
5789: @end menu
5790:
5791: @node CREATE, Variables, Defining Words, Defining Words
5792: @subsection @code{CREATE}
5793: @cindex simple defining words
5794: @cindex defining words, simple
5795:
5796: Defining words are used to create new entries in the dictionary. The
5797: simplest defining word is @code{CREATE}. @code{CREATE} is used like
5798: this:
5799:
5800: @example
5801: CREATE new-word1
5802: @end example
5803:
5804: @code{CREATE} is a parsing word, i.e., it takes an argument from the
5805: input stream (@code{new-word1} in our example). It generates a
5806: dictionary entry for @code{new-word1}. When @code{new-word1} is
5807: executed, all that it does is leave an address on the stack. The address
5808: represents the value of the data space pointer (@code{HERE}) at the time
5809: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
5810: associating a name with the address of a region of memory.
5811:
5812: doc-create
5813:
5814: Note that in ANS Forth guarantees only for @code{create} that its body
5815: is in dictionary data space (i.e., where @code{here}, @code{allot}
5816: etc. work, @pxref{Dictionary allocation}). Also, in ANS Forth only
5817: @code{create}d words can be modified with @code{does>}
5818: (@pxref{User-defined Defining Words}). And in ANS Forth @code{>body}
5819: can only be applied to @code{create}d words.
5820:
5821: By extending this example to reserve some memory in data space, we end
5822: up with something like a @i{variable}. Here are two different ways to do
5823: it:
5824:
5825: @example
5826: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
5827: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
5828: @end example
5829:
5830: The variable can be examined and modified using @code{@@} (``fetch'') and
5831: @code{!} (``store'') like this:
5832:
5833: @example
5834: new-word2 @@ . \ get address, fetch from it and display
5835: 1234 new-word2 ! \ new value, get address, store to it
5836: @end example
5837:
5838: @cindex arrays
5839: A similar mechanism can be used to create arrays. For example, an
5840: 80-character text input buffer:
5841:
5842: @example
5843: CREATE text-buf 80 chars allot
5844:
5845: text-buf 0 chars c@@ \ the 1st character (offset 0)
5846: text-buf 3 chars c@@ \ the 4th character (offset 3)
5847: @end example
5848:
5849: You can build arbitrarily complex data structures by allocating
5850: appropriate areas of memory. For further discussions of this, and to
5851: learn about some Gforth tools that make it easier,
5852: @xref{Structures}.
5853:
5854:
5855: @node Variables, Constants, CREATE, Defining Words
5856: @subsection Variables
5857: @cindex variables
5858:
5859: The previous section showed how a sequence of commands could be used to
5860: generate a variable. As a final refinement, the whole code sequence can
5861: be wrapped up in a defining word (pre-empting the subject of the next
5862: section), making it easier to create new variables:
5863:
5864: @example
5865: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
5866: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
5867:
5868: myvariableX foo \ variable foo starts off with an unknown value
5869: myvariable0 joe \ whilst joe is initialised to 0
5870:
5871: 45 3 * foo ! \ set foo to 135
5872: 1234 joe ! \ set joe to 1234
5873: 3 joe +! \ increment joe by 3.. to 1237
5874: @end example
5875:
5876: Not surprisingly, there is no need to define @code{myvariable}, since
5877: Forth already has a definition @code{Variable}. ANS Forth does not
5878: guarantee that a @code{Variable} is initialised when it is created
5879: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
5880: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
5881: like @code{myvariable0}). Forth also provides @code{2Variable} and
5882: @code{fvariable} for double and floating-point variables, respectively
5883: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
5884: store a boolean, you can use @code{on} and @code{off} to toggle its
5885: state.
5886:
5887: doc-variable
5888: doc-2variable
5889: doc-fvariable
5890:
5891: @cindex user variables
5892: @cindex user space
5893: The defining word @code{User} behaves in the same way as @code{Variable}.
5894: The difference is that it reserves space in @i{user (data) space} rather
5895: than normal data space. In a Forth system that has a multi-tasker, each
5896: task has its own set of user variables.
5897:
5898: doc-user
5899: @c doc-udp
5900: @c doc-uallot
5901:
5902: @comment TODO is that stuff about user variables strictly correct? Is it
5903: @comment just terminal tasks that have user variables?
5904: @comment should document tasker.fs (with some examples) elsewhere
5905: @comment in this manual, then expand on user space and user variables.
5906:
5907: @node Constants, Values, Variables, Defining Words
5908: @subsection Constants
5909: @cindex constants
5910:
5911: @code{Constant} allows you to declare a fixed value and refer to it by
5912: name. For example:
5913:
5914: @example
5915: 12 Constant INCHES-PER-FOOT
5916: 3E+08 fconstant SPEED-O-LIGHT
5917: @end example
5918:
5919: A @code{Variable} can be both read and written, so its run-time
5920: behaviour is to supply an address through which its current value can be
5921: manipulated. In contrast, the value of a @code{Constant} cannot be
5922: changed once it has been declared@footnote{Well, often it can be -- but
5923: not in a Standard, portable way. It's safer to use a @code{Value} (read
5924: on).} so it's not necessary to supply the address -- it is more
5925: efficient to return the value of the constant directly. That's exactly
5926: what happens; the run-time effect of a constant is to put its value on
5927: the top of the stack (You can find one
5928: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
5929:
5930: Forth also provides @code{2Constant} and @code{fconstant} for defining
5931: double and floating-point constants, respectively.
5932:
5933: doc-constant
5934: doc-2constant
5935: doc-fconstant
5936:
5937: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
5938: @c nac-> How could that not be true in an ANS Forth? You can't define a
5939: @c constant, use it and then delete the definition of the constant..
5940:
5941: @c anton->An ANS Forth system can compile a constant to a literal; On
5942: @c decompilation you would see only the number, just as if it had been used
5943: @c in the first place. The word will stay, of course, but it will only be
5944: @c used by the text interpreter (no run-time duties, except when it is
5945: @c POSTPONEd or somesuch).
5946:
5947: @c nac:
5948: @c I agree that it's rather deep, but IMO it is an important difference
5949: @c relative to other programming languages.. often it's annoying: it
5950: @c certainly changes my programming style relative to C.
5951:
5952: @c anton: In what way?
5953:
5954: Constants in Forth behave differently from their equivalents in other
5955: programming languages. In other languages, a constant (such as an EQU in
5956: assembler or a #define in C) only exists at compile-time; in the
5957: executable program the constant has been translated into an absolute
5958: number and, unless you are using a symbolic debugger, it's impossible to
5959: know what abstract thing that number represents. In Forth a constant has
5960: an entry in the header space and remains there after the code that uses
5961: it has been defined. In fact, it must remain in the dictionary since it
5962: has run-time duties to perform. For example:
5963:
5964: @example
5965: 12 Constant INCHES-PER-FOOT
5966: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
5967: @end example
5968:
5969: @cindex in-lining of constants
5970: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
5971: associated with the constant @code{INCHES-PER-FOOT}. If you use
5972: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
5973: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
5974: attempt to optimise constants by in-lining them where they are used. You
5975: can force Gforth to in-line a constant like this:
5976:
5977: @example
5978: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
5979: @end example
5980:
5981: If you use @code{see} to decompile @i{this} version of
5982: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
5983: longer present. To understand how this works, read
5984: @ref{Interpret/Compile states}, and @ref{Literals}.
5985:
5986: In-lining constants in this way might improve execution time
5987: fractionally, and can ensure that a constant is now only referenced at
5988: compile-time. However, the definition of the constant still remains in
5989: the dictionary. Some Forth compilers provide a mechanism for controlling
5990: a second dictionary for holding transient words such that this second
5991: dictionary can be deleted later in order to recover memory
5992: space. However, there is no standard way of doing this.
5993:
5994:
5995: @node Values, Colon Definitions, Constants, Defining Words
5996: @subsection Values
5997: @cindex values
5998:
5999: A @code{Value} behaves like a @code{Constant}, but it can be changed.
6000: @code{TO} is a parsing word that changes a @code{Values}. In Gforth
6001: (not in ANS Forth) you can access (and change) a @code{value} also with
6002: @code{>body}.
6003:
6004: Here are some
6005: examples:
6006:
6007: @example
6008: 12 Value APPLES \ Define APPLES with an initial value of 12
6009: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6010: 1 ' APPLES >body +! \ Increment APPLES. Non-standard usage.
6011: APPLES \ puts 35 on the top of the stack.
6012: @end example
6013:
6014: doc-value
6015: doc-to
6016:
6017:
6018:
6019: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6020: @subsection Colon Definitions
6021: @cindex colon definitions
6022:
6023: @example
6024: : name ( ... -- ... )
6025: word1 word2 word3 ;
6026: @end example
6027:
6028: @noindent
6029: Creates a word called @code{name} that, upon execution, executes
6030: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
6031:
6032: The explanation above is somewhat superficial. For simple examples of
6033: colon definitions see @ref{Your first definition}. For an in-depth
6034: discussion of some of the issues involved, @xref{Interpretation and
6035: Compilation Semantics}.
6036:
6037: doc-:
6038: doc-;
6039:
6040:
6041: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
6042: @subsection Anonymous Definitions
6043: @cindex colon definitions
6044: @cindex defining words without name
6045:
6046: Sometimes you want to define an @dfn{anonymous word}; a word without a
6047: name. You can do this with:
6048:
6049: doc-:noname
6050:
6051: This leaves the execution token for the word on the stack after the
6052: closing @code{;}. Here's an example in which a deferred word is
6053: initialised with an @code{xt} from an anonymous colon definition:
6054:
6055: @example
6056: Defer deferred
6057: :noname ( ... -- ... )
6058: ... ;
6059: IS deferred
6060: @end example
6061:
6062: @noindent
6063: Gforth provides an alternative way of doing this, using two separate
6064: words:
6065:
6066: doc-noname
6067: @cindex execution token of last defined word
6068: doc-latestxt
6069:
6070: @noindent
6071: The previous example can be rewritten using @code{noname} and
6072: @code{latestxt}:
6073:
6074: @example
6075: Defer deferred
6076: noname : ( ... -- ... )
6077: ... ;
6078: latestxt IS deferred
6079: @end example
6080:
6081: @noindent
6082: @code{noname} works with any defining word, not just @code{:}.
6083:
6084: @code{latestxt} also works when the last word was not defined as
6085: @code{noname}. It does not work for combined words, though. It also has
6086: the useful property that is is valid as soon as the header for a
6087: definition has been built. Thus:
6088:
6089: @example
6090: latestxt . : foo [ latestxt . ] ; ' foo .
6091: @end example
6092:
6093: @noindent
6094: prints 3 numbers; the last two are the same.
6095:
6096: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
6097: @subsection Supplying the name of a defined word
6098: @cindex names for defined words
6099: @cindex defining words, name given in a string
6100:
6101: By default, a defining word takes the name for the defined word from the
6102: input stream. Sometimes you want to supply the name from a string. You
6103: can do this with:
6104:
6105: doc-nextname
6106:
6107: For example:
6108:
6109: @example
6110: s" foo" nextname create
6111: @end example
6112:
6113: @noindent
6114: is equivalent to:
6115:
6116: @example
6117: create foo
6118: @end example
6119:
6120: @noindent
6121: @code{nextname} works with any defining word.
6122:
6123:
6124: @node User-defined Defining Words, Deferred words, Supplying names, Defining Words
6125: @subsection User-defined Defining Words
6126: @cindex user-defined defining words
6127: @cindex defining words, user-defined
6128:
6129: You can create a new defining word by wrapping defining-time code around
6130: an existing defining word and putting the sequence in a colon
6131: definition.
6132:
6133: @c anton: This example is very complex and leads in a quite different
6134: @c direction from the CREATE-DOES> stuff that follows. It should probably
6135: @c be done elsewhere, or as a subsubsection of this subsection (or as a
6136: @c subsection of Defining Words)
6137:
6138: For example, suppose that you have a word @code{stats} that
6139: gathers statistics about colon definitions given the @i{xt} of the
6140: definition, and you want every colon definition in your application to
6141: make a call to @code{stats}. You can define and use a new version of
6142: @code{:} like this:
6143:
6144: @example
6145: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6146: ... ; \ other code
6147:
6148: : my: : latestxt postpone literal ['] stats compile, ;
6149:
6150: my: foo + - ;
6151: @end example
6152:
6153: When @code{foo} is defined using @code{my:} these steps occur:
6154:
6155: @itemize @bullet
6156: @item
6157: @code{my:} is executed.
6158: @item
6159: The @code{:} within the definition (the one between @code{my:} and
6160: @code{latestxt}) is executed, and does just what it always does; it parses
6161: the input stream for a name, builds a dictionary header for the name
6162: @code{foo} and switches @code{state} from interpret to compile.
6163: @item
6164: The word @code{latestxt} is executed. It puts the @i{xt} for the word that is
6165: being defined -- @code{foo} -- onto the stack.
6166: @item
6167: The code that was produced by @code{postpone literal} is executed; this
6168: causes the value on the stack to be compiled as a literal in the code
6169: area of @code{foo}.
6170: @item
6171: The code @code{['] stats} compiles a literal into the definition of
6172: @code{my:}. When @code{compile,} is executed, that literal -- the
6173: execution token for @code{stats} -- is layed down in the code area of
6174: @code{foo} , following the literal@footnote{Strictly speaking, the
6175: mechanism that @code{compile,} uses to convert an @i{xt} into something
6176: in the code area is implementation-dependent. A threaded implementation
6177: might spit out the execution token directly whilst another
6178: implementation might spit out a native code sequence.}.
6179: @item
6180: At this point, the execution of @code{my:} is complete, and control
6181: returns to the text interpreter. The text interpreter is in compile
6182: state, so subsequent text @code{+ -} is compiled into the definition of
6183: @code{foo} and the @code{;} terminates the definition as always.
6184: @end itemize
6185:
6186: You can use @code{see} to decompile a word that was defined using
6187: @code{my:} and see how it is different from a normal @code{:}
6188: definition. For example:
6189:
6190: @example
6191: : bar + - ; \ like foo but using : rather than my:
6192: see bar
6193: : bar
6194: + - ;
6195: see foo
6196: : foo
6197: 107645672 stats + - ;
6198:
6199: \ use ' foo . to show that 107645672 is the xt for foo
6200: @end example
6201:
6202: You can use techniques like this to make new defining words in terms of
6203: @i{any} existing defining word.
6204:
6205:
6206: @cindex defining defining words
6207: @cindex @code{CREATE} ... @code{DOES>}
6208: If you want the words defined with your defining words to behave
6209: differently from words defined with standard defining words, you can
6210: write your defining word like this:
6211:
6212: @example
6213: : def-word ( "name" -- )
6214: CREATE @i{code1}
6215: DOES> ( ... -- ... )
6216: @i{code2} ;
6217:
6218: def-word name
6219: @end example
6220:
6221: @cindex child words
6222: This fragment defines a @dfn{defining word} @code{def-word} and then
6223: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6224: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6225: is not executed at this time. The word @code{name} is sometimes called a
6226: @dfn{child} of @code{def-word}.
6227:
6228: When you execute @code{name}, the address of the body of @code{name} is
6229: put on the data stack and @i{code2} is executed (the address of the body
6230: of @code{name} is the address @code{HERE} returns immediately after the
6231: @code{CREATE}, i.e., the address a @code{create}d word returns by
6232: default).
6233:
6234: @c anton:
6235: @c www.dictionary.com says:
6236: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
6237: @c several generations of absence, usually caused by the chance
6238: @c recombination of genes. 2.An individual or a part that exhibits
6239: @c atavism. Also called throwback. 3.The return of a trait or recurrence
6240: @c of previous behavior after a period of absence.
6241: @c
6242: @c Doesn't seem to fit.
6243:
6244: @c @cindex atavism in child words
6245: You can use @code{def-word} to define a set of child words that behave
6246: similarly; they all have a common run-time behaviour determined by
6247: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
6248: body of the child word. The structure of the data is common to all
6249: children of @code{def-word}, but the data values are specific -- and
6250: private -- to each child word. When a child word is executed, the
6251: address of its private data area is passed as a parameter on TOS to be
6252: used and manipulated@footnote{It is legitimate both to read and write to
6253: this data area.} by @i{code2}.
6254:
6255: The two fragments of code that make up the defining words act (are
6256: executed) at two completely separate times:
6257:
6258: @itemize @bullet
6259: @item
6260: At @i{define time}, the defining word executes @i{code1} to generate a
6261: child word
6262: @item
6263: At @i{child execution time}, when a child word is invoked, @i{code2}
6264: is executed, using parameters (data) that are private and specific to
6265: the child word.
6266: @end itemize
6267:
6268: Another way of understanding the behaviour of @code{def-word} and
6269: @code{name} is to say that, if you make the following definitions:
6270: @example
6271: : def-word1 ( "name" -- )
6272: CREATE @i{code1} ;
6273:
6274: : action1 ( ... -- ... )
6275: @i{code2} ;
6276:
6277: def-word1 name1
6278: @end example
6279:
6280: @noindent
6281: Then using @code{name1 action1} is equivalent to using @code{name}.
6282:
6283: The classic example is that you can define @code{CONSTANT} in this way:
6284:
6285: @example
6286: : CONSTANT ( w "name" -- )
6287: CREATE ,
6288: DOES> ( -- w )
6289: @@ ;
6290: @end example
6291:
6292: @comment There is a beautiful description of how this works and what
6293: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6294: @comment commentary on the Counting Fruits problem.
6295:
6296: When you create a constant with @code{5 CONSTANT five}, a set of
6297: define-time actions take place; first a new word @code{five} is created,
6298: then the value 5 is laid down in the body of @code{five} with
6299: @code{,}. When @code{five} is executed, the address of the body is put on
6300: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6301: no code of its own; it simply contains a data field and a pointer to the
6302: code that follows @code{DOES>} in its defining word. That makes words
6303: created in this way very compact.
6304:
6305: The final example in this section is intended to remind you that space
6306: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6307: both read and written by a Standard program@footnote{Exercise: use this
6308: example as a starting point for your own implementation of @code{Value}
6309: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6310: @code{[']}.}:
6311:
6312: @example
6313: : foo ( "name" -- )
6314: CREATE -1 ,
6315: DOES> ( -- )
6316: @@ . ;
6317:
6318: foo first-word
6319: foo second-word
6320:
6321: 123 ' first-word >BODY !
6322: @end example
6323:
6324: If @code{first-word} had been a @code{CREATE}d word, we could simply
6325: have executed it to get the address of its data field. However, since it
6326: was defined to have @code{DOES>} actions, its execution semantics are to
6327: perform those @code{DOES>} actions. To get the address of its data field
6328: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6329: translate the xt into the address of the data field. When you execute
6330: @code{first-word}, it will display @code{123}. When you execute
6331: @code{second-word} it will display @code{-1}.
6332:
6333: @cindex stack effect of @code{DOES>}-parts
6334: @cindex @code{DOES>}-parts, stack effect
6335: In the examples above the stack comment after the @code{DOES>} specifies
6336: the stack effect of the defined words, not the stack effect of the
6337: following code (the following code expects the address of the body on
6338: the top of stack, which is not reflected in the stack comment). This is
6339: the convention that I use and recommend (it clashes a bit with using
6340: locals declarations for stack effect specification, though).
6341:
6342: @menu
6343: * CREATE..DOES> applications::
6344: * CREATE..DOES> details::
6345: * Advanced does> usage example::
6346: * Const-does>::
6347: @end menu
6348:
6349: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
6350: @subsubsection Applications of @code{CREATE..DOES>}
6351: @cindex @code{CREATE} ... @code{DOES>}, applications
6352:
6353: You may wonder how to use this feature. Here are some usage patterns:
6354:
6355: @cindex factoring similar colon definitions
6356: When you see a sequence of code occurring several times, and you can
6357: identify a meaning, you will factor it out as a colon definition. When
6358: you see similar colon definitions, you can factor them using
6359: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6360: that look very similar:
6361: @example
6362: : ori, ( reg-target reg-source n -- )
6363: 0 asm-reg-reg-imm ;
6364: : andi, ( reg-target reg-source n -- )
6365: 1 asm-reg-reg-imm ;
6366: @end example
6367:
6368: @noindent
6369: This could be factored with:
6370: @example
6371: : reg-reg-imm ( op-code -- )
6372: CREATE ,
6373: DOES> ( reg-target reg-source n -- )
6374: @@ asm-reg-reg-imm ;
6375:
6376: 0 reg-reg-imm ori,
6377: 1 reg-reg-imm andi,
6378: @end example
6379:
6380: @cindex currying
6381: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6382: supply a part of the parameters for a word (known as @dfn{currying} in
6383: the functional language community). E.g., @code{+} needs two
6384: parameters. Creating versions of @code{+} with one parameter fixed can
6385: be done like this:
6386:
6387: @example
6388: : curry+ ( n1 "name" -- )
6389: CREATE ,
6390: DOES> ( n2 -- n1+n2 )
6391: @@ + ;
6392:
6393: 3 curry+ 3+
6394: -2 curry+ 2-
6395: @end example
6396:
6397:
6398: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
6399: @subsubsection The gory details of @code{CREATE..DOES>}
6400: @cindex @code{CREATE} ... @code{DOES>}, details
6401:
6402: doc-does>
6403:
6404: @cindex @code{DOES>} in a separate definition
6405: This means that you need not use @code{CREATE} and @code{DOES>} in the
6406: same definition; you can put the @code{DOES>}-part in a separate
6407: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
6408: @example
6409: : does1
6410: DOES> ( ... -- ... )
6411: ... ;
6412:
6413: : does2
6414: DOES> ( ... -- ... )
6415: ... ;
6416:
6417: : def-word ( ... -- ... )
6418: create ...
6419: IF
6420: does1
6421: ELSE
6422: does2
6423: ENDIF ;
6424: @end example
6425:
6426: In this example, the selection of whether to use @code{does1} or
6427: @code{does2} is made at definition-time; at the time that the child word is
6428: @code{CREATE}d.
6429:
6430: @cindex @code{DOES>} in interpretation state
6431: In a standard program you can apply a @code{DOES>}-part only if the last
6432: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6433: will override the behaviour of the last word defined in any case. In a
6434: standard program, you can use @code{DOES>} only in a colon
6435: definition. In Gforth, you can also use it in interpretation state, in a
6436: kind of one-shot mode; for example:
6437: @example
6438: CREATE name ( ... -- ... )
6439: @i{initialization}
6440: DOES>
6441: @i{code} ;
6442: @end example
6443:
6444: @noindent
6445: is equivalent to the standard:
6446: @example
6447: :noname
6448: DOES>
6449: @i{code} ;
6450: CREATE name EXECUTE ( ... -- ... )
6451: @i{initialization}
6452: @end example
6453:
6454: doc->body
6455:
6456: @node Advanced does> usage example, Const-does>, CREATE..DOES> details, User-defined Defining Words
6457: @subsubsection Advanced does> usage example
6458:
6459: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6460: for disassembling instructions, that follow a very repetetive scheme:
6461:
6462: @example
6463: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6464: @var{entry-num} cells @var{table} + !
6465: @end example
6466:
6467: Of course, this inspires the idea to factor out the commonalities to
6468: allow a definition like
6469:
6470: @example
6471: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6472: @end example
6473:
6474: The parameters @var{disasm-operands} and @var{table} are usually
6475: correlated. Moreover, before I wrote the disassembler, there already
6476: existed code that defines instructions like this:
6477:
6478: @example
6479: @var{entry-num} @var{inst-format} @var{inst-name}
6480: @end example
6481:
6482: This code comes from the assembler and resides in
6483: @file{arch/mips/insts.fs}.
6484:
6485: So I had to define the @var{inst-format} words that performed the scheme
6486: above when executed. At first I chose to use run-time code-generation:
6487:
6488: @example
6489: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6490: :noname Postpone @var{disasm-operands}
6491: name Postpone sliteral Postpone type Postpone ;
6492: swap cells @var{table} + ! ;
6493: @end example
6494:
6495: Note that this supplies the other two parameters of the scheme above.
6496:
6497: An alternative would have been to write this using
6498: @code{create}/@code{does>}:
6499:
6500: @example
6501: : @var{inst-format} ( entry-num "name" -- )
6502: here name string, ( entry-num c-addr ) \ parse and save "name"
6503: noname create , ( entry-num )
6504: latestxt swap cells @var{table} + !
6505: does> ( addr w -- )
6506: \ disassemble instruction w at addr
6507: @@ >r
6508: @var{disasm-operands}
6509: r> count type ;
6510: @end example
6511:
6512: Somehow the first solution is simpler, mainly because it's simpler to
6513: shift a string from definition-time to use-time with @code{sliteral}
6514: than with @code{string,} and friends.
6515:
6516: I wrote a lot of words following this scheme and soon thought about
6517: factoring out the commonalities among them. Note that this uses a
6518: two-level defining word, i.e., a word that defines ordinary defining
6519: words.
6520:
6521: This time a solution involving @code{postpone} and friends seemed more
6522: difficult (try it as an exercise), so I decided to use a
6523: @code{create}/@code{does>} word; since I was already at it, I also used
6524: @code{create}/@code{does>} for the lower level (try using
6525: @code{postpone} etc. as an exercise), resulting in the following
6526: definition:
6527:
6528: @example
6529: : define-format ( disasm-xt table-xt -- )
6530: \ define an instruction format that uses disasm-xt for
6531: \ disassembling and enters the defined instructions into table
6532: \ table-xt
6533: create 2,
6534: does> ( u "inst" -- )
6535: \ defines an anonymous word for disassembling instruction inst,
6536: \ and enters it as u-th entry into table-xt
6537: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6538: noname create 2, \ define anonymous word
6539: execute latestxt swap ! \ enter xt of defined word into table-xt
6540: does> ( addr w -- )
6541: \ disassemble instruction w at addr
6542: 2@@ >r ( addr w disasm-xt R: c-addr )
6543: execute ( R: c-addr ) \ disassemble operands
6544: r> count type ; \ print name
6545: @end example
6546:
6547: Note that the tables here (in contrast to above) do the @code{cells +}
6548: by themselves (that's why you have to pass an xt). This word is used in
6549: the following way:
6550:
6551: @example
6552: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6553: @end example
6554:
6555: As shown above, the defined instruction format is then used like this:
6556:
6557: @example
6558: @var{entry-num} @var{inst-format} @var{inst-name}
6559: @end example
6560:
6561: In terms of currying, this kind of two-level defining word provides the
6562: parameters in three stages: first @var{disasm-operands} and @var{table},
6563: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6564: the instruction to be disassembled.
6565:
6566: Of course this did not quite fit all the instruction format names used
6567: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6568: the parameters into the right form.
6569:
6570: If you have trouble following this section, don't worry. First, this is
6571: involved and takes time (and probably some playing around) to
6572: understand; second, this is the first two-level
6573: @code{create}/@code{does>} word I have written in seventeen years of
6574: Forth; and if I did not have @file{insts.fs} to start with, I may well
6575: have elected to use just a one-level defining word (with some repeating
6576: of parameters when using the defining word). So it is not necessary to
6577: understand this, but it may improve your understanding of Forth.
6578:
6579:
6580: @node Const-does>, , Advanced does> usage example, User-defined Defining Words
6581: @subsubsection @code{Const-does>}
6582:
6583: A frequent use of @code{create}...@code{does>} is for transferring some
6584: values from definition-time to run-time. Gforth supports this use with
6585:
6586: doc-const-does>
6587:
6588: A typical use of this word is:
6589:
6590: @example
6591: : curry+ ( n1 "name" -- )
6592: 1 0 CONST-DOES> ( n2 -- n1+n2 )
6593: + ;
6594:
6595: 3 curry+ 3+
6596: @end example
6597:
6598: Here the @code{1 0} means that 1 cell and 0 floats are transferred from
6599: definition to run-time.
6600:
6601: The advantages of using @code{const-does>} are:
6602:
6603: @itemize
6604:
6605: @item
6606: You don't have to deal with storing and retrieving the values, i.e.,
6607: your program becomes more writable and readable.
6608:
6609: @item
6610: When using @code{does>}, you have to introduce a @code{@@} that cannot
6611: be optimized away (because you could change the data using
6612: @code{>body}...@code{!}); @code{const-does>} avoids this problem.
6613:
6614: @end itemize
6615:
6616: An ANS Forth implementation of @code{const-does>} is available in
6617: @file{compat/const-does.fs}.
6618:
6619:
6620: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6621: @subsection Deferred words
6622: @cindex deferred words
6623:
6624: The defining word @code{Defer} allows you to define a word by name
6625: without defining its behaviour; the definition of its behaviour is
6626: deferred. Here are two situation where this can be useful:
6627:
6628: @itemize @bullet
6629: @item
6630: Where you want to allow the behaviour of a word to be altered later, and
6631: for all precompiled references to the word to change when its behaviour
6632: is changed.
6633: @item
6634: For mutual recursion; @xref{Calls and returns}.
6635: @end itemize
6636:
6637: In the following example, @code{foo} always invokes the version of
6638: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6639: always invokes the version that prints ``@code{Hello}''. There is no way
6640: of getting @code{foo} to use the later version without re-ordering the
6641: source code and recompiling it.
6642:
6643: @example
6644: : greet ." Good morning" ;
6645: : foo ... greet ... ;
6646: : greet ." Hello" ;
6647: : bar ... greet ... ;
6648: @end example
6649:
6650: This problem can be solved by defining @code{greet} as a @code{Defer}red
6651: word. The behaviour of a @code{Defer}red word can be defined and
6652: redefined at any time by using @code{IS} to associate the xt of a
6653: previously-defined word with it. The previous example becomes:
6654:
6655: @example
6656: Defer greet ( -- )
6657: : foo ... greet ... ;
6658: : bar ... greet ... ;
6659: : greet1 ( -- ) ." Good morning" ;
6660: : greet2 ( -- ) ." Hello" ;
6661: ' greet2 IS greet \ make greet behave like greet2
6662: @end example
6663:
6664: @progstyle
6665: You should write a stack comment for every deferred word, and put only
6666: XTs into deferred words that conform to this stack effect. Otherwise
6667: it's too difficult to use the deferred word.
6668:
6669: A deferred word can be used to improve the statistics-gathering example
6670: from @ref{User-defined Defining Words}; rather than edit the
6671: application's source code to change every @code{:} to a @code{my:}, do
6672: this:
6673:
6674: @example
6675: : real: : ; \ retain access to the original
6676: defer : \ redefine as a deferred word
6677: ' my: IS : \ use special version of :
6678: \
6679: \ load application here
6680: \
6681: ' real: IS : \ go back to the original
6682: @end example
6683:
6684:
6685: One thing to note is that @code{IS} has special compilation semantics,
6686: such that it parses the name at compile time (like @code{TO}):
6687:
6688: @example
6689: : set-greet ( xt -- )
6690: IS greet ;
6691:
6692: ' greet1 set-greet
6693: @end example
6694:
6695: In situations where @code{IS} does not fit, use @code{defer!} instead.
6696:
6697: A deferred word can only inherit execution semantics from the xt
6698: (because that is all that an xt can represent -- for more discussion of
6699: this @pxref{Tokens for Words}); by default it will have default
6700: interpretation and compilation semantics deriving from this execution
6701: semantics. However, you can change the interpretation and compilation
6702: semantics of the deferred word in the usual ways:
6703:
6704: @example
6705: : bar .... ; immediate
6706: Defer fred immediate
6707: Defer jim
6708:
6709: ' bar IS jim \ jim has default semantics
6710: ' bar IS fred \ fred is immediate
6711: @end example
6712:
6713: doc-defer
6714: doc-defer!
6715: doc-is
6716: doc-defer@
6717: doc-action-of
6718: @comment TODO document these: what's defers [is]
6719: doc-defers
6720:
6721: @c Use @code{words-deferred} to see a list of deferred words.
6722:
6723: Definitions of these words (except @code{defers}) in ANS Forth are
6724: provided in @file{compat/defer.fs}.
6725:
6726:
6727: @node Aliases, , Deferred words, Defining Words
6728: @subsection Aliases
6729: @cindex aliases
6730:
6731: The defining word @code{Alias} allows you to define a word by name that
6732: has the same behaviour as some other word. Here are two situation where
6733: this can be useful:
6734:
6735: @itemize @bullet
6736: @item
6737: When you want access to a word's definition from a different word list
6738: (for an example of this, see the definition of the @code{Root} word list
6739: in the Gforth source).
6740: @item
6741: When you want to create a synonym; a definition that can be known by
6742: either of two names (for example, @code{THEN} and @code{ENDIF} are
6743: aliases).
6744: @end itemize
6745:
6746: Like deferred words, an alias has default compilation and interpretation
6747: semantics at the beginning (not the modifications of the other word),
6748: but you can change them in the usual ways (@code{immediate},
6749: @code{compile-only}). For example:
6750:
6751: @example
6752: : foo ... ; immediate
6753:
6754: ' foo Alias bar \ bar is not an immediate word
6755: ' foo Alias fooby immediate \ fooby is an immediate word
6756: @end example
6757:
6758: Words that are aliases have the same xt, different headers in the
6759: dictionary, and consequently different name tokens (@pxref{Tokens for
6760: Words}) and possibly different immediate flags. An alias can only have
6761: default or immediate compilation semantics; you can define aliases for
6762: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
6763:
6764: doc-alias
6765:
6766:
6767: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6768: @section Interpretation and Compilation Semantics
6769: @cindex semantics, interpretation and compilation
6770:
6771: @c !! state and ' are used without explanation
6772: @c example for immediate/compile-only? or is the tutorial enough
6773:
6774: @cindex interpretation semantics
6775: The @dfn{interpretation semantics} of a (named) word are what the text
6776: interpreter does when it encounters the word in interpret state. It also
6777: appears in some other contexts, e.g., the execution token returned by
6778: @code{' @i{word}} identifies the interpretation semantics of @i{word}
6779: (in other words, @code{' @i{word} execute} is equivalent to
6780: interpret-state text interpretation of @code{@i{word}}).
6781:
6782: @cindex compilation semantics
6783: The @dfn{compilation semantics} of a (named) word are what the text
6784: interpreter does when it encounters the word in compile state. It also
6785: appears in other contexts, e.g, @code{POSTPONE @i{word}}
6786: compiles@footnote{In standard terminology, ``appends to the current
6787: definition''.} the compilation semantics of @i{word}.
6788:
6789: @cindex execution semantics
6790: The standard also talks about @dfn{execution semantics}. They are used
6791: only for defining the interpretation and compilation semantics of many
6792: words. By default, the interpretation semantics of a word are to
6793: @code{execute} its execution semantics, and the compilation semantics of
6794: a word are to @code{compile,} its execution semantics.@footnote{In
6795: standard terminology: The default interpretation semantics are its
6796: execution semantics; the default compilation semantics are to append its
6797: execution semantics to the execution semantics of the current
6798: definition.}
6799:
6800: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
6801: the text interpreter, ticked, or @code{postpone}d, so they have no
6802: interpretation or compilation semantics. Their behaviour is represented
6803: by their XT (@pxref{Tokens for Words}), and we call it execution
6804: semantics, too.
6805:
6806: @comment TODO expand, make it co-operate with new sections on text interpreter.
6807:
6808: @cindex immediate words
6809: @cindex compile-only words
6810: You can change the semantics of the most-recently defined word:
6811:
6812:
6813: doc-immediate
6814: doc-compile-only
6815: doc-restrict
6816:
6817: By convention, words with non-default compilation semantics (e.g.,
6818: immediate words) often have names surrounded with brackets (e.g.,
6819: @code{[']}, @pxref{Execution token}).
6820:
6821: Note that ticking (@code{'}) a compile-only word gives an error
6822: (``Interpreting a compile-only word'').
6823:
6824: @menu
6825: * Combined words::
6826: @end menu
6827:
6828:
6829: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
6830: @subsection Combined Words
6831: @cindex combined words
6832:
6833: Gforth allows you to define @dfn{combined words} -- words that have an
6834: arbitrary combination of interpretation and compilation semantics.
6835:
6836: doc-interpret/compile:
6837:
6838: This feature was introduced for implementing @code{TO} and @code{S"}. I
6839: recommend that you do not define such words, as cute as they may be:
6840: they make it hard to get at both parts of the word in some contexts.
6841: E.g., assume you want to get an execution token for the compilation
6842: part. Instead, define two words, one that embodies the interpretation
6843: part, and one that embodies the compilation part. Once you have done
6844: that, you can define a combined word with @code{interpret/compile:} for
6845: the convenience of your users.
6846:
6847: You might try to use this feature to provide an optimizing
6848: implementation of the default compilation semantics of a word. For
6849: example, by defining:
6850: @example
6851: :noname
6852: foo bar ;
6853: :noname
6854: POSTPONE foo POSTPONE bar ;
6855: interpret/compile: opti-foobar
6856: @end example
6857:
6858: @noindent
6859: as an optimizing version of:
6860:
6861: @example
6862: : foobar
6863: foo bar ;
6864: @end example
6865:
6866: Unfortunately, this does not work correctly with @code{[compile]},
6867: because @code{[compile]} assumes that the compilation semantics of all
6868: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
6869: opti-foobar} would compile compilation semantics, whereas
6870: @code{[compile] foobar} would compile interpretation semantics.
6871:
6872: @cindex state-smart words (are a bad idea)
6873: @anchor{state-smartness}
6874: Some people try to use @dfn{state-smart} words to emulate the feature provided
6875: by @code{interpret/compile:} (words are state-smart if they check
6876: @code{STATE} during execution). E.g., they would try to code
6877: @code{foobar} like this:
6878:
6879: @example
6880: : foobar
6881: STATE @@
6882: IF ( compilation state )
6883: POSTPONE foo POSTPONE bar
6884: ELSE
6885: foo bar
6886: ENDIF ; immediate
6887: @end example
6888:
6889: Although this works if @code{foobar} is only processed by the text
6890: interpreter, it does not work in other contexts (like @code{'} or
6891: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
6892: for a state-smart word, not for the interpretation semantics of the
6893: original @code{foobar}; when you execute this execution token (directly
6894: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
6895: state, the result will not be what you expected (i.e., it will not
6896: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
6897: write them@footnote{For a more detailed discussion of this topic, see
6898: M. Anton Ertl,
6899: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
6900: it is Evil and How to Exorcise it}}, EuroForth '98.}!
6901:
6902: @cindex defining words with arbitrary semantics combinations
6903: It is also possible to write defining words that define words with
6904: arbitrary combinations of interpretation and compilation semantics. In
6905: general, they look like this:
6906:
6907: @example
6908: : def-word
6909: create-interpret/compile
6910: @i{code1}
6911: interpretation>
6912: @i{code2}
6913: <interpretation
6914: compilation>
6915: @i{code3}
6916: <compilation ;
6917: @end example
6918:
6919: For a @i{word} defined with @code{def-word}, the interpretation
6920: semantics are to push the address of the body of @i{word} and perform
6921: @i{code2}, and the compilation semantics are to push the address of
6922: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
6923: can also be defined like this (except that the defined constants don't
6924: behave correctly when @code{[compile]}d):
6925:
6926: @example
6927: : constant ( n "name" -- )
6928: create-interpret/compile
6929: ,
6930: interpretation> ( -- n )
6931: @@
6932: <interpretation
6933: compilation> ( compilation. -- ; run-time. -- n )
6934: @@ postpone literal
6935: <compilation ;
6936: @end example
6937:
6938:
6939: doc-create-interpret/compile
6940: doc-interpretation>
6941: doc-<interpretation
6942: doc-compilation>
6943: doc-<compilation
6944:
6945:
6946: Words defined with @code{interpret/compile:} and
6947: @code{create-interpret/compile} have an extended header structure that
6948: differs from other words; however, unless you try to access them with
6949: plain address arithmetic, you should not notice this. Words for
6950: accessing the header structure usually know how to deal with this; e.g.,
6951: @code{'} @i{word} @code{>body} also gives you the body of a word created
6952: with @code{create-interpret/compile}.
6953:
6954:
6955: @c -------------------------------------------------------------
6956: @node Tokens for Words, Compiling words, Interpretation and Compilation Semantics, Words
6957: @section Tokens for Words
6958: @cindex tokens for words
6959:
6960: This section describes the creation and use of tokens that represent
6961: words.
6962:
6963: @menu
6964: * Execution token:: represents execution/interpretation semantics
6965: * Compilation token:: represents compilation semantics
6966: * Name token:: represents named words
6967: @end menu
6968:
6969: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
6970: @subsection Execution token
6971:
6972: @cindex xt
6973: @cindex execution token
6974: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
6975: You can use @code{execute} to invoke this behaviour.
6976:
6977: @cindex tick (')
6978: You can use @code{'} to get an execution token that represents the
6979: interpretation semantics of a named word:
6980:
6981: @example
6982: 5 ' . ( n xt )
6983: execute ( ) \ execute the xt (i.e., ".")
6984: @end example
6985:
6986: doc-'
6987:
6988: @code{'} parses at run-time; there is also a word @code{[']} that parses
6989: when it is compiled, and compiles the resulting XT:
6990:
6991: @example
6992: : foo ['] . execute ;
6993: 5 foo
6994: : bar ' execute ; \ by contrast,
6995: 5 bar . \ ' parses "." when bar executes
6996: @end example
6997:
6998: doc-[']
6999:
7000: If you want the execution token of @i{word}, write @code{['] @i{word}}
7001: in compiled code and @code{' @i{word}} in interpreted code. Gforth's
7002: @code{'} and @code{[']} behave somewhat unusually by complaining about
7003: compile-only words (because these words have no interpretation
7004: semantics). You might get what you want by using @code{COMP' @i{word}
7005: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
7006: token}).
7007:
7008: Another way to get an XT is @code{:noname} or @code{latestxt}
7009: (@pxref{Anonymous Definitions}). For anonymous words this gives an xt
7010: for the only behaviour the word has (the execution semantics). For
7011: named words, @code{latestxt} produces an XT for the same behaviour it
7012: would produce if the word was defined anonymously.
7013:
7014: @example
7015: :noname ." hello" ;
7016: execute
7017: @end example
7018:
7019: An XT occupies one cell and can be manipulated like any other cell.
7020:
7021: @cindex code field address
7022: @cindex CFA
7023: In ANS Forth the XT is just an abstract data type (i.e., defined by the
7024: operations that produce or consume it). For old hands: In Gforth, the
7025: XT is implemented as a code field address (CFA).
7026:
7027: doc-execute
7028: doc-perform
7029:
7030: @node Compilation token, Name token, Execution token, Tokens for Words
7031: @subsection Compilation token
7032:
7033: @cindex compilation token
7034: @cindex CT (compilation token)
7035: Gforth represents the compilation semantics of a named word by a
7036: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
7037: @i{xt} is an execution token. The compilation semantics represented by
7038: the compilation token can be performed with @code{execute}, which
7039: consumes the whole compilation token, with an additional stack effect
7040: determined by the represented compilation semantics.
7041:
7042: At present, the @i{w} part of a compilation token is an execution token,
7043: and the @i{xt} part represents either @code{execute} or
7044: @code{compile,}@footnote{Depending upon the compilation semantics of the
7045: word. If the word has default compilation semantics, the @i{xt} will
7046: represent @code{compile,}. Otherwise (e.g., for immediate words), the
7047: @i{xt} will represent @code{execute}.}. However, don't rely on that
7048: knowledge, unless necessary; future versions of Gforth may introduce
7049: unusual compilation tokens (e.g., a compilation token that represents
7050: the compilation semantics of a literal).
7051:
7052: You can perform the compilation semantics represented by the compilation
7053: token with @code{execute}. You can compile the compilation semantics
7054: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
7055: equivalent to @code{postpone @i{word}}.
7056:
7057: doc-[comp']
7058: doc-comp'
7059: doc-postpone,
7060:
7061: @node Name token, , Compilation token, Tokens for Words
7062: @subsection Name token
7063:
7064: @cindex name token
7065: Gforth represents named words by the @dfn{name token}, (@i{nt}). Name
7066: token is an abstract data type that occurs as argument or result of the
7067: words below.
7068:
7069: @c !! put this elswhere?
7070: @cindex name field address
7071: @cindex NFA
7072: The closest thing to the nt in older Forth systems is the name field
7073: address (NFA), but there are significant differences: in older Forth
7074: systems each word had a unique NFA, LFA, CFA and PFA (in this order, or
7075: LFA, NFA, CFA, PFA) and there were words for getting from one to the
7076: next. In contrast, in Gforth 0@dots{}n nts correspond to one xt; there
7077: is a link field in the structure identified by the name token, but
7078: searching usually uses a hash table external to these structures; the
7079: name in Gforth has a cell-wide count-and-flags field, and the nt is not
7080: implemented as the address of that count field.
7081:
7082: doc-find-name
7083: doc-latest
7084: doc->name
7085: doc-name>int
7086: doc-name?int
7087: doc-name>comp
7088: doc-name>string
7089: doc-id.
7090: doc-.name
7091: doc-.id
7092:
7093: @c ----------------------------------------------------------
7094: @node Compiling words, The Text Interpreter, Tokens for Words, Words
7095: @section Compiling words
7096: @cindex compiling words
7097: @cindex macros
7098:
7099: In contrast to most other languages, Forth has no strict boundary
7100: between compilation and run-time. E.g., you can run arbitrary code
7101: between defining words (or for computing data used by defining words
7102: like @code{constant}). Moreover, @code{Immediate} (@pxref{Interpretation
7103: and Compilation Semantics} and @code{[}...@code{]} (see below) allow
7104: running arbitrary code while compiling a colon definition (exception:
7105: you must not allot dictionary space).
7106:
7107: @menu
7108: * Literals:: Compiling data values
7109: * Macros:: Compiling words
7110: @end menu
7111:
7112: @node Literals, Macros, Compiling words, Compiling words
7113: @subsection Literals
7114: @cindex Literals
7115:
7116: The simplest and most frequent example is to compute a literal during
7117: compilation. E.g., the following definition prints an array of strings,
7118: one string per line:
7119:
7120: @example
7121: : .strings ( addr u -- ) \ gforth
7122: 2* cells bounds U+DO
7123: cr i 2@@ type
7124: 2 cells +LOOP ;
7125: @end example
7126:
7127: With a simple-minded compiler like Gforth's, this computes @code{2
7128: cells} on every loop iteration. You can compute this value once and for
7129: all at compile time and compile it into the definition like this:
7130:
7131: @example
7132: : .strings ( addr u -- ) \ gforth
7133: 2* cells bounds U+DO
7134: cr i 2@@ type
7135: [ 2 cells ] literal +LOOP ;
7136: @end example
7137:
7138: @code{[} switches the text interpreter to interpret state (you will get
7139: an @code{ok} prompt if you type this example interactively and insert a
7140: newline between @code{[} and @code{]}), so it performs the
7141: interpretation semantics of @code{2 cells}; this computes a number.
7142: @code{]} switches the text interpreter back into compile state. It then
7143: performs @code{Literal}'s compilation semantics, which are to compile
7144: this number into the current word. You can decompile the word with
7145: @code{see .strings} to see the effect on the compiled code.
7146:
7147: You can also optimize the @code{2* cells} into @code{[ 2 cells ] literal
7148: *} in this way.
7149:
7150: doc-[
7151: doc-]
7152: doc-literal
7153: doc-]L
7154:
7155: There are also words for compiling other data types than single cells as
7156: literals:
7157:
7158: doc-2literal
7159: doc-fliteral
7160: doc-sliteral
7161:
7162: @cindex colon-sys, passing data across @code{:}
7163: @cindex @code{:}, passing data across
7164: You might be tempted to pass data from outside a colon definition to the
7165: inside on the data stack. This does not work, because @code{:} puhes a
7166: colon-sys, making stuff below unaccessible. E.g., this does not work:
7167:
7168: @example
7169: 5 : foo literal ; \ error: "unstructured"
7170: @end example
7171:
7172: Instead, you have to pass the value in some other way, e.g., through a
7173: variable:
7174:
7175: @example
7176: variable temp
7177: 5 temp !
7178: : foo [ temp @@ ] literal ;
7179: @end example
7180:
7181:
7182: @node Macros, , Literals, Compiling words
7183: @subsection Macros
7184: @cindex Macros
7185: @cindex compiling compilation semantics
7186:
7187: @code{Literal} and friends compile data values into the current
7188: definition. You can also write words that compile other words into the
7189: current definition. E.g.,
7190:
7191: @example
7192: : compile-+ ( -- ) \ compiled code: ( n1 n2 -- n )
7193: POSTPONE + ;
7194:
7195: : foo ( n1 n2 -- n )
7196: [ compile-+ ] ;
7197: 1 2 foo .
7198: @end example
7199:
7200: This is equivalent to @code{: foo + ;} (@code{see foo} to check this).
7201: What happens in this example? @code{Postpone} compiles the compilation
7202: semantics of @code{+} into @code{compile-+}; later the text interpreter
7203: executes @code{compile-+} and thus the compilation semantics of +, which
7204: compile (the execution semantics of) @code{+} into
7205: @code{foo}.@footnote{A recent RFI answer requires that compiling words
7206: should only be executed in compile state, so this example is not
7207: guaranteed to work on all standard systems, but on any decent system it
7208: will work.}
7209:
7210: doc-postpone
7211: doc-[compile]
7212:
7213: Compiling words like @code{compile-+} are usually immediate (or similar)
7214: so you do not have to switch to interpret state to execute them;
7215: mopifying the last example accordingly produces:
7216:
7217: @example
7218: : [compile-+] ( compilation: --; interpretation: -- )
7219: \ compiled code: ( n1 n2 -- n )
7220: POSTPONE + ; immediate
7221:
7222: : foo ( n1 n2 -- n )
7223: [compile-+] ;
7224: 1 2 foo .
7225: @end example
7226:
7227: Immediate compiling words are similar to macros in other languages (in
7228: particular, Lisp). The important differences to macros in, e.g., C are:
7229:
7230: @itemize @bullet
7231:
7232: @item
7233: You use the same language for defining and processing macros, not a
7234: separate preprocessing language and processor.
7235:
7236: @item
7237: Consequently, the full power of Forth is available in macro definitions.
7238: E.g., you can perform arbitrarily complex computations, or generate
7239: different code conditionally or in a loop (e.g., @pxref{Advanced macros
7240: Tutorial}). This power is very useful when writing a parser generators
7241: or other code-generating software.
7242:
7243: @item
7244: Macros defined using @code{postpone} etc. deal with the language at a
7245: higher level than strings; name binding happens at macro definition
7246: time, so you can avoid the pitfalls of name collisions that can happen
7247: in C macros. Of course, Forth is a liberal language and also allows to
7248: shoot yourself in the foot with text-interpreted macros like
7249:
7250: @example
7251: : [compile-+] s" +" evaluate ; immediate
7252: @end example
7253:
7254: Apart from binding the name at macro use time, using @code{evaluate}
7255: also makes your definition @code{state}-smart (@pxref{state-smartness}).
7256: @end itemize
7257:
7258: You may want the macro to compile a number into a word. The word to do
7259: it is @code{literal}, but you have to @code{postpone} it, so its
7260: compilation semantics take effect when the macro is executed, not when
7261: it is compiled:
7262:
7263: @example
7264: : [compile-5] ( -- ) \ compiled code: ( -- n )
7265: 5 POSTPONE literal ; immediate
7266:
7267: : foo [compile-5] ;
7268: foo .
7269: @end example
7270:
7271: You may want to pass parameters to a macro, that the macro should
7272: compile into the current definition. If the parameter is a number, then
7273: you can use @code{postpone literal} (similar for other values).
7274:
7275: If you want to pass a word that is to be compiled, the usual way is to
7276: pass an execution token and @code{compile,} it:
7277:
7278: @example
7279: : twice1 ( xt -- ) \ compiled code: ... -- ...
7280: dup compile, compile, ;
7281:
7282: : 2+ ( n1 -- n2 )
7283: [ ' 1+ twice1 ] ;
7284: @end example
7285:
7286: doc-compile,
7287:
7288: An alternative available in Gforth, that allows you to pass compile-only
7289: words as parameters is to use the compilation token (@pxref{Compilation
7290: token}). The same example in this technique:
7291:
7292: @example
7293: : twice ( ... ct -- ... ) \ compiled code: ... -- ...
7294: 2dup 2>r execute 2r> execute ;
7295:
7296: : 2+ ( n1 -- n2 )
7297: [ comp' 1+ twice ] ;
7298: @end example
7299:
7300: In the example above @code{2>r} and @code{2r>} ensure that @code{twice}
7301: works even if the executed compilation semantics has an effect on the
7302: data stack.
7303:
7304: You can also define complete definitions with these words; this provides
7305: an alternative to using @code{does>} (@pxref{User-defined Defining
7306: Words}). E.g., instead of
7307:
7308: @example
7309: : curry+ ( n1 "name" -- )
7310: CREATE ,
7311: DOES> ( n2 -- n1+n2 )
7312: @@ + ;
7313: @end example
7314:
7315: you could define
7316:
7317: @example
7318: : curry+ ( n1 "name" -- )
7319: \ name execution: ( n2 -- n1+n2 )
7320: >r : r> POSTPONE literal POSTPONE + POSTPONE ; ;
7321:
7322: -3 curry+ 3-
7323: see 3-
7324: @end example
7325:
7326: The sequence @code{>r : r>} is necessary, because @code{:} puts a
7327: colon-sys on the data stack that makes everything below it unaccessible.
7328:
7329: This way of writing defining words is sometimes more, sometimes less
7330: convenient than using @code{does>} (@pxref{Advanced does> usage
7331: example}). One advantage of this method is that it can be optimized
7332: better, because the compiler knows that the value compiled with
7333: @code{literal} is fixed, whereas the data associated with a
7334: @code{create}d word can be changed.
7335:
7336: @c ----------------------------------------------------------
7337: @node The Text Interpreter, The Input Stream, Compiling words, Words
7338: @section The Text Interpreter
7339: @cindex interpreter - outer
7340: @cindex text interpreter
7341: @cindex outer interpreter
7342:
7343: @c Should we really describe all these ugly details? IMO the text
7344: @c interpreter should be much cleaner, but that may not be possible within
7345: @c ANS Forth. - anton
7346: @c nac-> I wanted to explain how it works to show how you can exploit
7347: @c it in your own programs. When I was writing a cross-compiler, figuring out
7348: @c some of these gory details was very helpful to me. None of the textbooks
7349: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
7350: @c seems to positively avoid going into too much detail for some of
7351: @c the internals.
7352:
7353: @c anton: ok. I wonder, though, if this is the right place; for some stuff
7354: @c it is; for the ugly details, I would prefer another place. I wonder
7355: @c whether we should have a chapter before "Words" that describes some
7356: @c basic concepts referred to in words, and a chapter after "Words" that
7357: @c describes implementation details.
7358:
7359: The text interpreter@footnote{This is an expanded version of the
7360: material in @ref{Introducing the Text Interpreter}.} is an endless loop
7361: that processes input from the current input device. It is also called
7362: the outer interpreter, in contrast to the inner interpreter
7363: (@pxref{Engine}) which executes the compiled Forth code on interpretive
7364: implementations.
7365:
7366: @cindex interpret state
7367: @cindex compile state
7368: The text interpreter operates in one of two states: @dfn{interpret
7369: state} and @dfn{compile state}. The current state is defined by the
7370: aptly-named variable @code{state}.
7371:
7372: This section starts by describing how the text interpreter behaves when
7373: it is in interpret state, processing input from the user input device --
7374: the keyboard. This is the mode that a Forth system is in after it starts
7375: up.
7376:
7377: @cindex input buffer
7378: @cindex terminal input buffer
7379: The text interpreter works from an area of memory called the @dfn{input
7380: buffer}@footnote{When the text interpreter is processing input from the
7381: keyboard, this area of memory is called the @dfn{terminal input buffer}
7382: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7383: @code{#TIB}.}, which stores your keyboard input when you press the
7384: @key{RET} key. Starting at the beginning of the input buffer, it skips
7385: leading spaces (called @dfn{delimiters}) then parses a string (a
7386: sequence of non-space characters) until it reaches either a space
7387: character or the end of the buffer. Having parsed a string, it makes two
7388: attempts to process it:
7389:
7390: @cindex dictionary
7391: @itemize @bullet
7392: @item
7393: It looks for the string in a @dfn{dictionary} of definitions. If the
7394: string is found, the string names a @dfn{definition} (also known as a
7395: @dfn{word}) and the dictionary search returns information that allows
7396: the text interpreter to perform the word's @dfn{interpretation
7397: semantics}. In most cases, this simply means that the word will be
7398: executed.
7399: @item
7400: If the string is not found in the dictionary, the text interpreter
7401: attempts to treat it as a number, using the rules described in
7402: @ref{Number Conversion}. If the string represents a legal number in the
7403: current radix, the number is pushed onto a parameter stack (the data
7404: stack for integers, the floating-point stack for floating-point
7405: numbers).
7406: @end itemize
7407:
7408: If both attempts fail, or if the word is found in the dictionary but has
7409: no interpretation semantics@footnote{This happens if the word was
7410: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7411: remainder of the input buffer, issues an error message and waits for
7412: more input. If one of the attempts succeeds, the text interpreter
7413: repeats the parsing process until the whole of the input buffer has been
7414: processed, at which point it prints the status message ``@code{ ok}''
7415: and waits for more input.
7416:
7417: @c anton: this should be in the input stream subsection (or below it)
7418:
7419: @cindex parse area
7420: The text interpreter keeps track of its position in the input buffer by
7421: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7422: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7423: of the input buffer. The region from offset @code{>IN @@} to the end of
7424: the input buffer is called the @dfn{parse area}@footnote{In other words,
7425: the text interpreter processes the contents of the input buffer by
7426: parsing strings from the parse area until the parse area is empty.}.
7427: This example shows how @code{>IN} changes as the text interpreter parses
7428: the input buffer:
7429:
7430: @example
7431: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7432: CR ." ->" TYPE ." <-" ; IMMEDIATE
7433:
7434: 1 2 3 remaining + remaining .
7435:
7436: : foo 1 2 3 remaining SWAP remaining ;
7437: @end example
7438:
7439: @noindent
7440: The result is:
7441:
7442: @example
7443: ->+ remaining .<-
7444: ->.<-5 ok
7445:
7446: ->SWAP remaining ;-<
7447: ->;<- ok
7448: @end example
7449:
7450: @cindex parsing words
7451: The value of @code{>IN} can also be modified by a word in the input
7452: buffer that is executed by the text interpreter. This means that a word
7453: can ``trick'' the text interpreter into either skipping a section of the
7454: input buffer@footnote{This is how parsing words work.} or into parsing a
7455: section twice. For example:
7456:
7457: @example
7458: : lat ." <<foo>>" ;
7459: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
7460: @end example
7461:
7462: @noindent
7463: When @code{flat} is executed, this output is produced@footnote{Exercise
7464: for the reader: what would happen if the @code{3} were replaced with
7465: @code{4}?}:
7466:
7467: @example
7468: <<bar>><<foo>>
7469: @end example
7470:
7471: This technique can be used to work around some of the interoperability
7472: problems of parsing words. Of course, it's better to avoid parsing
7473: words where possible.
7474:
7475: @noindent
7476: Two important notes about the behaviour of the text interpreter:
7477:
7478: @itemize @bullet
7479: @item
7480: It processes each input string to completion before parsing additional
7481: characters from the input buffer.
7482: @item
7483: It treats the input buffer as a read-only region (and so must your code).
7484: @end itemize
7485:
7486: @noindent
7487: When the text interpreter is in compile state, its behaviour changes in
7488: these ways:
7489:
7490: @itemize @bullet
7491: @item
7492: If a parsed string is found in the dictionary, the text interpreter will
7493: perform the word's @dfn{compilation semantics}. In most cases, this
7494: simply means that the execution semantics of the word will be appended
7495: to the current definition.
7496: @item
7497: When a number is encountered, it is compiled into the current definition
7498: (as a literal) rather than being pushed onto a parameter stack.
7499: @item
7500: If an error occurs, @code{state} is modified to put the text interpreter
7501: back into interpret state.
7502: @item
7503: Each time a line is entered from the keyboard, Gforth prints
7504: ``@code{ compiled}'' rather than `` @code{ok}''.
7505: @end itemize
7506:
7507: @cindex text interpreter - input sources
7508: When the text interpreter is using an input device other than the
7509: keyboard, its behaviour changes in these ways:
7510:
7511: @itemize @bullet
7512: @item
7513: When the parse area is empty, the text interpreter attempts to refill
7514: the input buffer from the input source. When the input source is
7515: exhausted, the input source is set back to the previous input source.
7516: @item
7517: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7518: time the parse area is emptied.
7519: @item
7520: If an error occurs, the input source is set back to the user input
7521: device.
7522: @end itemize
7523:
7524: You can read about this in more detail in @ref{Input Sources}.
7525:
7526: doc->in
7527: doc-source
7528:
7529: doc-tib
7530: doc-#tib
7531:
7532:
7533: @menu
7534: * Input Sources::
7535: * Number Conversion::
7536: * Interpret/Compile states::
7537: * Interpreter Directives::
7538: @end menu
7539:
7540: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7541: @subsection Input Sources
7542: @cindex input sources
7543: @cindex text interpreter - input sources
7544:
7545: By default, the text interpreter processes input from the user input
7546: device (the keyboard) when Forth starts up. The text interpreter can
7547: process input from any of these sources:
7548:
7549: @itemize @bullet
7550: @item
7551: The user input device -- the keyboard.
7552: @item
7553: A file, using the words described in @ref{Forth source files}.
7554: @item
7555: A block, using the words described in @ref{Blocks}.
7556: @item
7557: A text string, using @code{evaluate}.
7558: @end itemize
7559:
7560: A program can identify the current input device from the values of
7561: @code{source-id} and @code{blk}.
7562:
7563:
7564: doc-source-id
7565: doc-blk
7566:
7567: doc-save-input
7568: doc-restore-input
7569:
7570: doc-evaluate
7571: doc-query
7572:
7573:
7574:
7575: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
7576: @subsection Number Conversion
7577: @cindex number conversion
7578: @cindex double-cell numbers, input format
7579: @cindex input format for double-cell numbers
7580: @cindex single-cell numbers, input format
7581: @cindex input format for single-cell numbers
7582: @cindex floating-point numbers, input format
7583: @cindex input format for floating-point numbers
7584:
7585: This section describes the rules that the text interpreter uses when it
7586: tries to convert a string into a number.
7587:
7588: Let <digit> represent any character that is a legal digit in the current
7589: number base@footnote{For example, 0-9 when the number base is decimal or
7590: 0-9, A-F when the number base is hexadecimal.}.
7591:
7592: Let <decimal digit> represent any character in the range 0-9.
7593:
7594: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7595: in the braces (@i{a} or @i{b} or neither).
7596:
7597: Let * represent any number of instances of the previous character
7598: (including none).
7599:
7600: Let any other character represent itself.
7601:
7602: @noindent
7603: Now, the conversion rules are:
7604:
7605: @itemize @bullet
7606: @item
7607: A string of the form <digit><digit>* is treated as a single-precision
7608: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
7609: @item
7610: A string of the form -<digit><digit>* is treated as a single-precision
7611: (cell-sized) negative integer, and is represented using 2's-complement
7612: arithmetic. Examples are -45 -5681 -0
7613: @item
7614: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
7615: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7616: (all three of these represent the same number).
7617: @item
7618: A string of the form -<digit><digit>*.<digit>* is treated as a
7619: double-precision (double-cell-sized) negative integer, and is
7620: represented using 2's-complement arithmetic. Examples are -3465. -3.465
7621: -34.65 (all three of these represent the same number).
7622: @item
7623: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7624: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
7625: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
7626: number) +12.E-4
7627: @end itemize
7628:
7629: By default, the number base used for integer number conversion is given
7630: by the contents of the variable @code{base}. Note that a lot of
7631: confusion can result from unexpected values of @code{base}. If you
7632: change @code{base} anywhere, make sure to save the old value and restore
7633: it afterwards. In general I recommend keeping @code{base} decimal, and
7634: using the prefixes described below for the popular non-decimal bases.
7635:
7636: doc-dpl
7637: doc-base
7638: doc-hex
7639: doc-decimal
7640:
7641: @cindex '-prefix for character strings
7642: @cindex &-prefix for decimal numbers
7643: @cindex #-prefix for decimal numbers
7644: @cindex %-prefix for binary numbers
7645: @cindex $-prefix for hexadecimal numbers
7646: @cindex 0x-prefix for hexadecimal numbers
7647: Gforth allows you to override the value of @code{base} by using a
7648: prefix@footnote{Some Forth implementations provide a similar scheme by
7649: implementing @code{$} etc. as parsing words that process the subsequent
7650: number in the input stream and push it onto the stack. For example, see
7651: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7652: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7653: is required between the prefix and the number.} before the first digit
7654: of an (integer) number. The following prefixes are supported:
7655:
7656: @itemize @bullet
7657: @item
7658: @code{&} -- decimal
7659: @item
7660: @code{#} -- decimal
7661: @item
7662: @code{%} -- binary
7663: @item
7664: @code{$} -- hexadecimal
7665: @item
7666: @code{0x} -- hexadecimal, if base<33.
7667: @item
7668: @code{'} -- numeric value (e.g., ASCII code) of next character; an
7669: optional @code{'} may be present after the character.
7670: @end itemize
7671:
7672: Here are some examples, with the equivalent decimal number shown after
7673: in braces:
7674:
7675: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7676: 'A (65),
7677: -'a' (-97),
7678: &905 (905), $abc (2478), $ABC (2478).
7679:
7680: @cindex number conversion - traps for the unwary
7681: @noindent
7682: Number conversion has a number of traps for the unwary:
7683:
7684: @itemize @bullet
7685: @item
7686: You cannot determine the current number base using the code sequence
7687: @code{base @@ .} -- the number base is always 10 in the current number
7688: base. Instead, use something like @code{base @@ dec.}
7689: @item
7690: If the number base is set to a value greater than 14 (for example,
7691: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7692: it to be intepreted as either a single-precision integer or a
7693: floating-point number (Gforth treats it as an integer). The ambiguity
7694: can be resolved by explicitly stating the sign of the mantissa and/or
7695: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7696: ambiguity arises; either representation will be treated as a
7697: floating-point number.
7698: @item
7699: There is a word @code{bin} but it does @i{not} set the number base!
7700: It is used to specify file types.
7701: @item
7702: ANS Forth requires the @code{.} of a double-precision number to be the
7703: final character in the string. Gforth allows the @code{.} to be
7704: anywhere after the first digit.
7705: @item
7706: The number conversion process does not check for overflow.
7707: @item
7708: In an ANS Forth program @code{base} is required to be decimal when
7709: converting floating-point numbers. In Gforth, number conversion to
7710: floating-point numbers always uses base &10, irrespective of the value
7711: of @code{base}.
7712: @end itemize
7713:
7714: You can read numbers into your programs with the words described in
7715: @ref{Input}.
7716:
7717: @node Interpret/Compile states, Interpreter Directives, Number Conversion, The Text Interpreter
7718: @subsection Interpret/Compile states
7719: @cindex Interpret/Compile states
7720:
7721: A standard program is not permitted to change @code{state}
7722: explicitly. However, it can change @code{state} implicitly, using the
7723: words @code{[} and @code{]}. When @code{[} is executed it switches
7724: @code{state} to interpret state, and therefore the text interpreter
7725: starts interpreting. When @code{]} is executed it switches @code{state}
7726: to compile state and therefore the text interpreter starts
7727: compiling. The most common usage for these words is for switching into
7728: interpret state and back from within a colon definition; this technique
7729: can be used to compile a literal (for an example, @pxref{Literals}) or
7730: for conditional compilation (for an example, @pxref{Interpreter
7731: Directives}).
7732:
7733:
7734: @c This is a bad example: It's non-standard, and it's not necessary.
7735: @c However, I can't think of a good example for switching into compile
7736: @c state when there is no current word (@code{state}-smart words are not a
7737: @c good reason). So maybe we should use an example for switching into
7738: @c interpret @code{state} in a colon def. - anton
7739: @c nac-> I agree. I started out by putting in the example, then realised
7740: @c that it was non-ANS, so wrote more words around it. I hope this
7741: @c re-written version is acceptable to you. I do want to keep the example
7742: @c as it is helpful for showing what is and what is not portable, particularly
7743: @c where it outlaws a style in common use.
7744:
7745: @c anton: it's more important to show what's portable. After we have done
7746: @c that, we can also show what's not. In any case, I have written a
7747: @c section Compiling Words which also deals with [ ].
7748:
7749: @c !! The following example does not work in Gforth 0.5.9 or later.
7750:
7751: @c @code{[} and @code{]} also give you the ability to switch into compile
7752: @c state and back, but we cannot think of any useful Standard application
7753: @c for this ability. Pre-ANS Forth textbooks have examples like this:
7754:
7755: @c @example
7756: @c : AA ." this is A" ;
7757: @c : BB ." this is B" ;
7758: @c : CC ." this is C" ;
7759:
7760: @c create table ] aa bb cc [
7761:
7762: @c : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7763: @c cells table + @@ execute ;
7764: @c @end example
7765:
7766: @c This example builds a jump table; @code{0 go} will display ``@code{this
7767: @c is A}''. Using @code{[} and @code{]} in this example is equivalent to
7768: @c defining @code{table} like this:
7769:
7770: @c @example
7771: @c create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
7772: @c @end example
7773:
7774: @c The problem with this code is that the definition of @code{table} is not
7775: @c portable -- it @i{compile}s execution tokens into code space. Whilst it
7776: @c @i{may} work on systems where code space and data space co-incide, the
7777: @c Standard only allows data space to be assigned for a @code{CREATE}d
7778: @c word. In addition, the Standard only allows @code{@@} to access data
7779: @c space, whilst this example is using it to access code space. The only
7780: @c portable, Standard way to build this table is to build it in data space,
7781: @c like this:
7782:
7783: @c @example
7784: @c create table ' aa , ' bb , ' cc ,
7785: @c @end example
7786:
7787: @c doc-state
7788:
7789:
7790: @node Interpreter Directives, , Interpret/Compile states, The Text Interpreter
7791: @subsection Interpreter Directives
7792: @cindex interpreter directives
7793: @cindex conditional compilation
7794:
7795: These words are usually used in interpret state; typically to control
7796: which parts of a source file are processed by the text
7797: interpreter. There are only a few ANS Forth Standard words, but Gforth
7798: supplements these with a rich set of immediate control structure words
7799: to compensate for the fact that the non-immediate versions can only be
7800: used in compile state (@pxref{Control Structures}). Typical usages:
7801:
7802: @example
7803: FALSE Constant HAVE-ASSEMBLER
7804: .
7805: .
7806: HAVE-ASSEMBLER [IF]
7807: : ASSEMBLER-FEATURE
7808: ...
7809: ;
7810: [ENDIF]
7811: .
7812: .
7813: : SEE
7814: ... \ general-purpose SEE code
7815: [ HAVE-ASSEMBLER [IF] ]
7816: ... \ assembler-specific SEE code
7817: [ [ENDIF] ]
7818: ;
7819: @end example
7820:
7821:
7822: doc-[IF]
7823: doc-[ELSE]
7824: doc-[THEN]
7825: doc-[ENDIF]
7826:
7827: doc-[IFDEF]
7828: doc-[IFUNDEF]
7829:
7830: doc-[?DO]
7831: doc-[DO]
7832: doc-[FOR]
7833: doc-[LOOP]
7834: doc-[+LOOP]
7835: doc-[NEXT]
7836:
7837: doc-[BEGIN]
7838: doc-[UNTIL]
7839: doc-[AGAIN]
7840: doc-[WHILE]
7841: doc-[REPEAT]
7842:
7843:
7844: @c -------------------------------------------------------------
7845: @node The Input Stream, Word Lists, The Text Interpreter, Words
7846: @section The Input Stream
7847: @cindex input stream
7848:
7849: @c !! integrate this better with the "Text Interpreter" section
7850: The text interpreter reads from the input stream, which can come from
7851: several sources (@pxref{Input Sources}). Some words, in particular
7852: defining words, but also words like @code{'}, read parameters from the
7853: input stream instead of from the stack.
7854:
7855: Such words are called parsing words, because they parse the input
7856: stream. Parsing words are hard to use in other words, because it is
7857: hard to pass program-generated parameters through the input stream.
7858: They also usually have an unintuitive combination of interpretation and
7859: compilation semantics when implemented naively, leading to various
7860: approaches that try to produce a more intuitive behaviour
7861: (@pxref{Combined words}).
7862:
7863: It should be obvious by now that parsing words are a bad idea. If you
7864: want to implement a parsing word for convenience, also provide a factor
7865: of the word that does not parse, but takes the parameters on the stack.
7866: To implement the parsing word on top if it, you can use the following
7867: words:
7868:
7869: @c anton: these belong in the input stream section
7870: doc-parse
7871: doc-parse-name
7872: doc-parse-word
7873: doc-name
7874: doc-word
7875: doc-\"-parse
7876: doc-refill
7877:
7878: Conversely, if you have the bad luck (or lack of foresight) to have to
7879: deal with parsing words without having such factors, how do you pass a
7880: string that is not in the input stream to it?
7881:
7882: doc-execute-parsing
7883:
7884: A definition of this word in ANS Forth is provided in
7885: @file{compat/execute-parsing.fs}.
7886:
7887: If you want to run a parsing word on a file, the following word should
7888: help:
7889:
7890: doc-execute-parsing-file
7891:
7892: @c -------------------------------------------------------------
7893: @node Word Lists, Environmental Queries, The Input Stream, Words
7894: @section Word Lists
7895: @cindex word lists
7896: @cindex header space
7897:
7898: A wordlist is a list of named words; you can add new words and look up
7899: words by name (and you can remove words in a restricted way with
7900: markers). Every named (and @code{reveal}ed) word is in one wordlist.
7901:
7902: @cindex search order stack
7903: The text interpreter searches the wordlists present in the search order
7904: (a stack of wordlists), from the top to the bottom. Within each
7905: wordlist, the search starts conceptually at the newest word; i.e., if
7906: two words in a wordlist have the same name, the newer word is found.
7907:
7908: @cindex compilation word list
7909: New words are added to the @dfn{compilation wordlist} (aka current
7910: wordlist).
7911:
7912: @cindex wid
7913: A word list is identified by a cell-sized word list identifier (@i{wid})
7914: in much the same way as a file is identified by a file handle. The
7915: numerical value of the wid has no (portable) meaning, and might change
7916: from session to session.
7917:
7918: The ANS Forth ``Search order'' word set is intended to provide a set of
7919: low-level tools that allow various different schemes to be
7920: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
7921: word. @file{compat/vocabulary.fs} provides an implementation in ANS
7922: Forth.
7923:
7924: @comment TODO: locals section refers to here, saying that every word list (aka
7925: @comment vocabulary) has its own methods for searching etc. Need to document that.
7926: @c anton: but better in a separate subsection on wordlist internals
7927:
7928: @comment TODO: document markers, reveal, tables, mappedwordlist
7929:
7930: @comment the gforthman- prefix is used to pick out the true definition of a
7931: @comment word from the source files, rather than some alias.
7932:
7933: doc-forth-wordlist
7934: doc-definitions
7935: doc-get-current
7936: doc-set-current
7937: doc-get-order
7938: doc---gforthman-set-order
7939: doc-wordlist
7940: doc-table
7941: doc->order
7942: doc-previous
7943: doc-also
7944: doc---gforthman-forth
7945: doc-only
7946: doc---gforthman-order
7947:
7948: doc-find
7949: doc-search-wordlist
7950:
7951: doc-words
7952: doc-vlist
7953: @c doc-words-deferred
7954:
7955: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
7956: doc-root
7957: doc-vocabulary
7958: doc-seal
7959: doc-vocs
7960: doc-current
7961: doc-context
7962:
7963:
7964: @menu
7965: * Vocabularies::
7966: * Why use word lists?::
7967: * Word list example::
7968: @end menu
7969:
7970: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
7971: @subsection Vocabularies
7972: @cindex Vocabularies, detailed explanation
7973:
7974: Here is an example of creating and using a new wordlist using ANS
7975: Forth words:
7976:
7977: @example
7978: wordlist constant my-new-words-wordlist
7979: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
7980:
7981: \ add it to the search order
7982: also my-new-words
7983:
7984: \ alternatively, add it to the search order and make it
7985: \ the compilation word list
7986: also my-new-words definitions
7987: \ type "order" to see the problem
7988: @end example
7989:
7990: The problem with this example is that @code{order} has no way to
7991: associate the name @code{my-new-words} with the wid of the word list (in
7992: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
7993: that has no associated name). There is no Standard way of associating a
7994: name with a wid.
7995:
7996: In Gforth, this example can be re-coded using @code{vocabulary}, which
7997: associates a name with a wid:
7998:
7999: @example
8000: vocabulary my-new-words
8001:
8002: \ add it to the search order
8003: also my-new-words
8004:
8005: \ alternatively, add it to the search order and make it
8006: \ the compilation word list
8007: my-new-words definitions
8008: \ type "order" to see that the problem is solved
8009: @end example
8010:
8011:
8012: @node Why use word lists?, Word list example, Vocabularies, Word Lists
8013: @subsection Why use word lists?
8014: @cindex word lists - why use them?
8015:
8016: Here are some reasons why people use wordlists:
8017:
8018: @itemize @bullet
8019:
8020: @c anton: Gforth's hashing implementation makes the search speed
8021: @c independent from the number of words. But it is linear with the number
8022: @c of wordlists that have to be searched, so in effect using more wordlists
8023: @c actually slows down compilation.
8024:
8025: @c @item
8026: @c To improve compilation speed by reducing the number of header space
8027: @c entries that must be searched. This is achieved by creating a new
8028: @c word list that contains all of the definitions that are used in the
8029: @c definition of a Forth system but which would not usually be used by
8030: @c programs running on that system. That word list would be on the search
8031: @c list when the Forth system was compiled but would be removed from the
8032: @c search list for normal operation. This can be a useful technique for
8033: @c low-performance systems (for example, 8-bit processors in embedded
8034: @c systems) but is unlikely to be necessary in high-performance desktop
8035: @c systems.
8036:
8037: @item
8038: To prevent a set of words from being used outside the context in which
8039: they are valid. Two classic examples of this are an integrated editor
8040: (all of the edit commands are defined in a separate word list; the
8041: search order is set to the editor word list when the editor is invoked;
8042: the old search order is restored when the editor is terminated) and an
8043: integrated assembler (the op-codes for the machine are defined in a
8044: separate word list which is used when a @code{CODE} word is defined).
8045:
8046: @item
8047: To organize the words of an application or library into a user-visible
8048: set (in @code{forth-wordlist} or some other common wordlist) and a set
8049: of helper words used just for the implementation (hidden in a separate
8050: wordlist). This keeps @code{words}' output smaller, separates
8051: implementation and interface, and reduces the chance of name conflicts
8052: within the common wordlist.
8053:
8054: @item
8055: To prevent a name-space clash between multiple definitions with the same
8056: name. For example, when building a cross-compiler you might have a word
8057: @code{IF} that generates conditional code for your target system. By
8058: placing this definition in a different word list you can control whether
8059: the host system's @code{IF} or the target system's @code{IF} get used in
8060: any particular context by controlling the order of the word lists on the
8061: search order stack.
8062:
8063: @end itemize
8064:
8065: The downsides of using wordlists are:
8066:
8067: @itemize
8068:
8069: @item
8070: Debugging becomes more cumbersome.
8071:
8072: @item
8073: Name conflicts worked around with wordlists are still there, and you
8074: have to arrange the search order carefully to get the desired results;
8075: if you forget to do that, you get hard-to-find errors (as in any case
8076: where you read the code differently from the compiler; @code{see} can
8077: help seeing which of several possible words the name resolves to in such
8078: cases). @code{See} displays just the name of the words, not what
8079: wordlist they belong to, so it might be misleading. Using unique names
8080: is a better approach to avoid name conflicts.
8081:
8082: @item
8083: You have to explicitly undo any changes to the search order. In many
8084: cases it would be more convenient if this happened implicitly. Gforth
8085: currently does not provide such a feature, but it may do so in the
8086: future.
8087: @end itemize
8088:
8089:
8090: @node Word list example, , Why use word lists?, Word Lists
8091: @subsection Word list example
8092: @cindex word lists - example
8093:
8094: The following example is from the
8095: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
8096: garbage collector} and uses wordlists to separate public words from
8097: helper words:
8098:
8099: @example
8100: get-current ( wid )
8101: vocabulary garbage-collector also garbage-collector definitions
8102: ... \ define helper words
8103: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
8104: ... \ define the public (i.e., API) words
8105: \ they can refer to the helper words
8106: previous \ restore original search order (helper words become invisible)
8107: @end example
8108:
8109: @c -------------------------------------------------------------
8110: @node Environmental Queries, Files, Word Lists, Words
8111: @section Environmental Queries
8112: @cindex environmental queries
8113:
8114: ANS Forth introduced the idea of ``environmental queries'' as a way
8115: for a program running on a system to determine certain characteristics of the system.
8116: The Standard specifies a number of strings that might be recognised by a system.
8117:
8118: The Standard requires that the header space used for environmental queries
8119: be distinct from the header space used for definitions.
8120:
8121: Typically, environmental queries are supported by creating a set of
8122: definitions in a word list that is @i{only} used during environmental
8123: queries; that is what Gforth does. There is no Standard way of adding
8124: definitions to the set of recognised environmental queries, but any
8125: implementation that supports the loading of optional word sets must have
8126: some mechanism for doing this (after loading the word set, the
8127: associated environmental query string must return @code{true}). In
8128: Gforth, the word list used to honour environmental queries can be
8129: manipulated just like any other word list.
8130:
8131:
8132: doc-environment?
8133: doc-environment-wordlist
8134:
8135: doc-gforth
8136: doc-os-class
8137:
8138:
8139: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
8140: returning two items on the stack, querying it using @code{environment?}
8141: will return an additional item; the @code{true} flag that shows that the
8142: string was recognised.
8143:
8144: @comment TODO Document the standard strings or note where they are documented herein
8145:
8146: Here are some examples of using environmental queries:
8147:
8148: @example
8149: s" address-unit-bits" environment? 0=
8150: [IF]
8151: cr .( environmental attribute address-units-bits unknown... ) cr
8152: [ELSE]
8153: drop \ ensure balanced stack effect
8154: [THEN]
8155:
8156: \ this might occur in the prelude of a standard program that uses THROW
8157: s" exception" environment? [IF]
8158: 0= [IF]
8159: : throw abort" exception thrown" ;
8160: [THEN]
8161: [ELSE] \ we don't know, so make sure
8162: : throw abort" exception thrown" ;
8163: [THEN]
8164:
8165: s" gforth" environment? [IF] .( Gforth version ) TYPE
8166: [ELSE] .( Not Gforth..) [THEN]
8167:
8168: \ a program using v*
8169: s" gforth" environment? [IF]
8170: s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
8171: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8172: >r swap 2swap swap 0e r> 0 ?DO
8173: dup f@ over + 2swap dup f@ f* f+ over + 2swap
8174: LOOP
8175: 2drop 2drop ;
8176: [THEN]
8177: [ELSE] \
8178: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8179: ...
8180: [THEN]
8181: @end example
8182:
8183: Here is an example of adding a definition to the environment word list:
8184:
8185: @example
8186: get-current environment-wordlist set-current
8187: true constant block
8188: true constant block-ext
8189: set-current
8190: @end example
8191:
8192: You can see what definitions are in the environment word list like this:
8193:
8194: @example
8195: environment-wordlist >order words previous
8196: @end example
8197:
8198:
8199: @c -------------------------------------------------------------
8200: @node Files, Blocks, Environmental Queries, Words
8201: @section Files
8202: @cindex files
8203: @cindex I/O - file-handling
8204:
8205: Gforth provides facilities for accessing files that are stored in the
8206: host operating system's file-system. Files that are processed by Gforth
8207: can be divided into two categories:
8208:
8209: @itemize @bullet
8210: @item
8211: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
8212: @item
8213: Files that are processed by some other program (@dfn{general files}).
8214: @end itemize
8215:
8216: @menu
8217: * Forth source files::
8218: * General files::
8219: * Redirection::
8220: * Search Paths::
8221: @end menu
8222:
8223: @c -------------------------------------------------------------
8224: @node Forth source files, General files, Files, Files
8225: @subsection Forth source files
8226: @cindex including files
8227: @cindex Forth source files
8228:
8229: The simplest way to interpret the contents of a file is to use one of
8230: these two formats:
8231:
8232: @example
8233: include mysource.fs
8234: s" mysource.fs" included
8235: @end example
8236:
8237: You usually want to include a file only if it is not included already
8238: (by, say, another source file). In that case, you can use one of these
8239: three formats:
8240:
8241: @example
8242: require mysource.fs
8243: needs mysource.fs
8244: s" mysource.fs" required
8245: @end example
8246:
8247: @cindex stack effect of included files
8248: @cindex including files, stack effect
8249: It is good practice to write your source files such that interpreting them
8250: does not change the stack. Source files designed in this way can be used with
8251: @code{required} and friends without complications. For example:
8252:
8253: @example
8254: 1024 require foo.fs drop
8255: @end example
8256:
8257: Here you want to pass the argument 1024 (e.g., a buffer size) to
8258: @file{foo.fs}. Interpreting @file{foo.fs} has the stack effect ( n -- n
8259: ), which allows its use with @code{require}. Of course with such
8260: parameters to required files, you have to ensure that the first
8261: @code{require} fits for all uses (i.e., @code{require} it early in the
8262: master load file).
8263:
8264: doc-include-file
8265: doc-included
8266: doc-included?
8267: doc-include
8268: doc-required
8269: doc-require
8270: doc-needs
8271: @c doc-init-included-files @c internal
8272: doc-sourcefilename
8273: doc-sourceline#
8274:
8275: A definition in ANS Forth for @code{required} is provided in
8276: @file{compat/required.fs}.
8277:
8278: @c -------------------------------------------------------------
8279: @node General files, Redirection, Forth source files, Files
8280: @subsection General files
8281: @cindex general files
8282: @cindex file-handling
8283:
8284: Files are opened/created by name and type. The following file access
8285: methods (FAMs) are recognised:
8286:
8287: @cindex fam (file access method)
8288: doc-r/o
8289: doc-r/w
8290: doc-w/o
8291: doc-bin
8292:
8293:
8294: When a file is opened/created, it returns a file identifier,
8295: @i{wfileid} that is used for all other file commands. All file
8296: commands also return a status value, @i{wior}, that is 0 for a
8297: successful operation and an implementation-defined non-zero value in the
8298: case of an error.
8299:
8300:
8301: doc-open-file
8302: doc-create-file
8303:
8304: doc-close-file
8305: doc-delete-file
8306: doc-rename-file
8307: doc-read-file
8308: doc-read-line
8309: doc-key-file
8310: doc-key?-file
8311: doc-write-file
8312: doc-write-line
8313: doc-emit-file
8314: doc-flush-file
8315:
8316: doc-file-status
8317: doc-file-position
8318: doc-reposition-file
8319: doc-file-size
8320: doc-resize-file
8321:
8322: doc-slurp-file
8323: doc-slurp-fid
8324: doc-stdin
8325: doc-stdout
8326: doc-stderr
8327:
8328: @c ---------------------------------------------------------
8329: @node Redirection, Search Paths, General files, Files
8330: @subsection Redirection
8331: @cindex Redirection
8332: @cindex Input Redirection
8333: @cindex Output Redirection
8334:
8335: You can redirect the output of @code{type} and @code{emit} and all the
8336: words that use them (all output words that don't have an explicit
8337: target file) to an arbitrary file with the @code{>outfile
8338: ... outfile<} construct, used like this:
8339:
8340: @example
8341: : print-some-warning ( n -- )
8342: stderr >outfile cr ." warning# " . outfile< ;
8343: @end example
8344:
8345: After the @code{outfile<}, the original output direction is restored;
8346: this construct is nestable and safe against exceptions. Similarly,
8347: there is a construct @code{>infile ... infile<} for redirecting the
8348: input of @code{key} and its users (any input word that does not take a
8349: file explicitly).
8350:
8351: If you do not want to redirect the input or output to a file, you can
8352: also make use of the fact that @code{key}, @code{emit} and @code{type}
8353: are deferred words (@pxref{Deferred Words}). However, in that case
8354: you have to worry about the restoration and the protection against
8355: exceptions yourself; also, note that for redirecting the output in
8356: this way, you have to redirect both @code{emit} and @code{type}.
8357:
8358: doc->outfile
8359: doc-outfile<
8360: doc->infile
8361: doc-infile<
8362:
8363: @c ---------------------------------------------------------
8364: @node Search Paths, , Redirection, Files
8365: @subsection Search Paths
8366: @cindex path for @code{included}
8367: @cindex file search path
8368: @cindex @code{include} search path
8369: @cindex search path for files
8370:
8371: If you specify an absolute filename (i.e., a filename starting with
8372: @file{/} or @file{~}, or with @file{:} in the second position (as in
8373: @samp{C:...})) for @code{included} and friends, that file is included
8374: just as you would expect.
8375:
8376: If the filename starts with @file{./}, this refers to the directory that
8377: the present file was @code{included} from. This allows files to include
8378: other files relative to their own position (irrespective of the current
8379: working directory or the absolute position). This feature is essential
8380: for libraries consisting of several files, where a file may include
8381: other files from the library. It corresponds to @code{#include "..."}
8382: in C. If the current input source is not a file, @file{.} refers to the
8383: directory of the innermost file being included, or, if there is no file
8384: being included, to the current working directory.
8385:
8386: For relative filenames (not starting with @file{./}), Gforth uses a
8387: search path similar to Forth's search order (@pxref{Word Lists}). It
8388: tries to find the given filename in the directories present in the path,
8389: and includes the first one it finds. There are separate search paths for
8390: Forth source files and general files. If the search path contains the
8391: directory @file{.}, this refers to the directory of the current file, or
8392: the working directory, as if the file had been specified with @file{./}.
8393:
8394: Use @file{~+} to refer to the current working directory (as in the
8395: @code{bash}).
8396:
8397: @c anton: fold the following subsubsections into this subsection?
8398:
8399: @menu
8400: * Source Search Paths::
8401: * General Search Paths::
8402: @end menu
8403:
8404: @c ---------------------------------------------------------
8405: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
8406: @subsubsection Source Search Paths
8407: @cindex search path control, source files
8408:
8409: The search path is initialized when you start Gforth (@pxref{Invoking
8410: Gforth}). You can display it and change it using @code{fpath} in
8411: combination with the general path handling words.
8412:
8413: doc-fpath
8414: @c the functionality of the following words is easily available through
8415: @c fpath and the general path words. The may go away.
8416: @c doc-.fpath
8417: @c doc-fpath+
8418: @c doc-fpath=
8419: @c doc-open-fpath-file
8420:
8421: @noindent
8422: Here is an example of using @code{fpath} and @code{require}:
8423:
8424: @example
8425: fpath path= /usr/lib/forth/|./
8426: require timer.fs
8427: @end example
8428:
8429:
8430: @c ---------------------------------------------------------
8431: @node General Search Paths, , Source Search Paths, Search Paths
8432: @subsubsection General Search Paths
8433: @cindex search path control, source files
8434:
8435: Your application may need to search files in several directories, like
8436: @code{included} does. To facilitate this, Gforth allows you to define
8437: and use your own search paths, by providing generic equivalents of the
8438: Forth search path words:
8439:
8440: doc-open-path-file
8441: doc-path-allot
8442: doc-clear-path
8443: doc-also-path
8444: doc-.path
8445: doc-path+
8446: doc-path=
8447:
8448: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
8449:
8450: Here's an example of creating an empty search path:
8451: @c
8452: @example
8453: create mypath 500 path-allot \ maximum length 500 chars (is checked)
8454: @end example
8455:
8456: @c -------------------------------------------------------------
8457: @node Blocks, Other I/O, Files, Words
8458: @section Blocks
8459: @cindex I/O - blocks
8460: @cindex blocks
8461:
8462: When you run Gforth on a modern desk-top computer, it runs under the
8463: control of an operating system which provides certain services. One of
8464: these services is @var{file services}, which allows Forth source code
8465: and data to be stored in files and read into Gforth (@pxref{Files}).
8466:
8467: Traditionally, Forth has been an important programming language on
8468: systems where it has interfaced directly to the underlying hardware with
8469: no intervening operating system. Forth provides a mechanism, called
8470: @dfn{blocks}, for accessing mass storage on such systems.
8471:
8472: A block is a 1024-byte data area, which can be used to hold data or
8473: Forth source code. No structure is imposed on the contents of the
8474: block. A block is identified by its number; blocks are numbered
8475: contiguously from 1 to an implementation-defined maximum.
8476:
8477: A typical system that used blocks but no operating system might use a
8478: single floppy-disk drive for mass storage, with the disks formatted to
8479: provide 256-byte sectors. Blocks would be implemented by assigning the
8480: first four sectors of the disk to block 1, the second four sectors to
8481: block 2 and so on, up to the limit of the capacity of the disk. The disk
8482: would not contain any file system information, just the set of blocks.
8483:
8484: @cindex blocks file
8485: On systems that do provide file services, blocks are typically
8486: implemented by storing a sequence of blocks within a single @dfn{blocks
8487: file}. The size of the blocks file will be an exact multiple of 1024
8488: bytes, corresponding to the number of blocks it contains. This is the
8489: mechanism that Gforth uses.
8490:
8491: @cindex @file{blocks.fb}
8492: Only one blocks file can be open at a time. If you use block words without
8493: having specified a blocks file, Gforth defaults to the blocks file
8494: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
8495: locate a blocks file (@pxref{Source Search Paths}).
8496:
8497: @cindex block buffers
8498: When you read and write blocks under program control, Gforth uses a
8499: number of @dfn{block buffers} as intermediate storage. These buffers are
8500: not used when you use @code{load} to interpret the contents of a block.
8501:
8502: The behaviour of the block buffers is analagous to that of a cache.
8503: Each block buffer has three states:
8504:
8505: @itemize @bullet
8506: @item
8507: Unassigned
8508: @item
8509: Assigned-clean
8510: @item
8511: Assigned-dirty
8512: @end itemize
8513:
8514: Initially, all block buffers are @i{unassigned}. In order to access a
8515: block, the block (specified by its block number) must be assigned to a
8516: block buffer.
8517:
8518: The assignment of a block to a block buffer is performed by @code{block}
8519: or @code{buffer}. Use @code{block} when you wish to modify the existing
8520: contents of a block. Use @code{buffer} when you don't care about the
8521: existing contents of the block@footnote{The ANS Forth definition of
8522: @code{buffer} is intended not to cause disk I/O; if the data associated
8523: with the particular block is already stored in a block buffer due to an
8524: earlier @code{block} command, @code{buffer} will return that block
8525: buffer and the existing contents of the block will be
8526: available. Otherwise, @code{buffer} will simply assign a new, empty
8527: block buffer for the block.}.
8528:
8529: Once a block has been assigned to a block buffer using @code{block} or
8530: @code{buffer}, that block buffer becomes the @i{current block
8531: buffer}. Data may only be manipulated (read or written) within the
8532: current block buffer.
8533:
8534: When the contents of the current block buffer has been modified it is
8535: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
8536: either abandon the changes (by doing nothing) or mark the block as
8537: changed (assigned-dirty), using @code{update}. Using @code{update} does
8538: not change the blocks file; it simply changes a block buffer's state to
8539: @i{assigned-dirty}. The block will be written implicitly when it's
8540: buffer is needed for another block, or explicitly by @code{flush} or
8541: @code{save-buffers}.
8542:
8543: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
8544: blocks file on disk. Leaving Gforth with @code{bye} also performs a
8545: @code{flush}.
8546:
8547: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
8548: algorithm to assign a block buffer to a block. That means that any
8549: particular block can only be assigned to one specific block buffer,
8550: called (for the particular operation) the @i{victim buffer}. If the
8551: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8552: the new block immediately. If it is @i{assigned-dirty} its current
8553: contents are written back to the blocks file on disk before it is
8554: allocated to the new block.
8555:
8556: Although no structure is imposed on the contents of a block, it is
8557: traditional to display the contents as 16 lines each of 64 characters. A
8558: block provides a single, continuous stream of input (for example, it
8559: acts as a single parse area) -- there are no end-of-line characters
8560: within a block, and no end-of-file character at the end of a
8561: block. There are two consequences of this:
8562:
8563: @itemize @bullet
8564: @item
8565: The last character of one line wraps straight into the first character
8566: of the following line
8567: @item
8568: The word @code{\} -- comment to end of line -- requires special
8569: treatment; in the context of a block it causes all characters until the
8570: end of the current 64-character ``line'' to be ignored.
8571: @end itemize
8572:
8573: In Gforth, when you use @code{block} with a non-existent block number,
8574: the current blocks file will be extended to the appropriate size and the
8575: block buffer will be initialised with spaces.
8576:
8577: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8578: for details) but doesn't encourage the use of blocks; the mechanism is
8579: only provided for backward compatibility -- ANS Forth requires blocks to
8580: be available when files are.
8581:
8582: Common techniques that are used when working with blocks include:
8583:
8584: @itemize @bullet
8585: @item
8586: A screen editor that allows you to edit blocks without leaving the Forth
8587: environment.
8588: @item
8589: Shadow screens; where every code block has an associated block
8590: containing comments (for example: code in odd block numbers, comments in
8591: even block numbers). Typically, the block editor provides a convenient
8592: mechanism to toggle between code and comments.
8593: @item
8594: Load blocks; a single block (typically block 1) contains a number of
8595: @code{thru} commands which @code{load} the whole of the application.
8596: @end itemize
8597:
8598: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8599: integrated into a Forth programming environment.
8600:
8601: @comment TODO what about errors on open-blocks?
8602:
8603: doc-open-blocks
8604: doc-use
8605: doc-block-offset
8606: doc-get-block-fid
8607: doc-block-position
8608:
8609: doc-list
8610: doc-scr
8611:
8612: doc---gforthman-block
8613: doc-buffer
8614:
8615: doc-empty-buffers
8616: doc-empty-buffer
8617: doc-update
8618: doc-updated?
8619: doc-save-buffers
8620: doc-save-buffer
8621: doc-flush
8622:
8623: doc-load
8624: doc-thru
8625: doc-+load
8626: doc-+thru
8627: doc---gforthman--->
8628: doc-block-included
8629:
8630:
8631: @c -------------------------------------------------------------
8632: @node Other I/O, OS command line arguments, Blocks, Words
8633: @section Other I/O
8634: @cindex I/O - keyboard and display
8635:
8636: @menu
8637: * Simple numeric output:: Predefined formats
8638: * Formatted numeric output:: Formatted (pictured) output
8639: * String Formats:: How Forth stores strings in memory
8640: * Displaying characters and strings:: Other stuff
8641: * Input:: Input
8642: * Pipes:: How to create your own pipes
8643: * Xchars and Unicode:: Non-ASCII characters
8644: @end menu
8645:
8646: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8647: @subsection Simple numeric output
8648: @cindex numeric output - simple/free-format
8649:
8650: The simplest output functions are those that display numbers from the
8651: data or floating-point stacks. Floating-point output is always displayed
8652: using base 10. Numbers displayed from the data stack use the value stored
8653: in @code{base}.
8654:
8655:
8656: doc-.
8657: doc-dec.
8658: doc-hex.
8659: doc-u.
8660: doc-.r
8661: doc-u.r
8662: doc-d.
8663: doc-ud.
8664: doc-d.r
8665: doc-ud.r
8666: doc-f.
8667: doc-fe.
8668: doc-fs.
8669: doc-f.rdp
8670:
8671: Examples of printing the number 1234.5678E23 in the different floating-point output
8672: formats are shown below:
8673:
8674: @example
8675: f. 123456779999999000000000000.
8676: fe. 123.456779999999E24
8677: fs. 1.23456779999999E26
8678: @end example
8679:
8680:
8681: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8682: @subsection Formatted numeric output
8683: @cindex formatted numeric output
8684: @cindex pictured numeric output
8685: @cindex numeric output - formatted
8686:
8687: Forth traditionally uses a technique called @dfn{pictured numeric
8688: output} for formatted printing of integers. In this technique, digits
8689: are extracted from the number (using the current output radix defined by
8690: @code{base}), converted to ASCII codes and appended to a string that is
8691: built in a scratch-pad area of memory (@pxref{core-idef,
8692: Implementation-defined options, Implementation-defined
8693: options}). Arbitrary characters can be appended to the string during the
8694: extraction process. The completed string is specified by an address
8695: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8696: under program control.
8697:
8698: All of the integer output words described in the previous section
8699: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
8700: numeric output.
8701:
8702: Three important things to remember about pictured numeric output:
8703:
8704: @itemize @bullet
8705: @item
8706: It always operates on double-precision numbers; to display a
8707: single-precision number, convert it first (for ways of doing this
8708: @pxref{Double precision}).
8709: @item
8710: It always treats the double-precision number as though it were
8711: unsigned. The examples below show ways of printing signed numbers.
8712: @item
8713: The string is built up from right to left; least significant digit first.
8714: @end itemize
8715:
8716:
8717: doc-<#
8718: doc-<<#
8719: doc-#
8720: doc-#s
8721: doc-hold
8722: doc-sign
8723: doc-#>
8724: doc-#>>
8725:
8726: doc-represent
8727: doc-f>str-rdp
8728: doc-f>buf-rdp
8729:
8730:
8731: @noindent
8732: Here are some examples of using pictured numeric output:
8733:
8734: @example
8735: : my-u. ( u -- )
8736: \ Simplest use of pns.. behaves like Standard u.
8737: 0 \ convert to unsigned double
8738: <<# \ start conversion
8739: #s \ convert all digits
8740: #> \ complete conversion
8741: TYPE SPACE \ display, with trailing space
8742: #>> ; \ release hold area
8743:
8744: : cents-only ( u -- )
8745: 0 \ convert to unsigned double
8746: <<# \ start conversion
8747: # # \ convert two least-significant digits
8748: #> \ complete conversion, discard other digits
8749: TYPE SPACE \ display, with trailing space
8750: #>> ; \ release hold area
8751:
8752: : dollars-and-cents ( u -- )
8753: 0 \ convert to unsigned double
8754: <<# \ start conversion
8755: # # \ convert two least-significant digits
8756: [char] . hold \ insert decimal point
8757: #s \ convert remaining digits
8758: [char] $ hold \ append currency symbol
8759: #> \ complete conversion
8760: TYPE SPACE \ display, with trailing space
8761: #>> ; \ release hold area
8762:
8763: : my-. ( n -- )
8764: \ handling negatives.. behaves like Standard .
8765: s>d \ convert to signed double
8766: swap over dabs \ leave sign byte followed by unsigned double
8767: <<# \ start conversion
8768: #s \ convert all digits
8769: rot sign \ get at sign byte, append "-" if needed
8770: #> \ complete conversion
8771: TYPE SPACE \ display, with trailing space
8772: #>> ; \ release hold area
8773:
8774: : account. ( n -- )
8775: \ accountants don't like minus signs, they use parentheses
8776: \ for negative numbers
8777: s>d \ convert to signed double
8778: swap over dabs \ leave sign byte followed by unsigned double
8779: <<# \ start conversion
8780: 2 pick \ get copy of sign byte
8781: 0< IF [char] ) hold THEN \ right-most character of output
8782: #s \ convert all digits
8783: rot \ get at sign byte
8784: 0< IF [char] ( hold THEN
8785: #> \ complete conversion
8786: TYPE SPACE \ display, with trailing space
8787: #>> ; \ release hold area
8788:
8789: @end example
8790:
8791: Here are some examples of using these words:
8792:
8793: @example
8794: 1 my-u. 1
8795: hex -1 my-u. decimal FFFFFFFF
8796: 1 cents-only 01
8797: 1234 cents-only 34
8798: 2 dollars-and-cents $0.02
8799: 1234 dollars-and-cents $12.34
8800: 123 my-. 123
8801: -123 my. -123
8802: 123 account. 123
8803: -456 account. (456)
8804: @end example
8805:
8806:
8807: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8808: @subsection String Formats
8809: @cindex strings - see character strings
8810: @cindex character strings - formats
8811: @cindex I/O - see character strings
8812: @cindex counted strings
8813:
8814: @c anton: this does not really belong here; maybe the memory section,
8815: @c or the principles chapter
8816:
8817: Forth commonly uses two different methods for representing character
8818: strings:
8819:
8820: @itemize @bullet
8821: @item
8822: @cindex address of counted string
8823: @cindex counted string
8824: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8825: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8826: string and the string occupies the subsequent @i{n} char addresses in
8827: memory.
8828: @item
8829: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8830: of the string in characters, and @i{c-addr} is the address of the
8831: first byte of the string.
8832: @end itemize
8833:
8834: ANS Forth encourages the use of the second format when representing
8835: strings.
8836:
8837:
8838: doc-count
8839:
8840:
8841: For words that move, copy and search for strings see @ref{Memory
8842: Blocks}. For words that display characters and strings see
8843: @ref{Displaying characters and strings}.
8844:
8845: @node Displaying characters and strings, Input, String Formats, Other I/O
8846: @subsection Displaying characters and strings
8847: @cindex characters - compiling and displaying
8848: @cindex character strings - compiling and displaying
8849:
8850: This section starts with a glossary of Forth words and ends with a set
8851: of examples.
8852:
8853:
8854: doc-bl
8855: doc-space
8856: doc-spaces
8857: doc-emit
8858: doc-toupper
8859: doc-."
8860: doc-.(
8861: doc-.\"
8862: doc-type
8863: doc-typewhite
8864: doc-cr
8865: @cindex cursor control
8866: doc-at-xy
8867: doc-page
8868: doc-s"
8869: doc-s\"
8870: doc-c"
8871: doc-char
8872: doc-[char]
8873:
8874:
8875: @noindent
8876: As an example, consider the following text, stored in a file @file{test.fs}:
8877:
8878: @example
8879: .( text-1)
8880: : my-word
8881: ." text-2" cr
8882: .( text-3)
8883: ;
8884:
8885: ." text-4"
8886:
8887: : my-char
8888: [char] ALPHABET emit
8889: char emit
8890: ;
8891: @end example
8892:
8893: When you load this code into Gforth, the following output is generated:
8894:
8895: @example
8896: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
8897: @end example
8898:
8899: @itemize @bullet
8900: @item
8901: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
8902: is an immediate word; it behaves in the same way whether it is used inside
8903: or outside a colon definition.
8904: @item
8905: Message @code{text-4} is displayed because of Gforth's added interpretation
8906: semantics for @code{."}.
8907: @item
8908: Message @code{text-2} is @i{not} displayed, because the text interpreter
8909: performs the compilation semantics for @code{."} within the definition of
8910: @code{my-word}.
8911: @end itemize
8912:
8913: Here are some examples of executing @code{my-word} and @code{my-char}:
8914:
8915: @example
8916: @kbd{my-word @key{RET}} text-2
8917: ok
8918: @kbd{my-char fred @key{RET}} Af ok
8919: @kbd{my-char jim @key{RET}} Aj ok
8920: @end example
8921:
8922: @itemize @bullet
8923: @item
8924: Message @code{text-2} is displayed because of the run-time behaviour of
8925: @code{."}.
8926: @item
8927: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
8928: on the stack at run-time. @code{emit} always displays the character
8929: when @code{my-char} is executed.
8930: @item
8931: @code{char} parses a string at run-time and the second @code{emit} displays
8932: the first character of the string.
8933: @item
8934: If you type @code{see my-char} you can see that @code{[char]} discarded
8935: the text ``LPHABET'' and only compiled the display code for ``A'' into the
8936: definition of @code{my-char}.
8937: @end itemize
8938:
8939:
8940:
8941: @node Input, Pipes, Displaying characters and strings, Other I/O
8942: @subsection Input
8943: @cindex input
8944: @cindex I/O - see input
8945: @cindex parsing a string
8946:
8947: For ways of storing character strings in memory see @ref{String Formats}.
8948:
8949: @comment TODO examples for >number >float accept key key? pad parse word refill
8950: @comment then index them
8951:
8952:
8953: doc-key
8954: doc-key?
8955: doc-ekey
8956: doc-ekey>char
8957: doc-ekey?
8958:
8959: Gforth recognizes various keys available on ANSI terminals (in MS-DOS
8960: you need the ANSI.SYS driver to get that behaviour). These are the
8961: keyboard events produced by various common keys:
8962:
8963: doc-k-left
8964: doc-k-right
8965: doc-k-up
8966: doc-k-down
8967: doc-k-home
8968: doc-k-end
8969: doc-k-prior
8970: doc-k-next
8971: doc-k-insert
8972: doc-k-delete
8973:
8974: The function keys (aka keypad keys) are:
8975:
8976: doc-k1
8977: doc-k2
8978: doc-k3
8979: doc-k4
8980: doc-k5
8981: doc-k6
8982: doc-k7
8983: doc-k8
8984: doc-k9
8985: doc-k10
8986: doc-k11
8987: doc-k12
8988:
8989: Note that K11 and K12 are not as widely available. The shifted
8990: function keys are also not very widely available:
8991:
8992: doc-s-k1
8993: doc-s-k2
8994: doc-s-k3
8995: doc-s-k4
8996: doc-s-k5
8997: doc-s-k6
8998: doc-s-k7
8999: doc-s-k8
9000: doc-s-k9
9001: doc-s-k10
9002: doc-s-k11
9003: doc-s-k12
9004:
9005: Words for inputting one line from the keyboard:
9006:
9007: doc-accept
9008: doc-edit-line
9009:
9010: Conversion words:
9011:
9012: doc-s>number?
9013: doc-s>unumber?
9014: doc->number
9015: doc->float
9016:
9017:
9018: @comment obsolescent words..
9019: Obsolescent input and conversion words:
9020:
9021: doc-convert
9022: doc-expect
9023: doc-span
9024:
9025:
9026: @node Pipes, Xchars and Unicode, Input, Other I/O
9027: @subsection Pipes
9028: @cindex pipes, creating your own
9029:
9030: In addition to using Gforth in pipes created by other processes
9031: (@pxref{Gforth in pipes}), you can create your own pipe with
9032: @code{open-pipe}, and read from or write to it.
9033:
9034: doc-open-pipe
9035: doc-close-pipe
9036:
9037: If you write to a pipe, Gforth can throw a @code{broken-pipe-error}; if
9038: you don't catch this exception, Gforth will catch it and exit, usually
9039: silently (@pxref{Gforth in pipes}). Since you probably do not want
9040: this, you should wrap a @code{catch} or @code{try} block around the code
9041: from @code{open-pipe} to @code{close-pipe}, so you can deal with the
9042: problem yourself, and then return to regular processing.
9043:
9044: doc-broken-pipe-error
9045:
9046: @node Xchars and Unicode, , Pipes, Other I/O
9047: @subsection Xchars and Unicode
9048:
9049: This chapter needs completion
9050:
9051: @node OS command line arguments, Locals, Other I/O, Words
9052: @section OS command line arguments
9053: @cindex OS command line arguments
9054: @cindex command line arguments, OS
9055: @cindex arguments, OS command line
9056:
9057: The usual way to pass arguments to Gforth programs on the command line
9058: is via the @option{-e} option, e.g.
9059:
9060: @example
9061: gforth -e "123 456" foo.fs -e bye
9062: @end example
9063:
9064: However, you may want to interpret the command-line arguments directly.
9065: In that case, you can access the (image-specific) command-line arguments
9066: through @code{next-arg}:
9067:
9068: doc-next-arg
9069:
9070: Here's an example program @file{echo.fs} for @code{next-arg}:
9071:
9072: @example
9073: : echo ( -- )
9074: begin
9075: next-arg 2dup 0 0 d<> while
9076: type space
9077: repeat
9078: 2drop ;
9079:
9080: echo cr bye
9081: @end example
9082:
9083: This can be invoked with
9084:
9085: @example
9086: gforth echo.fs hello world
9087: @end example
9088:
9089: and it will print
9090:
9091: @example
9092: hello world
9093: @end example
9094:
9095: The next lower level of dealing with the OS command line are the
9096: following words:
9097:
9098: doc-arg
9099: doc-shift-args
9100:
9101: Finally, at the lowest level Gforth provides the following words:
9102:
9103: doc-argc
9104: doc-argv
9105:
9106: @c -------------------------------------------------------------
9107: @node Locals, Structures, OS command line arguments, Words
9108: @section Locals
9109: @cindex locals
9110:
9111: Local variables can make Forth programming more enjoyable and Forth
9112: programs easier to read. Unfortunately, the locals of ANS Forth are
9113: laden with restrictions. Therefore, we provide not only the ANS Forth
9114: locals wordset, but also our own, more powerful locals wordset (we
9115: implemented the ANS Forth locals wordset through our locals wordset).
9116:
9117: The ideas in this section have also been published in M. Anton Ertl,
9118: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
9119: Automatic Scoping of Local Variables}}, EuroForth '94.
9120:
9121: @menu
9122: * Gforth locals::
9123: * ANS Forth locals::
9124: @end menu
9125:
9126: @node Gforth locals, ANS Forth locals, Locals, Locals
9127: @subsection Gforth locals
9128: @cindex Gforth locals
9129: @cindex locals, Gforth style
9130:
9131: Locals can be defined with
9132:
9133: @example
9134: @{ local1 local2 ... -- comment @}
9135: @end example
9136: or
9137: @example
9138: @{ local1 local2 ... @}
9139: @end example
9140:
9141: E.g.,
9142: @example
9143: : max @{ n1 n2 -- n3 @}
9144: n1 n2 > if
9145: n1
9146: else
9147: n2
9148: endif ;
9149: @end example
9150:
9151: The similarity of locals definitions with stack comments is intended. A
9152: locals definition often replaces the stack comment of a word. The order
9153: of the locals corresponds to the order in a stack comment and everything
9154: after the @code{--} is really a comment.
9155:
9156: This similarity has one disadvantage: It is too easy to confuse locals
9157: declarations with stack comments, causing bugs and making them hard to
9158: find. However, this problem can be avoided by appropriate coding
9159: conventions: Do not use both notations in the same program. If you do,
9160: they should be distinguished using additional means, e.g. by position.
9161:
9162: @cindex types of locals
9163: @cindex locals types
9164: The name of the local may be preceded by a type specifier, e.g.,
9165: @code{F:} for a floating point value:
9166:
9167: @example
9168: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
9169: \ complex multiplication
9170: Ar Br f* Ai Bi f* f-
9171: Ar Bi f* Ai Br f* f+ ;
9172: @end example
9173:
9174: @cindex flavours of locals
9175: @cindex locals flavours
9176: @cindex value-flavoured locals
9177: @cindex variable-flavoured locals
9178: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9179: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9180: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9181: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9182: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9183: produces its address (which becomes invalid when the variable's scope is
9184: left). E.g., the standard word @code{emit} can be defined in terms of
9185: @code{type} like this:
9186:
9187: @example
9188: : emit @{ C^ char* -- @}
9189: char* 1 type ;
9190: @end example
9191:
9192: @cindex default type of locals
9193: @cindex locals, default type
9194: A local without type specifier is a @code{W:} local. Both flavours of
9195: locals are initialized with values from the data or FP stack.
9196:
9197: Currently there is no way to define locals with user-defined data
9198: structures, but we are working on it.
9199:
9200: Gforth allows defining locals everywhere in a colon definition. This
9201: poses the following questions:
9202:
9203: @menu
9204: * Where are locals visible by name?::
9205: * How long do locals live?::
9206: * Locals programming style::
9207: * Locals implementation::
9208: @end menu
9209:
9210: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9211: @subsubsection Where are locals visible by name?
9212: @cindex locals visibility
9213: @cindex visibility of locals
9214: @cindex scope of locals
9215:
9216: Basically, the answer is that locals are visible where you would expect
9217: it in block-structured languages, and sometimes a little longer. If you
9218: want to restrict the scope of a local, enclose its definition in
9219: @code{SCOPE}...@code{ENDSCOPE}.
9220:
9221:
9222: doc-scope
9223: doc-endscope
9224:
9225:
9226: These words behave like control structure words, so you can use them
9227: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9228: arbitrary ways.
9229:
9230: If you want a more exact answer to the visibility question, here's the
9231: basic principle: A local is visible in all places that can only be
9232: reached through the definition of the local@footnote{In compiler
9233: construction terminology, all places dominated by the definition of the
9234: local.}. In other words, it is not visible in places that can be reached
9235: without going through the definition of the local. E.g., locals defined
9236: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9237: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9238: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
9239:
9240: The reasoning behind this solution is: We want to have the locals
9241: visible as long as it is meaningful. The user can always make the
9242: visibility shorter by using explicit scoping. In a place that can
9243: only be reached through the definition of a local, the meaning of a
9244: local name is clear. In other places it is not: How is the local
9245: initialized at the control flow path that does not contain the
9246: definition? Which local is meant, if the same name is defined twice in
9247: two independent control flow paths?
9248:
9249: This should be enough detail for nearly all users, so you can skip the
9250: rest of this section. If you really must know all the gory details and
9251: options, read on.
9252:
9253: In order to implement this rule, the compiler has to know which places
9254: are unreachable. It knows this automatically after @code{AHEAD},
9255: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9256: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9257: compiler that the control flow never reaches that place. If
9258: @code{UNREACHABLE} is not used where it could, the only consequence is
9259: that the visibility of some locals is more limited than the rule above
9260: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9261: lie to the compiler), buggy code will be produced.
9262:
9263:
9264: doc-unreachable
9265:
9266:
9267: Another problem with this rule is that at @code{BEGIN}, the compiler
9268: does not know which locals will be visible on the incoming
9269: back-edge. All problems discussed in the following are due to this
9270: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9271: loops as examples; the discussion also applies to @code{?DO} and other
9272: loops). Perhaps the most insidious example is:
9273: @example
9274: AHEAD
9275: BEGIN
9276: x
9277: [ 1 CS-ROLL ] THEN
9278: @{ x @}
9279: ...
9280: UNTIL
9281: @end example
9282:
9283: This should be legal according to the visibility rule. The use of
9284: @code{x} can only be reached through the definition; but that appears
9285: textually below the use.
9286:
9287: From this example it is clear that the visibility rules cannot be fully
9288: implemented without major headaches. Our implementation treats common
9289: cases as advertised and the exceptions are treated in a safe way: The
9290: compiler makes a reasonable guess about the locals visible after a
9291: @code{BEGIN}; if it is too pessimistic, the
9292: user will get a spurious error about the local not being defined; if the
9293: compiler is too optimistic, it will notice this later and issue a
9294: warning. In the case above the compiler would complain about @code{x}
9295: being undefined at its use. You can see from the obscure examples in
9296: this section that it takes quite unusual control structures to get the
9297: compiler into trouble, and even then it will often do fine.
9298:
9299: If the @code{BEGIN} is reachable from above, the most optimistic guess
9300: is that all locals visible before the @code{BEGIN} will also be
9301: visible after the @code{BEGIN}. This guess is valid for all loops that
9302: are entered only through the @code{BEGIN}, in particular, for normal
9303: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9304: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9305: compiler. When the branch to the @code{BEGIN} is finally generated by
9306: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9307: warns the user if it was too optimistic:
9308: @example
9309: IF
9310: @{ x @}
9311: BEGIN
9312: \ x ?
9313: [ 1 cs-roll ] THEN
9314: ...
9315: UNTIL
9316: @end example
9317:
9318: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9319: optimistically assumes that it lives until the @code{THEN}. It notices
9320: this difference when it compiles the @code{UNTIL} and issues a
9321: warning. The user can avoid the warning, and make sure that @code{x}
9322: is not used in the wrong area by using explicit scoping:
9323: @example
9324: IF
9325: SCOPE
9326: @{ x @}
9327: ENDSCOPE
9328: BEGIN
9329: [ 1 cs-roll ] THEN
9330: ...
9331: UNTIL
9332: @end example
9333:
9334: Since the guess is optimistic, there will be no spurious error messages
9335: about undefined locals.
9336:
9337: If the @code{BEGIN} is not reachable from above (e.g., after
9338: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9339: optimistic guess, as the locals visible after the @code{BEGIN} may be
9340: defined later. Therefore, the compiler assumes that no locals are
9341: visible after the @code{BEGIN}. However, the user can use
9342: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9343: visible at the BEGIN as at the point where the top control-flow stack
9344: item was created.
9345:
9346:
9347: doc-assume-live
9348:
9349:
9350: @noindent
9351: E.g.,
9352: @example
9353: @{ x @}
9354: AHEAD
9355: ASSUME-LIVE
9356: BEGIN
9357: x
9358: [ 1 CS-ROLL ] THEN
9359: ...
9360: UNTIL
9361: @end example
9362:
9363: Other cases where the locals are defined before the @code{BEGIN} can be
9364: handled by inserting an appropriate @code{CS-ROLL} before the
9365: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9366: behind the @code{ASSUME-LIVE}).
9367:
9368: Cases where locals are defined after the @code{BEGIN} (but should be
9369: visible immediately after the @code{BEGIN}) can only be handled by
9370: rearranging the loop. E.g., the ``most insidious'' example above can be
9371: arranged into:
9372: @example
9373: BEGIN
9374: @{ x @}
9375: ... 0=
9376: WHILE
9377: x
9378: REPEAT
9379: @end example
9380:
9381: @node How long do locals live?, Locals programming style, Where are locals visible by name?, Gforth locals
9382: @subsubsection How long do locals live?
9383: @cindex locals lifetime
9384: @cindex lifetime of locals
9385:
9386: The right answer for the lifetime question would be: A local lives at
9387: least as long as it can be accessed. For a value-flavoured local this
9388: means: until the end of its visibility. However, a variable-flavoured
9389: local could be accessed through its address far beyond its visibility
9390: scope. Ultimately, this would mean that such locals would have to be
9391: garbage collected. Since this entails un-Forth-like implementation
9392: complexities, I adopted the same cowardly solution as some other
9393: languages (e.g., C): The local lives only as long as it is visible;
9394: afterwards its address is invalid (and programs that access it
9395: afterwards are erroneous).
9396:
9397: @node Locals programming style, Locals implementation, How long do locals live?, Gforth locals
9398: @subsubsection Locals programming style
9399: @cindex locals programming style
9400: @cindex programming style, locals
9401:
9402: The freedom to define locals anywhere has the potential to change
9403: programming styles dramatically. In particular, the need to use the
9404: return stack for intermediate storage vanishes. Moreover, all stack
9405: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9406: determined arguments) can be eliminated: If the stack items are in the
9407: wrong order, just write a locals definition for all of them; then
9408: write the items in the order you want.
9409:
9410: This seems a little far-fetched and eliminating stack manipulations is
9411: unlikely to become a conscious programming objective. Still, the number
9412: of stack manipulations will be reduced dramatically if local variables
9413: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
9414: a traditional implementation of @code{max}).
9415:
9416: This shows one potential benefit of locals: making Forth programs more
9417: readable. Of course, this benefit will only be realized if the
9418: programmers continue to honour the principle of factoring instead of
9419: using the added latitude to make the words longer.
9420:
9421: @cindex single-assignment style for locals
9422: Using @code{TO} can and should be avoided. Without @code{TO},
9423: every value-flavoured local has only a single assignment and many
9424: advantages of functional languages apply to Forth. I.e., programs are
9425: easier to analyse, to optimize and to read: It is clear from the
9426: definition what the local stands for, it does not turn into something
9427: different later.
9428:
9429: E.g., a definition using @code{TO} might look like this:
9430: @example
9431: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9432: u1 u2 min 0
9433: ?do
9434: addr1 c@@ addr2 c@@ -
9435: ?dup-if
9436: unloop exit
9437: then
9438: addr1 char+ TO addr1
9439: addr2 char+ TO addr2
9440: loop
9441: u1 u2 - ;
9442: @end example
9443: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9444: every loop iteration. @code{strcmp} is a typical example of the
9445: readability problems of using @code{TO}. When you start reading
9446: @code{strcmp}, you think that @code{addr1} refers to the start of the
9447: string. Only near the end of the loop you realize that it is something
9448: else.
9449:
9450: This can be avoided by defining two locals at the start of the loop that
9451: are initialized with the right value for the current iteration.
9452: @example
9453: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9454: addr1 addr2
9455: u1 u2 min 0
9456: ?do @{ s1 s2 @}
9457: s1 c@@ s2 c@@ -
9458: ?dup-if
9459: unloop exit
9460: then
9461: s1 char+ s2 char+
9462: loop
9463: 2drop
9464: u1 u2 - ;
9465: @end example
9466: Here it is clear from the start that @code{s1} has a different value
9467: in every loop iteration.
9468:
9469: @node Locals implementation, , Locals programming style, Gforth locals
9470: @subsubsection Locals implementation
9471: @cindex locals implementation
9472: @cindex implementation of locals
9473:
9474: @cindex locals stack
9475: Gforth uses an extra locals stack. The most compelling reason for
9476: this is that the return stack is not float-aligned; using an extra stack
9477: also eliminates the problems and restrictions of using the return stack
9478: as locals stack. Like the other stacks, the locals stack grows toward
9479: lower addresses. A few primitives allow an efficient implementation:
9480:
9481:
9482: doc-@local#
9483: doc-f@local#
9484: doc-laddr#
9485: doc-lp+!#
9486: doc-lp!
9487: doc->l
9488: doc-f>l
9489:
9490:
9491: In addition to these primitives, some specializations of these
9492: primitives for commonly occurring inline arguments are provided for
9493: efficiency reasons, e.g., @code{@@local0} as specialization of
9494: @code{@@local#} for the inline argument 0. The following compiling words
9495: compile the right specialized version, or the general version, as
9496: appropriate:
9497:
9498:
9499: @c doc-compile-@local
9500: @c doc-compile-f@local
9501: doc-compile-lp+!
9502:
9503:
9504: Combinations of conditional branches and @code{lp+!#} like
9505: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9506: is taken) are provided for efficiency and correctness in loops.
9507:
9508: A special area in the dictionary space is reserved for keeping the
9509: local variable names. @code{@{} switches the dictionary pointer to this
9510: area and @code{@}} switches it back and generates the locals
9511: initializing code. @code{W:} etc.@ are normal defining words. This
9512: special area is cleared at the start of every colon definition.
9513:
9514: @cindex word list for defining locals
9515: A special feature of Gforth's dictionary is used to implement the
9516: definition of locals without type specifiers: every word list (aka
9517: vocabulary) has its own methods for searching
9518: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9519: with a special search method: When it is searched for a word, it
9520: actually creates that word using @code{W:}. @code{@{} changes the search
9521: order to first search the word list containing @code{@}}, @code{W:} etc.,
9522: and then the word list for defining locals without type specifiers.
9523:
9524: The lifetime rules support a stack discipline within a colon
9525: definition: The lifetime of a local is either nested with other locals
9526: lifetimes or it does not overlap them.
9527:
9528: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9529: pointer manipulation is generated. Between control structure words
9530: locals definitions can push locals onto the locals stack. @code{AGAIN}
9531: is the simplest of the other three control flow words. It has to
9532: restore the locals stack depth of the corresponding @code{BEGIN}
9533: before branching. The code looks like this:
9534: @format
9535: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9536: @code{branch} <begin>
9537: @end format
9538:
9539: @code{UNTIL} is a little more complicated: If it branches back, it
9540: must adjust the stack just like @code{AGAIN}. But if it falls through,
9541: the locals stack must not be changed. The compiler generates the
9542: following code:
9543: @format
9544: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9545: @end format
9546: The locals stack pointer is only adjusted if the branch is taken.
9547:
9548: @code{THEN} can produce somewhat inefficient code:
9549: @format
9550: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9551: <orig target>:
9552: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9553: @end format
9554: The second @code{lp+!#} adjusts the locals stack pointer from the
9555: level at the @i{orig} point to the level after the @code{THEN}. The
9556: first @code{lp+!#} adjusts the locals stack pointer from the current
9557: level to the level at the orig point, so the complete effect is an
9558: adjustment from the current level to the right level after the
9559: @code{THEN}.
9560:
9561: @cindex locals information on the control-flow stack
9562: @cindex control-flow stack items, locals information
9563: In a conventional Forth implementation a dest control-flow stack entry
9564: is just the target address and an orig entry is just the address to be
9565: patched. Our locals implementation adds a word list to every orig or dest
9566: item. It is the list of locals visible (or assumed visible) at the point
9567: described by the entry. Our implementation also adds a tag to identify
9568: the kind of entry, in particular to differentiate between live and dead
9569: (reachable and unreachable) orig entries.
9570:
9571: A few unusual operations have to be performed on locals word lists:
9572:
9573:
9574: doc-common-list
9575: doc-sub-list?
9576: doc-list-size
9577:
9578:
9579: Several features of our locals word list implementation make these
9580: operations easy to implement: The locals word lists are organised as
9581: linked lists; the tails of these lists are shared, if the lists
9582: contain some of the same locals; and the address of a name is greater
9583: than the address of the names behind it in the list.
9584:
9585: Another important implementation detail is the variable
9586: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9587: determine if they can be reached directly or only through the branch
9588: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9589: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9590: definition, by @code{BEGIN} and usually by @code{THEN}.
9591:
9592: Counted loops are similar to other loops in most respects, but
9593: @code{LEAVE} requires special attention: It performs basically the same
9594: service as @code{AHEAD}, but it does not create a control-flow stack
9595: entry. Therefore the information has to be stored elsewhere;
9596: traditionally, the information was stored in the target fields of the
9597: branches created by the @code{LEAVE}s, by organizing these fields into a
9598: linked list. Unfortunately, this clever trick does not provide enough
9599: space for storing our extended control flow information. Therefore, we
9600: introduce another stack, the leave stack. It contains the control-flow
9601: stack entries for all unresolved @code{LEAVE}s.
9602:
9603: Local names are kept until the end of the colon definition, even if
9604: they are no longer visible in any control-flow path. In a few cases
9605: this may lead to increased space needs for the locals name area, but
9606: usually less than reclaiming this space would cost in code size.
9607:
9608:
9609: @node ANS Forth locals, , Gforth locals, Locals
9610: @subsection ANS Forth locals
9611: @cindex locals, ANS Forth style
9612:
9613: The ANS Forth locals wordset does not define a syntax for locals, but
9614: words that make it possible to define various syntaxes. One of the
9615: possible syntaxes is a subset of the syntax we used in the Gforth locals
9616: wordset, i.e.:
9617:
9618: @example
9619: @{ local1 local2 ... -- comment @}
9620: @end example
9621: @noindent
9622: or
9623: @example
9624: @{ local1 local2 ... @}
9625: @end example
9626:
9627: The order of the locals corresponds to the order in a stack comment. The
9628: restrictions are:
9629:
9630: @itemize @bullet
9631: @item
9632: Locals can only be cell-sized values (no type specifiers are allowed).
9633: @item
9634: Locals can be defined only outside control structures.
9635: @item
9636: Locals can interfere with explicit usage of the return stack. For the
9637: exact (and long) rules, see the standard. If you don't use return stack
9638: accessing words in a definition using locals, you will be all right. The
9639: purpose of this rule is to make locals implementation on the return
9640: stack easier.
9641: @item
9642: The whole definition must be in one line.
9643: @end itemize
9644:
9645: Locals defined in ANS Forth behave like @code{VALUE}s
9646: (@pxref{Values}). I.e., they are initialized from the stack. Using their
9647: name produces their value. Their value can be changed using @code{TO}.
9648:
9649: Since the syntax above is supported by Gforth directly, you need not do
9650: anything to use it. If you want to port a program using this syntax to
9651: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9652: syntax on the other system.
9653:
9654: Note that a syntax shown in the standard, section A.13 looks
9655: similar, but is quite different in having the order of locals
9656: reversed. Beware!
9657:
9658: The ANS Forth locals wordset itself consists of one word:
9659:
9660: doc-(local)
9661:
9662: The ANS Forth locals extension wordset defines a syntax using
9663: @code{locals|}, but it is so awful that we strongly recommend not to use
9664: it. We have implemented this syntax to make porting to Gforth easy, but
9665: do not document it here. The problem with this syntax is that the locals
9666: are defined in an order reversed with respect to the standard stack
9667: comment notation, making programs harder to read, and easier to misread
9668: and miswrite. The only merit of this syntax is that it is easy to
9669: implement using the ANS Forth locals wordset.
9670:
9671:
9672: @c ----------------------------------------------------------
9673: @node Structures, Object-oriented Forth, Locals, Words
9674: @section Structures
9675: @cindex structures
9676: @cindex records
9677:
9678: This section presents the structure package that comes with Gforth. A
9679: version of the package implemented in ANS Forth is available in
9680: @file{compat/struct.fs}. This package was inspired by a posting on
9681: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9682: possibly John Hayes). A version of this section has been published in
9683: M. Anton Ertl,
9684: @uref{http://www.complang.tuwien.ac.at/forth/objects/structs.html, Yet
9685: Another Forth Structures Package}, Forth Dimensions 19(3), pages
9686: 13--16. Marcel Hendrix provided helpful comments.
9687:
9688: @menu
9689: * Why explicit structure support?::
9690: * Structure Usage::
9691: * Structure Naming Convention::
9692: * Structure Implementation::
9693: * Structure Glossary::
9694: @end menu
9695:
9696: @node Why explicit structure support?, Structure Usage, Structures, Structures
9697: @subsection Why explicit structure support?
9698:
9699: @cindex address arithmetic for structures
9700: @cindex structures using address arithmetic
9701: If we want to use a structure containing several fields, we could simply
9702: reserve memory for it, and access the fields using address arithmetic
9703: (@pxref{Address arithmetic}). As an example, consider a structure with
9704: the following fields
9705:
9706: @table @code
9707: @item a
9708: is a float
9709: @item b
9710: is a cell
9711: @item c
9712: is a float
9713: @end table
9714:
9715: Given the (float-aligned) base address of the structure we get the
9716: address of the field
9717:
9718: @table @code
9719: @item a
9720: without doing anything further.
9721: @item b
9722: with @code{float+}
9723: @item c
9724: with @code{float+ cell+ faligned}
9725: @end table
9726:
9727: It is easy to see that this can become quite tiring.
9728:
9729: Moreover, it is not very readable, because seeing a
9730: @code{cell+} tells us neither which kind of structure is
9731: accessed nor what field is accessed; we have to somehow infer the kind
9732: of structure, and then look up in the documentation, which field of
9733: that structure corresponds to that offset.
9734:
9735: Finally, this kind of address arithmetic also causes maintenance
9736: troubles: If you add or delete a field somewhere in the middle of the
9737: structure, you have to find and change all computations for the fields
9738: afterwards.
9739:
9740: So, instead of using @code{cell+} and friends directly, how
9741: about storing the offsets in constants:
9742:
9743: @example
9744: 0 constant a-offset
9745: 0 float+ constant b-offset
9746: 0 float+ cell+ faligned c-offset
9747: @end example
9748:
9749: Now we can get the address of field @code{x} with @code{x-offset
9750: +}. This is much better in all respects. Of course, you still
9751: have to change all later offset definitions if you add a field. You can
9752: fix this by declaring the offsets in the following way:
9753:
9754: @example
9755: 0 constant a-offset
9756: a-offset float+ constant b-offset
9757: b-offset cell+ faligned constant c-offset
9758: @end example
9759:
9760: Since we always use the offsets with @code{+}, we could use a defining
9761: word @code{cfield} that includes the @code{+} in the action of the
9762: defined word:
9763:
9764: @example
9765: : cfield ( n "name" -- )
9766: create ,
9767: does> ( name execution: addr1 -- addr2 )
9768: @@ + ;
9769:
9770: 0 cfield a
9771: 0 a float+ cfield b
9772: 0 b cell+ faligned cfield c
9773: @end example
9774:
9775: Instead of @code{x-offset +}, we now simply write @code{x}.
9776:
9777: The structure field words now can be used quite nicely. However,
9778: their definition is still a bit cumbersome: We have to repeat the
9779: name, the information about size and alignment is distributed before
9780: and after the field definitions etc. The structure package presented
9781: here addresses these problems.
9782:
9783: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9784: @subsection Structure Usage
9785: @cindex structure usage
9786:
9787: @cindex @code{field} usage
9788: @cindex @code{struct} usage
9789: @cindex @code{end-struct} usage
9790: You can define a structure for a (data-less) linked list with:
9791: @example
9792: struct
9793: cell% field list-next
9794: end-struct list%
9795: @end example
9796:
9797: With the address of the list node on the stack, you can compute the
9798: address of the field that contains the address of the next node with
9799: @code{list-next}. E.g., you can determine the length of a list
9800: with:
9801:
9802: @example
9803: : list-length ( list -- n )
9804: \ "list" is a pointer to the first element of a linked list
9805: \ "n" is the length of the list
9806: 0 BEGIN ( list1 n1 )
9807: over
9808: WHILE ( list1 n1 )
9809: 1+ swap list-next @@ swap
9810: REPEAT
9811: nip ;
9812: @end example
9813:
9814: You can reserve memory for a list node in the dictionary with
9815: @code{list% %allot}, which leaves the address of the list node on the
9816: stack. For the equivalent allocation on the heap you can use @code{list%
9817: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9818: use @code{list% %allocate}). You can get the the size of a list
9819: node with @code{list% %size} and its alignment with @code{list%
9820: %alignment}.
9821:
9822: Note that in ANS Forth the body of a @code{create}d word is
9823: @code{aligned} but not necessarily @code{faligned};
9824: therefore, if you do a:
9825:
9826: @example
9827: create @emph{name} foo% %allot drop
9828: @end example
9829:
9830: @noindent
9831: then the memory alloted for @code{foo%} is guaranteed to start at the
9832: body of @code{@emph{name}} only if @code{foo%} contains only character,
9833: cell and double fields. Therefore, if your structure contains floats,
9834: better use
9835:
9836: @example
9837: foo% %allot constant @emph{name}
9838: @end example
9839:
9840: @cindex structures containing structures
9841: You can include a structure @code{foo%} as a field of
9842: another structure, like this:
9843: @example
9844: struct
9845: ...
9846: foo% field ...
9847: ...
9848: end-struct ...
9849: @end example
9850:
9851: @cindex structure extension
9852: @cindex extended records
9853: Instead of starting with an empty structure, you can extend an
9854: existing structure. E.g., a plain linked list without data, as defined
9855: above, is hardly useful; You can extend it to a linked list of integers,
9856: like this:@footnote{This feature is also known as @emph{extended
9857: records}. It is the main innovation in the Oberon language; in other
9858: words, adding this feature to Modula-2 led Wirth to create a new
9859: language, write a new compiler etc. Adding this feature to Forth just
9860: required a few lines of code.}
9861:
9862: @example
9863: list%
9864: cell% field intlist-int
9865: end-struct intlist%
9866: @end example
9867:
9868: @code{intlist%} is a structure with two fields:
9869: @code{list-next} and @code{intlist-int}.
9870:
9871: @cindex structures containing arrays
9872: You can specify an array type containing @emph{n} elements of
9873: type @code{foo%} like this:
9874:
9875: @example
9876: foo% @emph{n} *
9877: @end example
9878:
9879: You can use this array type in any place where you can use a normal
9880: type, e.g., when defining a @code{field}, or with
9881: @code{%allot}.
9882:
9883: @cindex first field optimization
9884: The first field is at the base address of a structure and the word for
9885: this field (e.g., @code{list-next}) actually does not change the address
9886: on the stack. You may be tempted to leave it away in the interest of
9887: run-time and space efficiency. This is not necessary, because the
9888: structure package optimizes this case: If you compile a first-field
9889: words, no code is generated. So, in the interest of readability and
9890: maintainability you should include the word for the field when accessing
9891: the field.
9892:
9893:
9894: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
9895: @subsection Structure Naming Convention
9896: @cindex structure naming convention
9897:
9898: The field names that come to (my) mind are often quite generic, and,
9899: if used, would cause frequent name clashes. E.g., many structures
9900: probably contain a @code{counter} field. The structure names
9901: that come to (my) mind are often also the logical choice for the names
9902: of words that create such a structure.
9903:
9904: Therefore, I have adopted the following naming conventions:
9905:
9906: @itemize @bullet
9907: @cindex field naming convention
9908: @item
9909: The names of fields are of the form
9910: @code{@emph{struct}-@emph{field}}, where
9911: @code{@emph{struct}} is the basic name of the structure, and
9912: @code{@emph{field}} is the basic name of the field. You can
9913: think of field words as converting the (address of the)
9914: structure into the (address of the) field.
9915:
9916: @cindex structure naming convention
9917: @item
9918: The names of structures are of the form
9919: @code{@emph{struct}%}, where
9920: @code{@emph{struct}} is the basic name of the structure.
9921: @end itemize
9922:
9923: This naming convention does not work that well for fields of extended
9924: structures; e.g., the integer list structure has a field
9925: @code{intlist-int}, but has @code{list-next}, not
9926: @code{intlist-next}.
9927:
9928: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
9929: @subsection Structure Implementation
9930: @cindex structure implementation
9931: @cindex implementation of structures
9932:
9933: The central idea in the implementation is to pass the data about the
9934: structure being built on the stack, not in some global
9935: variable. Everything else falls into place naturally once this design
9936: decision is made.
9937:
9938: The type description on the stack is of the form @emph{align
9939: size}. Keeping the size on the top-of-stack makes dealing with arrays
9940: very simple.
9941:
9942: @code{field} is a defining word that uses @code{Create}
9943: and @code{DOES>}. The body of the field contains the offset
9944: of the field, and the normal @code{DOES>} action is simply:
9945:
9946: @example
9947: @@ +
9948: @end example
9949:
9950: @noindent
9951: i.e., add the offset to the address, giving the stack effect
9952: @i{addr1 -- addr2} for a field.
9953:
9954: @cindex first field optimization, implementation
9955: This simple structure is slightly complicated by the optimization
9956: for fields with offset 0, which requires a different
9957: @code{DOES>}-part (because we cannot rely on there being
9958: something on the stack if such a field is invoked during
9959: compilation). Therefore, we put the different @code{DOES>}-parts
9960: in separate words, and decide which one to invoke based on the
9961: offset. For a zero offset, the field is basically a noop; it is
9962: immediate, and therefore no code is generated when it is compiled.
9963:
9964: @node Structure Glossary, , Structure Implementation, Structures
9965: @subsection Structure Glossary
9966: @cindex structure glossary
9967:
9968:
9969: doc-%align
9970: doc-%alignment
9971: doc-%alloc
9972: doc-%allocate
9973: doc-%allot
9974: doc-cell%
9975: doc-char%
9976: doc-dfloat%
9977: doc-double%
9978: doc-end-struct
9979: doc-field
9980: doc-float%
9981: doc-naligned
9982: doc-sfloat%
9983: doc-%size
9984: doc-struct
9985:
9986:
9987: @c -------------------------------------------------------------
9988: @node Object-oriented Forth, Programming Tools, Structures, Words
9989: @section Object-oriented Forth
9990:
9991: Gforth comes with three packages for object-oriented programming:
9992: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
9993: is preloaded, so you have to @code{include} them before use. The most
9994: important differences between these packages (and others) are discussed
9995: in @ref{Comparison with other object models}. All packages are written
9996: in ANS Forth and can be used with any other ANS Forth.
9997:
9998: @menu
9999: * Why object-oriented programming?::
10000: * Object-Oriented Terminology::
10001: * Objects::
10002: * OOF::
10003: * Mini-OOF::
10004: * Comparison with other object models::
10005: @end menu
10006:
10007: @c ----------------------------------------------------------------
10008: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
10009: @subsection Why object-oriented programming?
10010: @cindex object-oriented programming motivation
10011: @cindex motivation for object-oriented programming
10012:
10013: Often we have to deal with several data structures (@emph{objects}),
10014: that have to be treated similarly in some respects, but differently in
10015: others. Graphical objects are the textbook example: circles, triangles,
10016: dinosaurs, icons, and others, and we may want to add more during program
10017: development. We want to apply some operations to any graphical object,
10018: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
10019: has to do something different for every kind of object.
10020: @comment TODO add some other operations eg perimeter, area
10021: @comment and tie in to concrete examples later..
10022:
10023: We could implement @code{draw} as a big @code{CASE}
10024: control structure that executes the appropriate code depending on the
10025: kind of object to be drawn. This would be not be very elegant, and,
10026: moreover, we would have to change @code{draw} every time we add
10027: a new kind of graphical object (say, a spaceship).
10028:
10029: What we would rather do is: When defining spaceships, we would tell
10030: the system: ``Here's how you @code{draw} a spaceship; you figure
10031: out the rest''.
10032:
10033: This is the problem that all systems solve that (rightfully) call
10034: themselves object-oriented; the object-oriented packages presented here
10035: solve this problem (and not much else).
10036: @comment TODO ?list properties of oo systems.. oo vs o-based?
10037:
10038: @c ------------------------------------------------------------------------
10039: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
10040: @subsection Object-Oriented Terminology
10041: @cindex object-oriented terminology
10042: @cindex terminology for object-oriented programming
10043:
10044: This section is mainly for reference, so you don't have to understand
10045: all of it right away. The terminology is mainly Smalltalk-inspired. In
10046: short:
10047:
10048: @table @emph
10049: @cindex class
10050: @item class
10051: a data structure definition with some extras.
10052:
10053: @cindex object
10054: @item object
10055: an instance of the data structure described by the class definition.
10056:
10057: @cindex instance variables
10058: @item instance variables
10059: fields of the data structure.
10060:
10061: @cindex selector
10062: @cindex method selector
10063: @cindex virtual function
10064: @item selector
10065: (or @emph{method selector}) a word (e.g.,
10066: @code{draw}) that performs an operation on a variety of data
10067: structures (classes). A selector describes @emph{what} operation to
10068: perform. In C++ terminology: a (pure) virtual function.
10069:
10070: @cindex method
10071: @item method
10072: the concrete definition that performs the operation
10073: described by the selector for a specific class. A method specifies
10074: @emph{how} the operation is performed for a specific class.
10075:
10076: @cindex selector invocation
10077: @cindex message send
10078: @cindex invoking a selector
10079: @item selector invocation
10080: a call of a selector. One argument of the call (the TOS (top-of-stack))
10081: is used for determining which method is used. In Smalltalk terminology:
10082: a message (consisting of the selector and the other arguments) is sent
10083: to the object.
10084:
10085: @cindex receiving object
10086: @item receiving object
10087: the object used for determining the method executed by a selector
10088: invocation. In the @file{objects.fs} model, it is the object that is on
10089: the TOS when the selector is invoked. (@emph{Receiving} comes from
10090: the Smalltalk @emph{message} terminology.)
10091:
10092: @cindex child class
10093: @cindex parent class
10094: @cindex inheritance
10095: @item child class
10096: a class that has (@emph{inherits}) all properties (instance variables,
10097: selectors, methods) from a @emph{parent class}. In Smalltalk
10098: terminology: The subclass inherits from the superclass. In C++
10099: terminology: The derived class inherits from the base class.
10100:
10101: @end table
10102:
10103: @c If you wonder about the message sending terminology, it comes from
10104: @c a time when each object had it's own task and objects communicated via
10105: @c message passing; eventually the Smalltalk developers realized that
10106: @c they can do most things through simple (indirect) calls. They kept the
10107: @c terminology.
10108:
10109: @c --------------------------------------------------------------
10110: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
10111: @subsection The @file{objects.fs} model
10112: @cindex objects
10113: @cindex object-oriented programming
10114:
10115: @cindex @file{objects.fs}
10116: @cindex @file{oof.fs}
10117:
10118: This section describes the @file{objects.fs} package. This material also
10119: has been published in M. Anton Ertl,
10120: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
10121: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
10122: 37--43.
10123: @c McKewan's and Zsoter's packages
10124:
10125: This section assumes that you have read @ref{Structures}.
10126:
10127: The techniques on which this model is based have been used to implement
10128: the parser generator, Gray, and have also been used in Gforth for
10129: implementing the various flavours of word lists (hashed or not,
10130: case-sensitive or not, special-purpose word lists for locals etc.).
10131:
10132:
10133: @menu
10134: * Properties of the Objects model::
10135: * Basic Objects Usage::
10136: * The Objects base class::
10137: * Creating objects::
10138: * Object-Oriented Programming Style::
10139: * Class Binding::
10140: * Method conveniences::
10141: * Classes and Scoping::
10142: * Dividing classes::
10143: * Object Interfaces::
10144: * Objects Implementation::
10145: * Objects Glossary::
10146: @end menu
10147:
10148: Marcel Hendrix provided helpful comments on this section.
10149:
10150: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
10151: @subsubsection Properties of the @file{objects.fs} model
10152: @cindex @file{objects.fs} properties
10153:
10154: @itemize @bullet
10155: @item
10156: It is straightforward to pass objects on the stack. Passing
10157: selectors on the stack is a little less convenient, but possible.
10158:
10159: @item
10160: Objects are just data structures in memory, and are referenced by their
10161: address. You can create words for objects with normal defining words
10162: like @code{constant}. Likewise, there is no difference between instance
10163: variables that contain objects and those that contain other data.
10164:
10165: @item
10166: Late binding is efficient and easy to use.
10167:
10168: @item
10169: It avoids parsing, and thus avoids problems with state-smartness
10170: and reduced extensibility; for convenience there are a few parsing
10171: words, but they have non-parsing counterparts. There are also a few
10172: defining words that parse. This is hard to avoid, because all standard
10173: defining words parse (except @code{:noname}); however, such
10174: words are not as bad as many other parsing words, because they are not
10175: state-smart.
10176:
10177: @item
10178: It does not try to incorporate everything. It does a few things and does
10179: them well (IMO). In particular, this model was not designed to support
10180: information hiding (although it has features that may help); you can use
10181: a separate package for achieving this.
10182:
10183: @item
10184: It is layered; you don't have to learn and use all features to use this
10185: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10186: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
10187: are optional and independent of each other.
10188:
10189: @item
10190: An implementation in ANS Forth is available.
10191:
10192: @end itemize
10193:
10194:
10195: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10196: @subsubsection Basic @file{objects.fs} Usage
10197: @cindex basic objects usage
10198: @cindex objects, basic usage
10199:
10200: You can define a class for graphical objects like this:
10201:
10202: @cindex @code{class} usage
10203: @cindex @code{end-class} usage
10204: @cindex @code{selector} usage
10205: @example
10206: object class \ "object" is the parent class
10207: selector draw ( x y graphical -- )
10208: end-class graphical
10209: @end example
10210:
10211: This code defines a class @code{graphical} with an
10212: operation @code{draw}. We can perform the operation
10213: @code{draw} on any @code{graphical} object, e.g.:
10214:
10215: @example
10216: 100 100 t-rex draw
10217: @end example
10218:
10219: @noindent
10220: where @code{t-rex} is a word (say, a constant) that produces a
10221: graphical object.
10222:
10223: @comment TODO add a 2nd operation eg perimeter.. and use for
10224: @comment a concrete example
10225:
10226: @cindex abstract class
10227: How do we create a graphical object? With the present definitions,
10228: we cannot create a useful graphical object. The class
10229: @code{graphical} describes graphical objects in general, but not
10230: any concrete graphical object type (C++ users would call it an
10231: @emph{abstract class}); e.g., there is no method for the selector
10232: @code{draw} in the class @code{graphical}.
10233:
10234: For concrete graphical objects, we define child classes of the
10235: class @code{graphical}, e.g.:
10236:
10237: @cindex @code{overrides} usage
10238: @cindex @code{field} usage in class definition
10239: @example
10240: graphical class \ "graphical" is the parent class
10241: cell% field circle-radius
10242:
10243: :noname ( x y circle -- )
10244: circle-radius @@ draw-circle ;
10245: overrides draw
10246:
10247: :noname ( n-radius circle -- )
10248: circle-radius ! ;
10249: overrides construct
10250:
10251: end-class circle
10252: @end example
10253:
10254: Here we define a class @code{circle} as a child of @code{graphical},
10255: with field @code{circle-radius} (which behaves just like a field
10256: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10257: for the selectors @code{draw} and @code{construct} (@code{construct} is
10258: defined in @code{object}, the parent class of @code{graphical}).
10259:
10260: Now we can create a circle on the heap (i.e.,
10261: @code{allocate}d memory) with:
10262:
10263: @cindex @code{heap-new} usage
10264: @example
10265: 50 circle heap-new constant my-circle
10266: @end example
10267:
10268: @noindent
10269: @code{heap-new} invokes @code{construct}, thus
10270: initializing the field @code{circle-radius} with 50. We can draw
10271: this new circle at (100,100) with:
10272:
10273: @example
10274: 100 100 my-circle draw
10275: @end example
10276:
10277: @cindex selector invocation, restrictions
10278: @cindex class definition, restrictions
10279: Note: You can only invoke a selector if the object on the TOS
10280: (the receiving object) belongs to the class where the selector was
10281: defined or one of its descendents; e.g., you can invoke
10282: @code{draw} only for objects belonging to @code{graphical}
10283: or its descendents (e.g., @code{circle}). Immediately before
10284: @code{end-class}, the search order has to be the same as
10285: immediately after @code{class}.
10286:
10287: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10288: @subsubsection The @file{object.fs} base class
10289: @cindex @code{object} class
10290:
10291: When you define a class, you have to specify a parent class. So how do
10292: you start defining classes? There is one class available from the start:
10293: @code{object}. It is ancestor for all classes and so is the
10294: only class that has no parent. It has two selectors: @code{construct}
10295: and @code{print}.
10296:
10297: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10298: @subsubsection Creating objects
10299: @cindex creating objects
10300: @cindex object creation
10301: @cindex object allocation options
10302:
10303: @cindex @code{heap-new} discussion
10304: @cindex @code{dict-new} discussion
10305: @cindex @code{construct} discussion
10306: You can create and initialize an object of a class on the heap with
10307: @code{heap-new} ( ... class -- object ) and in the dictionary
10308: (allocation with @code{allot}) with @code{dict-new} (
10309: ... class -- object ). Both words invoke @code{construct}, which
10310: consumes the stack items indicated by "..." above.
10311:
10312: @cindex @code{init-object} discussion
10313: @cindex @code{class-inst-size} discussion
10314: If you want to allocate memory for an object yourself, you can get its
10315: alignment and size with @code{class-inst-size 2@@} ( class --
10316: align size ). Once you have memory for an object, you can initialize
10317: it with @code{init-object} ( ... class object -- );
10318: @code{construct} does only a part of the necessary work.
10319:
10320: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10321: @subsubsection Object-Oriented Programming Style
10322: @cindex object-oriented programming style
10323: @cindex programming style, object-oriented
10324:
10325: This section is not exhaustive.
10326:
10327: @cindex stack effects of selectors
10328: @cindex selectors and stack effects
10329: In general, it is a good idea to ensure that all methods for the
10330: same selector have the same stack effect: when you invoke a selector,
10331: you often have no idea which method will be invoked, so, unless all
10332: methods have the same stack effect, you will not know the stack effect
10333: of the selector invocation.
10334:
10335: One exception to this rule is methods for the selector
10336: @code{construct}. We know which method is invoked, because we
10337: specify the class to be constructed at the same place. Actually, I
10338: defined @code{construct} as a selector only to give the users a
10339: convenient way to specify initialization. The way it is used, a
10340: mechanism different from selector invocation would be more natural
10341: (but probably would take more code and more space to explain).
10342:
10343: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10344: @subsubsection Class Binding
10345: @cindex class binding
10346: @cindex early binding
10347:
10348: @cindex late binding
10349: Normal selector invocations determine the method at run-time depending
10350: on the class of the receiving object. This run-time selection is called
10351: @i{late binding}.
10352:
10353: Sometimes it's preferable to invoke a different method. For example,
10354: you might want to use the simple method for @code{print}ing
10355: @code{object}s instead of the possibly long-winded @code{print} method
10356: of the receiver class. You can achieve this by replacing the invocation
10357: of @code{print} with:
10358:
10359: @cindex @code{[bind]} usage
10360: @example
10361: [bind] object print
10362: @end example
10363:
10364: @noindent
10365: in compiled code or:
10366:
10367: @cindex @code{bind} usage
10368: @example
10369: bind object print
10370: @end example
10371:
10372: @cindex class binding, alternative to
10373: @noindent
10374: in interpreted code. Alternatively, you can define the method with a
10375: name (e.g., @code{print-object}), and then invoke it through the
10376: name. Class binding is just a (often more convenient) way to achieve
10377: the same effect; it avoids name clutter and allows you to invoke
10378: methods directly without naming them first.
10379:
10380: @cindex superclass binding
10381: @cindex parent class binding
10382: A frequent use of class binding is this: When we define a method
10383: for a selector, we often want the method to do what the selector does
10384: in the parent class, and a little more. There is a special word for
10385: this purpose: @code{[parent]}; @code{[parent]
10386: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10387: selector}}, where @code{@emph{parent}} is the parent
10388: class of the current class. E.g., a method definition might look like:
10389:
10390: @cindex @code{[parent]} usage
10391: @example
10392: :noname
10393: dup [parent] foo \ do parent's foo on the receiving object
10394: ... \ do some more
10395: ; overrides foo
10396: @end example
10397:
10398: @cindex class binding as optimization
10399: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10400: March 1997), Andrew McKewan presents class binding as an optimization
10401: technique. I recommend not using it for this purpose unless you are in
10402: an emergency. Late binding is pretty fast with this model anyway, so the
10403: benefit of using class binding is small; the cost of using class binding
10404: where it is not appropriate is reduced maintainability.
10405:
10406: While we are at programming style questions: You should bind
10407: selectors only to ancestor classes of the receiving object. E.g., say,
10408: you know that the receiving object is of class @code{foo} or its
10409: descendents; then you should bind only to @code{foo} and its
10410: ancestors.
10411:
10412: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10413: @subsubsection Method conveniences
10414: @cindex method conveniences
10415:
10416: In a method you usually access the receiving object pretty often. If
10417: you define the method as a plain colon definition (e.g., with
10418: @code{:noname}), you may have to do a lot of stack
10419: gymnastics. To avoid this, you can define the method with @code{m:
10420: ... ;m}. E.g., you could define the method for
10421: @code{draw}ing a @code{circle} with
10422:
10423: @cindex @code{this} usage
10424: @cindex @code{m:} usage
10425: @cindex @code{;m} usage
10426: @example
10427: m: ( x y circle -- )
10428: ( x y ) this circle-radius @@ draw-circle ;m
10429: @end example
10430:
10431: @cindex @code{exit} in @code{m: ... ;m}
10432: @cindex @code{exitm} discussion
10433: @cindex @code{catch} in @code{m: ... ;m}
10434: When this method is executed, the receiver object is removed from the
10435: stack; you can access it with @code{this} (admittedly, in this
10436: example the use of @code{m: ... ;m} offers no advantage). Note
10437: that I specify the stack effect for the whole method (i.e. including
10438: the receiver object), not just for the code between @code{m:}
10439: and @code{;m}. You cannot use @code{exit} in
10440: @code{m:...;m}; instead, use
10441: @code{exitm}.@footnote{Moreover, for any word that calls
10442: @code{catch} and was defined before loading
10443: @code{objects.fs}, you have to redefine it like I redefined
10444: @code{catch}: @code{: catch this >r catch r> to-this ;}}
10445:
10446: @cindex @code{inst-var} usage
10447: You will frequently use sequences of the form @code{this
10448: @emph{field}} (in the example above: @code{this
10449: circle-radius}). If you use the field only in this way, you can
10450: define it with @code{inst-var} and eliminate the
10451: @code{this} before the field name. E.g., the @code{circle}
10452: class above could also be defined with:
10453:
10454: @example
10455: graphical class
10456: cell% inst-var radius
10457:
10458: m: ( x y circle -- )
10459: radius @@ draw-circle ;m
10460: overrides draw
10461:
10462: m: ( n-radius circle -- )
10463: radius ! ;m
10464: overrides construct
10465:
10466: end-class circle
10467: @end example
10468:
10469: @code{radius} can only be used in @code{circle} and its
10470: descendent classes and inside @code{m:...;m}.
10471:
10472: @cindex @code{inst-value} usage
10473: You can also define fields with @code{inst-value}, which is
10474: to @code{inst-var} what @code{value} is to
10475: @code{variable}. You can change the value of such a field with
10476: @code{[to-inst]}. E.g., we could also define the class
10477: @code{circle} like this:
10478:
10479: @example
10480: graphical class
10481: inst-value radius
10482:
10483: m: ( x y circle -- )
10484: radius draw-circle ;m
10485: overrides draw
10486:
10487: m: ( n-radius circle -- )
10488: [to-inst] radius ;m
10489: overrides construct
10490:
10491: end-class circle
10492: @end example
10493:
10494: @c !! :m is easy to confuse with m:. Another name would be better.
10495:
10496: @c Finally, you can define named methods with @code{:m}. One use of this
10497: @c feature is the definition of words that occur only in one class and are
10498: @c not intended to be overridden, but which still need method context
10499: @c (e.g., for accessing @code{inst-var}s). Another use is for methods that
10500: @c would be bound frequently, if defined anonymously.
10501:
10502:
10503: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
10504: @subsubsection Classes and Scoping
10505: @cindex classes and scoping
10506: @cindex scoping and classes
10507:
10508: Inheritance is frequent, unlike structure extension. This exacerbates
10509: the problem with the field name convention (@pxref{Structure Naming
10510: Convention}): One always has to remember in which class the field was
10511: originally defined; changing a part of the class structure would require
10512: changes for renaming in otherwise unaffected code.
10513:
10514: @cindex @code{inst-var} visibility
10515: @cindex @code{inst-value} visibility
10516: To solve this problem, I added a scoping mechanism (which was not in my
10517: original charter): A field defined with @code{inst-var} (or
10518: @code{inst-value}) is visible only in the class where it is defined and in
10519: the descendent classes of this class. Using such fields only makes
10520: sense in @code{m:}-defined methods in these classes anyway.
10521:
10522: This scoping mechanism allows us to use the unadorned field name,
10523: because name clashes with unrelated words become much less likely.
10524:
10525: @cindex @code{protected} discussion
10526: @cindex @code{private} discussion
10527: Once we have this mechanism, we can also use it for controlling the
10528: visibility of other words: All words defined after
10529: @code{protected} are visible only in the current class and its
10530: descendents. @code{public} restores the compilation
10531: (i.e. @code{current}) word list that was in effect before. If you
10532: have several @code{protected}s without an intervening
10533: @code{public} or @code{set-current}, @code{public}
10534: will restore the compilation word list in effect before the first of
10535: these @code{protected}s.
10536:
10537: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10538: @subsubsection Dividing classes
10539: @cindex Dividing classes
10540: @cindex @code{methods}...@code{end-methods}
10541:
10542: You may want to do the definition of methods separate from the
10543: definition of the class, its selectors, fields, and instance variables,
10544: i.e., separate the implementation from the definition. You can do this
10545: in the following way:
10546:
10547: @example
10548: graphical class
10549: inst-value radius
10550: end-class circle
10551:
10552: ... \ do some other stuff
10553:
10554: circle methods \ now we are ready
10555:
10556: m: ( x y circle -- )
10557: radius draw-circle ;m
10558: overrides draw
10559:
10560: m: ( n-radius circle -- )
10561: [to-inst] radius ;m
10562: overrides construct
10563:
10564: end-methods
10565: @end example
10566:
10567: You can use several @code{methods}...@code{end-methods} sections. The
10568: only things you can do to the class in these sections are: defining
10569: methods, and overriding the class's selectors. You must not define new
10570: selectors or fields.
10571:
10572: Note that you often have to override a selector before using it. In
10573: particular, you usually have to override @code{construct} with a new
10574: method before you can invoke @code{heap-new} and friends. E.g., you
10575: must not create a circle before the @code{overrides construct} sequence
10576: in the example above.
10577:
10578: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
10579: @subsubsection Object Interfaces
10580: @cindex object interfaces
10581: @cindex interfaces for objects
10582:
10583: In this model you can only call selectors defined in the class of the
10584: receiving objects or in one of its ancestors. If you call a selector
10585: with a receiving object that is not in one of these classes, the
10586: result is undefined; if you are lucky, the program crashes
10587: immediately.
10588:
10589: @cindex selectors common to hardly-related classes
10590: Now consider the case when you want to have a selector (or several)
10591: available in two classes: You would have to add the selector to a
10592: common ancestor class, in the worst case to @code{object}. You
10593: may not want to do this, e.g., because someone else is responsible for
10594: this ancestor class.
10595:
10596: The solution for this problem is interfaces. An interface is a
10597: collection of selectors. If a class implements an interface, the
10598: selectors become available to the class and its descendents. A class
10599: can implement an unlimited number of interfaces. For the problem
10600: discussed above, we would define an interface for the selector(s), and
10601: both classes would implement the interface.
10602:
10603: As an example, consider an interface @code{storage} for
10604: writing objects to disk and getting them back, and a class
10605: @code{foo} that implements it. The code would look like this:
10606:
10607: @cindex @code{interface} usage
10608: @cindex @code{end-interface} usage
10609: @cindex @code{implementation} usage
10610: @example
10611: interface
10612: selector write ( file object -- )
10613: selector read1 ( file object -- )
10614: end-interface storage
10615:
10616: bar class
10617: storage implementation
10618:
10619: ... overrides write
10620: ... overrides read1
10621: ...
10622: end-class foo
10623: @end example
10624:
10625: @noindent
10626: (I would add a word @code{read} @i{( file -- object )} that uses
10627: @code{read1} internally, but that's beyond the point illustrated
10628: here.)
10629:
10630: Note that you cannot use @code{protected} in an interface; and
10631: of course you cannot define fields.
10632:
10633: In the Neon model, all selectors are available for all classes;
10634: therefore it does not need interfaces. The price you pay in this model
10635: is slower late binding, and therefore, added complexity to avoid late
10636: binding.
10637:
10638: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10639: @subsubsection @file{objects.fs} Implementation
10640: @cindex @file{objects.fs} implementation
10641:
10642: @cindex @code{object-map} discussion
10643: An object is a piece of memory, like one of the data structures
10644: described with @code{struct...end-struct}. It has a field
10645: @code{object-map} that points to the method map for the object's
10646: class.
10647:
10648: @cindex method map
10649: @cindex virtual function table
10650: The @emph{method map}@footnote{This is Self terminology; in C++
10651: terminology: virtual function table.} is an array that contains the
10652: execution tokens (@i{xt}s) of the methods for the object's class. Each
10653: selector contains an offset into a method map.
10654:
10655: @cindex @code{selector} implementation, class
10656: @code{selector} is a defining word that uses
10657: @code{CREATE} and @code{DOES>}. The body of the
10658: selector contains the offset; the @code{DOES>} action for a
10659: class selector is, basically:
10660:
10661: @example
10662: ( object addr ) @@ over object-map @@ + @@ execute
10663: @end example
10664:
10665: Since @code{object-map} is the first field of the object, it
10666: does not generate any code. As you can see, calling a selector has a
10667: small, constant cost.
10668:
10669: @cindex @code{current-interface} discussion
10670: @cindex class implementation and representation
10671: A class is basically a @code{struct} combined with a method
10672: map. During the class definition the alignment and size of the class
10673: are passed on the stack, just as with @code{struct}s, so
10674: @code{field} can also be used for defining class
10675: fields. However, passing more items on the stack would be
10676: inconvenient, so @code{class} builds a data structure in memory,
10677: which is accessed through the variable
10678: @code{current-interface}. After its definition is complete, the
10679: class is represented on the stack by a pointer (e.g., as parameter for
10680: a child class definition).
10681:
10682: A new class starts off with the alignment and size of its parent,
10683: and a copy of the parent's method map. Defining new fields extends the
10684: size and alignment; likewise, defining new selectors extends the
10685: method map. @code{overrides} just stores a new @i{xt} in the method
10686: map at the offset given by the selector.
10687:
10688: @cindex class binding, implementation
10689: Class binding just gets the @i{xt} at the offset given by the selector
10690: from the class's method map and @code{compile,}s (in the case of
10691: @code{[bind]}) it.
10692:
10693: @cindex @code{this} implementation
10694: @cindex @code{catch} and @code{this}
10695: @cindex @code{this} and @code{catch}
10696: I implemented @code{this} as a @code{value}. At the
10697: start of an @code{m:...;m} method the old @code{this} is
10698: stored to the return stack and restored at the end; and the object on
10699: the TOS is stored @code{TO this}. This technique has one
10700: disadvantage: If the user does not leave the method via
10701: @code{;m}, but via @code{throw} or @code{exit},
10702: @code{this} is not restored (and @code{exit} may
10703: crash). To deal with the @code{throw} problem, I have redefined
10704: @code{catch} to save and restore @code{this}; the same
10705: should be done with any word that can catch an exception. As for
10706: @code{exit}, I simply forbid it (as a replacement, there is
10707: @code{exitm}).
10708:
10709: @cindex @code{inst-var} implementation
10710: @code{inst-var} is just the same as @code{field}, with
10711: a different @code{DOES>} action:
10712: @example
10713: @@ this +
10714: @end example
10715: Similar for @code{inst-value}.
10716:
10717: @cindex class scoping implementation
10718: Each class also has a word list that contains the words defined with
10719: @code{inst-var} and @code{inst-value}, and its protected
10720: words. It also has a pointer to its parent. @code{class} pushes
10721: the word lists of the class and all its ancestors onto the search order stack,
10722: and @code{end-class} drops them.
10723:
10724: @cindex interface implementation
10725: An interface is like a class without fields, parent and protected
10726: words; i.e., it just has a method map. If a class implements an
10727: interface, its method map contains a pointer to the method map of the
10728: interface. The positive offsets in the map are reserved for class
10729: methods, therefore interface map pointers have negative
10730: offsets. Interfaces have offsets that are unique throughout the
10731: system, unlike class selectors, whose offsets are only unique for the
10732: classes where the selector is available (invokable).
10733:
10734: This structure means that interface selectors have to perform one
10735: indirection more than class selectors to find their method. Their body
10736: contains the interface map pointer offset in the class method map, and
10737: the method offset in the interface method map. The
10738: @code{does>} action for an interface selector is, basically:
10739:
10740: @example
10741: ( object selector-body )
10742: 2dup selector-interface @@ ( object selector-body object interface-offset )
10743: swap object-map @@ + @@ ( object selector-body map )
10744: swap selector-offset @@ + @@ execute
10745: @end example
10746:
10747: where @code{object-map} and @code{selector-offset} are
10748: first fields and generate no code.
10749:
10750: As a concrete example, consider the following code:
10751:
10752: @example
10753: interface
10754: selector if1sel1
10755: selector if1sel2
10756: end-interface if1
10757:
10758: object class
10759: if1 implementation
10760: selector cl1sel1
10761: cell% inst-var cl1iv1
10762:
10763: ' m1 overrides construct
10764: ' m2 overrides if1sel1
10765: ' m3 overrides if1sel2
10766: ' m4 overrides cl1sel2
10767: end-class cl1
10768:
10769: create obj1 object dict-new drop
10770: create obj2 cl1 dict-new drop
10771: @end example
10772:
10773: The data structure created by this code (including the data structure
10774: for @code{object}) is shown in the
10775: @uref{objects-implementation.eps,figure}, assuming a cell size of 4.
10776: @comment TODO add this diagram..
10777:
10778: @node Objects Glossary, , Objects Implementation, Objects
10779: @subsubsection @file{objects.fs} Glossary
10780: @cindex @file{objects.fs} Glossary
10781:
10782:
10783: doc---objects-bind
10784: doc---objects-<bind>
10785: doc---objects-bind'
10786: doc---objects-[bind]
10787: doc---objects-class
10788: doc---objects-class->map
10789: doc---objects-class-inst-size
10790: doc---objects-class-override!
10791: doc---objects-class-previous
10792: doc---objects-class>order
10793: doc---objects-construct
10794: doc---objects-current'
10795: doc---objects-[current]
10796: doc---objects-current-interface
10797: doc---objects-dict-new
10798: doc---objects-end-class
10799: doc---objects-end-class-noname
10800: doc---objects-end-interface
10801: doc---objects-end-interface-noname
10802: doc---objects-end-methods
10803: doc---objects-exitm
10804: doc---objects-heap-new
10805: doc---objects-implementation
10806: doc---objects-init-object
10807: doc---objects-inst-value
10808: doc---objects-inst-var
10809: doc---objects-interface
10810: doc---objects-m:
10811: doc---objects-:m
10812: doc---objects-;m
10813: doc---objects-method
10814: doc---objects-methods
10815: doc---objects-object
10816: doc---objects-overrides
10817: doc---objects-[parent]
10818: doc---objects-print
10819: doc---objects-protected
10820: doc---objects-public
10821: doc---objects-selector
10822: doc---objects-this
10823: doc---objects-<to-inst>
10824: doc---objects-[to-inst]
10825: doc---objects-to-this
10826: doc---objects-xt-new
10827:
10828:
10829: @c -------------------------------------------------------------
10830: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10831: @subsection The @file{oof.fs} model
10832: @cindex oof
10833: @cindex object-oriented programming
10834:
10835: @cindex @file{objects.fs}
10836: @cindex @file{oof.fs}
10837:
10838: This section describes the @file{oof.fs} package.
10839:
10840: The package described in this section has been used in bigFORTH since 1991, and
10841: used for two large applications: a chromatographic system used to
10842: create new medicaments, and a graphic user interface library (MINOS).
10843:
10844: You can find a description (in German) of @file{oof.fs} in @cite{Object
10845: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10846: 10(2), 1994.
10847:
10848: @menu
10849: * Properties of the OOF model::
10850: * Basic OOF Usage::
10851: * The OOF base class::
10852: * Class Declaration::
10853: * Class Implementation::
10854: @end menu
10855:
10856: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
10857: @subsubsection Properties of the @file{oof.fs} model
10858: @cindex @file{oof.fs} properties
10859:
10860: @itemize @bullet
10861: @item
10862: This model combines object oriented programming with information
10863: hiding. It helps you writing large application, where scoping is
10864: necessary, because it provides class-oriented scoping.
10865:
10866: @item
10867: Named objects, object pointers, and object arrays can be created,
10868: selector invocation uses the ``object selector'' syntax. Selector invocation
10869: to objects and/or selectors on the stack is a bit less convenient, but
10870: possible.
10871:
10872: @item
10873: Selector invocation and instance variable usage of the active object is
10874: straightforward, since both make use of the active object.
10875:
10876: @item
10877: Late binding is efficient and easy to use.
10878:
10879: @item
10880: State-smart objects parse selectors. However, extensibility is provided
10881: using a (parsing) selector @code{postpone} and a selector @code{'}.
10882:
10883: @item
10884: An implementation in ANS Forth is available.
10885:
10886: @end itemize
10887:
10888:
10889: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
10890: @subsubsection Basic @file{oof.fs} Usage
10891: @cindex @file{oof.fs} usage
10892:
10893: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
10894:
10895: You can define a class for graphical objects like this:
10896:
10897: @cindex @code{class} usage
10898: @cindex @code{class;} usage
10899: @cindex @code{method} usage
10900: @example
10901: object class graphical \ "object" is the parent class
10902: method draw ( x y -- )
10903: class;
10904: @end example
10905:
10906: This code defines a class @code{graphical} with an
10907: operation @code{draw}. We can perform the operation
10908: @code{draw} on any @code{graphical} object, e.g.:
10909:
10910: @example
10911: 100 100 t-rex draw
10912: @end example
10913:
10914: @noindent
10915: where @code{t-rex} is an object or object pointer, created with e.g.
10916: @code{graphical : t-rex}.
10917:
10918: @cindex abstract class
10919: How do we create a graphical object? With the present definitions,
10920: we cannot create a useful graphical object. The class
10921: @code{graphical} describes graphical objects in general, but not
10922: any concrete graphical object type (C++ users would call it an
10923: @emph{abstract class}); e.g., there is no method for the selector
10924: @code{draw} in the class @code{graphical}.
10925:
10926: For concrete graphical objects, we define child classes of the
10927: class @code{graphical}, e.g.:
10928:
10929: @example
10930: graphical class circle \ "graphical" is the parent class
10931: cell var circle-radius
10932: how:
10933: : draw ( x y -- )
10934: circle-radius @@ draw-circle ;
10935:
10936: : init ( n-radius -- )
10937: circle-radius ! ;
10938: class;
10939: @end example
10940:
10941: Here we define a class @code{circle} as a child of @code{graphical},
10942: with a field @code{circle-radius}; it defines new methods for the
10943: selectors @code{draw} and @code{init} (@code{init} is defined in
10944: @code{object}, the parent class of @code{graphical}).
10945:
10946: Now we can create a circle in the dictionary with:
10947:
10948: @example
10949: 50 circle : my-circle
10950: @end example
10951:
10952: @noindent
10953: @code{:} invokes @code{init}, thus initializing the field
10954: @code{circle-radius} with 50. We can draw this new circle at (100,100)
10955: with:
10956:
10957: @example
10958: 100 100 my-circle draw
10959: @end example
10960:
10961: @cindex selector invocation, restrictions
10962: @cindex class definition, restrictions
10963: Note: You can only invoke a selector if the receiving object belongs to
10964: the class where the selector was defined or one of its descendents;
10965: e.g., you can invoke @code{draw} only for objects belonging to
10966: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
10967: mechanism will check if you try to invoke a selector that is not
10968: defined in this class hierarchy, so you'll get an error at compilation
10969: time.
10970:
10971:
10972: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
10973: @subsubsection The @file{oof.fs} base class
10974: @cindex @file{oof.fs} base class
10975:
10976: When you define a class, you have to specify a parent class. So how do
10977: you start defining classes? There is one class available from the start:
10978: @code{object}. You have to use it as ancestor for all classes. It is the
10979: only class that has no parent. Classes are also objects, except that
10980: they don't have instance variables; class manipulation such as
10981: inheritance or changing definitions of a class is handled through
10982: selectors of the class @code{object}.
10983:
10984: @code{object} provides a number of selectors:
10985:
10986: @itemize @bullet
10987: @item
10988: @code{class} for subclassing, @code{definitions} to add definitions
10989: later on, and @code{class?} to get type informations (is the class a
10990: subclass of the class passed on the stack?).
10991:
10992: doc---object-class
10993: doc---object-definitions
10994: doc---object-class?
10995:
10996:
10997: @item
10998: @code{init} and @code{dispose} as constructor and destructor of the
10999: object. @code{init} is invocated after the object's memory is allocated,
11000: while @code{dispose} also handles deallocation. Thus if you redefine
11001: @code{dispose}, you have to call the parent's dispose with @code{super
11002: dispose}, too.
11003:
11004: doc---object-init
11005: doc---object-dispose
11006:
11007:
11008: @item
11009: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
11010: @code{[]} to create named and unnamed objects and object arrays or
11011: object pointers.
11012:
11013: doc---object-new
11014: doc---object-new[]
11015: doc---object-:
11016: doc---object-ptr
11017: doc---object-asptr
11018: doc---object-[]
11019:
11020:
11021: @item
11022: @code{::} and @code{super} for explicit scoping. You should use explicit
11023: scoping only for super classes or classes with the same set of instance
11024: variables. Explicitly-scoped selectors use early binding.
11025:
11026: doc---object-::
11027: doc---object-super
11028:
11029:
11030: @item
11031: @code{self} to get the address of the object
11032:
11033: doc---object-self
11034:
11035:
11036: @item
11037: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
11038: pointers and instance defers.
11039:
11040: doc---object-bind
11041: doc---object-bound
11042: doc---object-link
11043: doc---object-is
11044:
11045:
11046: @item
11047: @code{'} to obtain selector tokens, @code{send} to invocate selectors
11048: form the stack, and @code{postpone} to generate selector invocation code.
11049:
11050: doc---object-'
11051: doc---object-postpone
11052:
11053:
11054: @item
11055: @code{with} and @code{endwith} to select the active object from the
11056: stack, and enable its scope. Using @code{with} and @code{endwith}
11057: also allows you to create code using selector @code{postpone} without being
11058: trapped by the state-smart objects.
11059:
11060: doc---object-with
11061: doc---object-endwith
11062:
11063:
11064: @end itemize
11065:
11066: @node Class Declaration, Class Implementation, The OOF base class, OOF
11067: @subsubsection Class Declaration
11068: @cindex class declaration
11069:
11070: @itemize @bullet
11071: @item
11072: Instance variables
11073:
11074: doc---oof-var
11075:
11076:
11077: @item
11078: Object pointers
11079:
11080: doc---oof-ptr
11081: doc---oof-asptr
11082:
11083:
11084: @item
11085: Instance defers
11086:
11087: doc---oof-defer
11088:
11089:
11090: @item
11091: Method selectors
11092:
11093: doc---oof-early
11094: doc---oof-method
11095:
11096:
11097: @item
11098: Class-wide variables
11099:
11100: doc---oof-static
11101:
11102:
11103: @item
11104: End declaration
11105:
11106: doc---oof-how:
11107: doc---oof-class;
11108:
11109:
11110: @end itemize
11111:
11112: @c -------------------------------------------------------------
11113: @node Class Implementation, , Class Declaration, OOF
11114: @subsubsection Class Implementation
11115: @cindex class implementation
11116:
11117: @c -------------------------------------------------------------
11118: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
11119: @subsection The @file{mini-oof.fs} model
11120: @cindex mini-oof
11121:
11122: Gforth's third object oriented Forth package is a 12-liner. It uses a
11123: mixture of the @file{objects.fs} and the @file{oof.fs} syntax,
11124: and reduces to the bare minimum of features. This is based on a posting
11125: of Bernd Paysan in comp.lang.forth.
11126:
11127: @menu
11128: * Basic Mini-OOF Usage::
11129: * Mini-OOF Example::
11130: * Mini-OOF Implementation::
11131: @end menu
11132:
11133: @c -------------------------------------------------------------
11134: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
11135: @subsubsection Basic @file{mini-oof.fs} Usage
11136: @cindex mini-oof usage
11137:
11138: There is a base class (@code{class}, which allocates one cell for the
11139: object pointer) plus seven other words: to define a method, a variable,
11140: a class; to end a class, to resolve binding, to allocate an object and
11141: to compile a class method.
11142: @comment TODO better description of the last one
11143:
11144:
11145: doc-object
11146: doc-method
11147: doc-var
11148: doc-class
11149: doc-end-class
11150: doc-defines
11151: doc-new
11152: doc-::
11153:
11154:
11155:
11156: @c -------------------------------------------------------------
11157: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
11158: @subsubsection Mini-OOF Example
11159: @cindex mini-oof example
11160:
11161: A short example shows how to use this package. This example, in slightly
11162: extended form, is supplied as @file{moof-exm.fs}
11163: @comment TODO could flesh this out with some comments from the Forthwrite article
11164:
11165: @example
11166: object class
11167: method init
11168: method draw
11169: end-class graphical
11170: @end example
11171:
11172: This code defines a class @code{graphical} with an
11173: operation @code{draw}. We can perform the operation
11174: @code{draw} on any @code{graphical} object, e.g.:
11175:
11176: @example
11177: 100 100 t-rex draw
11178: @end example
11179:
11180: where @code{t-rex} is an object or object pointer, created with e.g.
11181: @code{graphical new Constant t-rex}.
11182:
11183: For concrete graphical objects, we define child classes of the
11184: class @code{graphical}, e.g.:
11185:
11186: @example
11187: graphical class
11188: cell var circle-radius
11189: end-class circle \ "graphical" is the parent class
11190:
11191: :noname ( x y -- )
11192: circle-radius @@ draw-circle ; circle defines draw
11193: :noname ( r -- )
11194: circle-radius ! ; circle defines init
11195: @end example
11196:
11197: There is no implicit init method, so we have to define one. The creation
11198: code of the object now has to call init explicitely.
11199:
11200: @example
11201: circle new Constant my-circle
11202: 50 my-circle init
11203: @end example
11204:
11205: It is also possible to add a function to create named objects with
11206: automatic call of @code{init}, given that all objects have @code{init}
11207: on the same place:
11208:
11209: @example
11210: : new: ( .. o "name" -- )
11211: new dup Constant init ;
11212: 80 circle new: large-circle
11213: @end example
11214:
11215: We can draw this new circle at (100,100) with:
11216:
11217: @example
11218: 100 100 my-circle draw
11219: @end example
11220:
11221: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
11222: @subsubsection @file{mini-oof.fs} Implementation
11223:
11224: Object-oriented systems with late binding typically use a
11225: ``vtable''-approach: the first variable in each object is a pointer to a
11226: table, which contains the methods as function pointers. The vtable
11227: may also contain other information.
11228:
11229: So first, let's declare selectors:
11230:
11231: @example
11232: : method ( m v "name" -- m' v ) Create over , swap cell+ swap
11233: DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
11234: @end example
11235:
11236: During selector declaration, the number of selectors and instance
11237: variables is on the stack (in address units). @code{method} creates one
11238: selector and increments the selector number. To execute a selector, it
11239: takes the object, fetches the vtable pointer, adds the offset, and
11240: executes the method @i{xt} stored there. Each selector takes the object
11241: it is invoked with as top of stack parameter; it passes the parameters
11242: (including the object) unchanged to the appropriate method which should
11243: consume that object.
11244:
11245: Now, we also have to declare instance variables
11246:
11247: @example
11248: : var ( m v size "name" -- m v' ) Create over , +
11249: DOES> ( o -- addr ) @@ + ;
11250: @end example
11251:
11252: As before, a word is created with the current offset. Instance
11253: variables can have different sizes (cells, floats, doubles, chars), so
11254: all we do is take the size and add it to the offset. If your machine
11255: has alignment restrictions, put the proper @code{aligned} or
11256: @code{faligned} before the variable, to adjust the variable
11257: offset. That's why it is on the top of stack.
11258:
11259: We need a starting point (the base object) and some syntactic sugar:
11260:
11261: @example
11262: Create object 1 cells , 2 cells ,
11263: : class ( class -- class selectors vars ) dup 2@@ ;
11264: @end example
11265:
11266: For inheritance, the vtable of the parent object has to be
11267: copied when a new, derived class is declared. This gives all the
11268: methods of the parent class, which can be overridden, though.
11269:
11270: @example
11271: : end-class ( class selectors vars "name" -- )
11272: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
11273: cell+ dup cell+ r> rot @@ 2 cells /string move ;
11274: @end example
11275:
11276: The first line creates the vtable, initialized with
11277: @code{noop}s. The second line is the inheritance mechanism, it
11278: copies the xts from the parent vtable.
11279:
11280: We still have no way to define new methods, let's do that now:
11281:
11282: @example
11283: : defines ( xt class "name" -- ) ' >body @@ + ! ;
11284: @end example
11285:
11286: To allocate a new object, we need a word, too:
11287:
11288: @example
11289: : new ( class -- o ) here over @@ allot swap over ! ;
11290: @end example
11291:
11292: Sometimes derived classes want to access the method of the
11293: parent object. There are two ways to achieve this with Mini-OOF:
11294: first, you could use named words, and second, you could look up the
11295: vtable of the parent object.
11296:
11297: @example
11298: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
11299: @end example
11300:
11301:
11302: Nothing can be more confusing than a good example, so here is
11303: one. First let's declare a text object (called
11304: @code{button}), that stores text and position:
11305:
11306: @example
11307: object class
11308: cell var text
11309: cell var len
11310: cell var x
11311: cell var y
11312: method init
11313: method draw
11314: end-class button
11315: @end example
11316:
11317: @noindent
11318: Now, implement the two methods, @code{draw} and @code{init}:
11319:
11320: @example
11321: :noname ( o -- )
11322: >r r@@ x @@ r@@ y @@ at-xy r@@ text @@ r> len @@ type ;
11323: button defines draw
11324: :noname ( addr u o -- )
11325: >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
11326: button defines init
11327: @end example
11328:
11329: @noindent
11330: To demonstrate inheritance, we define a class @code{bold-button}, with no
11331: new data and no new selectors:
11332:
11333: @example
11334: button class
11335: end-class bold-button
11336:
11337: : bold 27 emit ." [1m" ;
11338: : normal 27 emit ." [0m" ;
11339: @end example
11340:
11341: @noindent
11342: The class @code{bold-button} has a different draw method to
11343: @code{button}, but the new method is defined in terms of the draw method
11344: for @code{button}:
11345:
11346: @example
11347: :noname bold [ button :: draw ] normal ; bold-button defines draw
11348: @end example
11349:
11350: @noindent
11351: Finally, create two objects and apply selectors:
11352:
11353: @example
11354: button new Constant foo
11355: s" thin foo" foo init
11356: page
11357: foo draw
11358: bold-button new Constant bar
11359: s" fat bar" bar init
11360: 1 bar y !
11361: bar draw
11362: @end example
11363:
11364:
11365: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11366: @subsection Comparison with other object models
11367: @cindex comparison of object models
11368: @cindex object models, comparison
11369:
11370: Many object-oriented Forth extensions have been proposed (@cite{A survey
11371: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11372: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11373: relation of the object models described here to two well-known and two
11374: closely-related (by the use of method maps) models. Andras Zsoter
11375: helped us with this section.
11376:
11377: @cindex Neon model
11378: The most popular model currently seems to be the Neon model (see
11379: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11380: 1997) by Andrew McKewan) but this model has a number of limitations
11381: @footnote{A longer version of this critique can be
11382: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11383: Dimensions, May 1997) by Anton Ertl.}:
11384:
11385: @itemize @bullet
11386: @item
11387: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11388: to pass objects on the stack.
11389:
11390: @item
11391: It requires that the selector parses the input stream (at
11392: compile time); this leads to reduced extensibility and to bugs that are
11393: hard to find.
11394:
11395: @item
11396: It allows using every selector on every object; this eliminates the
11397: need for interfaces, but makes it harder to create efficient
11398: implementations.
11399: @end itemize
11400:
11401: @cindex Pountain's object-oriented model
11402: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11403: Press, London, 1987) by Dick Pountain. However, it is not really about
11404: object-oriented programming, because it hardly deals with late
11405: binding. Instead, it focuses on features like information hiding and
11406: overloading that are characteristic of modular languages like Ada (83).
11407:
11408: @cindex Zsoter's object-oriented model
11409: In @uref{http://www.forth.org/oopf.html, Does late binding have to be
11410: slow?} (Forth Dimensions 18(1) 1996, pages 31-35) Andras Zsoter
11411: describes a model that makes heavy use of an active object (like
11412: @code{this} in @file{objects.fs}): The active object is not only used
11413: for accessing all fields, but also specifies the receiving object of
11414: every selector invocation; you have to change the active object
11415: explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
11416: changes more or less implicitly at @code{m: ... ;m}. Such a change at
11417: the method entry point is unnecessary with Zsoter's model, because the
11418: receiving object is the active object already. On the other hand, the
11419: explicit change is absolutely necessary in that model, because otherwise
11420: no one could ever change the active object. An ANS Forth implementation
11421: of this model is available through
11422: @uref{http://www.forth.org/oopf.html}.
11423:
11424: @cindex @file{oof.fs}, differences to other models
11425: The @file{oof.fs} model combines information hiding and overloading
11426: resolution (by keeping names in various word lists) with object-oriented
11427: programming. It sets the active object implicitly on method entry, but
11428: also allows explicit changing (with @code{>o...o>} or with
11429: @code{with...endwith}). It uses parsing and state-smart objects and
11430: classes for resolving overloading and for early binding: the object or
11431: class parses the selector and determines the method from this. If the
11432: selector is not parsed by an object or class, it performs a call to the
11433: selector for the active object (late binding), like Zsoter's model.
11434: Fields are always accessed through the active object. The big
11435: disadvantage of this model is the parsing and the state-smartness, which
11436: reduces extensibility and increases the opportunities for subtle bugs;
11437: essentially, you are only safe if you never tick or @code{postpone} an
11438: object or class (Bernd disagrees, but I (Anton) am not convinced).
11439:
11440: @cindex @file{mini-oof.fs}, differences to other models
11441: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11442: version of the @file{objects.fs} model, but syntactically it is a
11443: mixture of the @file{objects.fs} and @file{oof.fs} models.
11444:
11445:
11446: @c -------------------------------------------------------------
11447: @node Programming Tools, C Interface, Object-oriented Forth, Words
11448: @section Programming Tools
11449: @cindex programming tools
11450:
11451: @c !! move this and assembler down below OO stuff.
11452:
11453: @menu
11454: * Examining:: Data and Code.
11455: * Forgetting words:: Usually before reloading.
11456: * Debugging:: Simple and quick.
11457: * Assertions:: Making your programs self-checking.
11458: * Singlestep Debugger:: Executing your program word by word.
11459: @end menu
11460:
11461: @node Examining, Forgetting words, Programming Tools, Programming Tools
11462: @subsection Examining data and code
11463: @cindex examining data and code
11464: @cindex data examination
11465: @cindex code examination
11466:
11467: The following words inspect the stack non-destructively:
11468:
11469: doc-.s
11470: doc-f.s
11471: doc-maxdepth-.s
11472:
11473: There is a word @code{.r} but it does @i{not} display the return stack!
11474: It is used for formatted numeric output (@pxref{Simple numeric output}).
11475:
11476: doc-depth
11477: doc-fdepth
11478: doc-clearstack
11479: doc-clearstacks
11480:
11481: The following words inspect memory.
11482:
11483: doc-?
11484: doc-dump
11485:
11486: And finally, @code{see} allows to inspect code:
11487:
11488: doc-see
11489: doc-xt-see
11490: doc-simple-see
11491: doc-simple-see-range
11492:
11493: @node Forgetting words, Debugging, Examining, Programming Tools
11494: @subsection Forgetting words
11495: @cindex words, forgetting
11496: @cindex forgeting words
11497:
11498: @c anton: other, maybe better places for this subsection: Defining Words;
11499: @c Dictionary allocation. At least a reference should be there.
11500:
11501: Forth allows you to forget words (and everything that was alloted in the
11502: dictonary after them) in a LIFO manner.
11503:
11504: doc-marker
11505:
11506: The most common use of this feature is during progam development: when
11507: you change a source file, forget all the words it defined and load it
11508: again (since you also forget everything defined after the source file
11509: was loaded, you have to reload that, too). Note that effects like
11510: storing to variables and destroyed system words are not undone when you
11511: forget words. With a system like Gforth, that is fast enough at
11512: starting up and compiling, I find it more convenient to exit and restart
11513: Gforth, as this gives me a clean slate.
11514:
11515: Here's an example of using @code{marker} at the start of a source file
11516: that you are debugging; it ensures that you only ever have one copy of
11517: the file's definitions compiled at any time:
11518:
11519: @example
11520: [IFDEF] my-code
11521: my-code
11522: [ENDIF]
11523:
11524: marker my-code
11525: init-included-files
11526:
11527: \ .. definitions start here
11528: \ .
11529: \ .
11530: \ end
11531: @end example
11532:
11533:
11534: @node Debugging, Assertions, Forgetting words, Programming Tools
11535: @subsection Debugging
11536: @cindex debugging
11537:
11538: Languages with a slow edit/compile/link/test development loop tend to
11539: require sophisticated tracing/stepping debuggers to facilate debugging.
11540:
11541: A much better (faster) way in fast-compiling languages is to add
11542: printing code at well-selected places, let the program run, look at
11543: the output, see where things went wrong, add more printing code, etc.,
11544: until the bug is found.
11545:
11546: The simple debugging aids provided in @file{debugs.fs}
11547: are meant to support this style of debugging.
11548:
11549: The word @code{~~} prints debugging information (by default the source
11550: location and the stack contents). It is easy to insert. If you use Emacs
11551: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
11552: query-replace them with nothing). The deferred words
11553: @code{printdebugdata} and @code{.debugline} control the output of
11554: @code{~~}. The default source location output format works well with
11555: Emacs' compilation mode, so you can step through the program at the
11556: source level using @kbd{C-x `} (the advantage over a stepping debugger
11557: is that you can step in any direction and you know where the crash has
11558: happened or where the strange data has occurred).
11559:
11560: doc-~~
11561: doc-printdebugdata
11562: doc-.debugline
11563:
11564: @cindex filenames in @code{~~} output
11565: @code{~~} (and assertions) will usually print the wrong file name if a
11566: marker is executed in the same file after their occurance. They will
11567: print @samp{*somewhere*} as file name if a marker is executed in the
11568: same file before their occurance.
11569:
11570:
11571: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
11572: @subsection Assertions
11573: @cindex assertions
11574:
11575: It is a good idea to make your programs self-checking, especially if you
11576: make an assumption that may become invalid during maintenance (for
11577: example, that a certain field of a data structure is never zero). Gforth
11578: supports @dfn{assertions} for this purpose. They are used like this:
11579:
11580: @example
11581: assert( @i{flag} )
11582: @end example
11583:
11584: The code between @code{assert(} and @code{)} should compute a flag, that
11585: should be true if everything is alright and false otherwise. It should
11586: not change anything else on the stack. The overall stack effect of the
11587: assertion is @code{( -- )}. E.g.
11588:
11589: @example
11590: assert( 1 1 + 2 = ) \ what we learn in school
11591: assert( dup 0<> ) \ assert that the top of stack is not zero
11592: assert( false ) \ this code should not be reached
11593: @end example
11594:
11595: The need for assertions is different at different times. During
11596: debugging, we want more checking, in production we sometimes care more
11597: for speed. Therefore, assertions can be turned off, i.e., the assertion
11598: becomes a comment. Depending on the importance of an assertion and the
11599: time it takes to check it, you may want to turn off some assertions and
11600: keep others turned on. Gforth provides several levels of assertions for
11601: this purpose:
11602:
11603:
11604: doc-assert0(
11605: doc-assert1(
11606: doc-assert2(
11607: doc-assert3(
11608: doc-assert(
11609: doc-)
11610:
11611:
11612: The variable @code{assert-level} specifies the highest assertions that
11613: are turned on. I.e., at the default @code{assert-level} of one,
11614: @code{assert0(} and @code{assert1(} assertions perform checking, while
11615: @code{assert2(} and @code{assert3(} assertions are treated as comments.
11616:
11617: The value of @code{assert-level} is evaluated at compile-time, not at
11618: run-time. Therefore you cannot turn assertions on or off at run-time;
11619: you have to set the @code{assert-level} appropriately before compiling a
11620: piece of code. You can compile different pieces of code at different
11621: @code{assert-level}s (e.g., a trusted library at level 1 and
11622: newly-written code at level 3).
11623:
11624:
11625: doc-assert-level
11626:
11627:
11628: If an assertion fails, a message compatible with Emacs' compilation mode
11629: is produced and the execution is aborted (currently with @code{ABORT"}.
11630: If there is interest, we will introduce a special throw code. But if you
11631: intend to @code{catch} a specific condition, using @code{throw} is
11632: probably more appropriate than an assertion).
11633:
11634: @cindex filenames in assertion output
11635: Assertions (and @code{~~}) will usually print the wrong file name if a
11636: marker is executed in the same file after their occurance. They will
11637: print @samp{*somewhere*} as file name if a marker is executed in the
11638: same file before their occurance.
11639:
11640: Definitions in ANS Forth for these assertion words are provided
11641: in @file{compat/assert.fs}.
11642:
11643:
11644: @node Singlestep Debugger, , Assertions, Programming Tools
11645: @subsection Singlestep Debugger
11646: @cindex singlestep Debugger
11647: @cindex debugging Singlestep
11648:
11649: The singlestep debugger works only with the engine @code{gforth-ditc}.
11650:
11651: When you create a new word there's often the need to check whether it
11652: behaves correctly or not. You can do this by typing @code{dbg
11653: badword}. A debug session might look like this:
11654:
11655: @example
11656: : badword 0 DO i . LOOP ; ok
11657: 2 dbg badword
11658: : badword
11659: Scanning code...
11660:
11661: Nesting debugger ready!
11662:
11663: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
11664: 400D4740 8049F68 DO -> [ 0 ]
11665: 400D4744 804A0C8 i -> [ 1 ] 00000
11666: 400D4748 400C5E60 . -> 0 [ 0 ]
11667: 400D474C 8049D0C LOOP -> [ 0 ]
11668: 400D4744 804A0C8 i -> [ 1 ] 00001
11669: 400D4748 400C5E60 . -> 1 [ 0 ]
11670: 400D474C 8049D0C LOOP -> [ 0 ]
11671: 400D4758 804B384 ; -> ok
11672: @end example
11673:
11674: Each line displayed is one step. You always have to hit return to
11675: execute the next word that is displayed. If you don't want to execute
11676: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
11677: an overview what keys are available:
11678:
11679: @table @i
11680:
11681: @item @key{RET}
11682: Next; Execute the next word.
11683:
11684: @item n
11685: Nest; Single step through next word.
11686:
11687: @item u
11688: Unnest; Stop debugging and execute rest of word. If we got to this word
11689: with nest, continue debugging with the calling word.
11690:
11691: @item d
11692: Done; Stop debugging and execute rest.
11693:
11694: @item s
11695: Stop; Abort immediately.
11696:
11697: @end table
11698:
11699: Debugging large application with this mechanism is very difficult, because
11700: you have to nest very deeply into the program before the interesting part
11701: begins. This takes a lot of time.
11702:
11703: To do it more directly put a @code{BREAK:} command into your source code.
11704: When program execution reaches @code{BREAK:} the single step debugger is
11705: invoked and you have all the features described above.
11706:
11707: If you have more than one part to debug it is useful to know where the
11708: program has stopped at the moment. You can do this by the
11709: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
11710: string is typed out when the ``breakpoint'' is reached.
11711:
11712:
11713: doc-dbg
11714: doc-break:
11715: doc-break"
11716:
11717: @c ------------------------------------------------------------
11718: @node C Interface, Assembler and Code Words, Programming Tools, Words
11719: @section C Interface
11720: @cindex C interface
11721: @cindex foreign language interface
11722: @cindex interface to C functions
11723:
11724: Note that the C interface is not yet complete; a better way of
11725: declaring C functions is planned, as well as a way of declaring
11726: structs, unions, and their fields.
11727:
11728: @menu
11729: * Calling C Functions::
11730: * Declaring C Functions::
11731: * Callbacks::
11732: * Low-Level C Interface Words::
11733: @end menu
11734:
11735: @node Calling C Functions, Declaring C Functions, C Interface, C Interface
11736: @subsection Calling C functions
11737: @cindex C functions, calls to
11738: @cindex calling C functions
11739:
11740: Once a C function is declared (see @pxref{Declaring C Functions}), you
11741: can call it as follows: You push the arguments on the stack(s), and
11742: then call the word for the C function. The arguments have to be
11743: pushed in the same order as the arguments appear in the C
11744: documentation (i.e., the first argument is deepest on the stack).
11745: Integer and pointer arguments have to be pushed on the data stack,
11746: floating-point arguments on the FP stack; these arguments are consumed
11747: by the called C function.
11748:
11749: On returning from the C function, the return value, if any, resides on
11750: the appropriate stack: an integer return value is pushed on the data
11751: stack, an FP return value on the FP stack, and a void return value
11752: results in not pushing anything. Note that most C functions have a
11753: return value, even if that is often not used in C; in Forth, you have
11754: to @code{drop} this return value explicitly if you do not use it.
11755:
11756: By default, an integer argument or return value corresponds to a
11757: single cell, and a floating-point argument or return value corresponds
11758: to a Forth float value; the C interface performs the appropriate
11759: conversions where necessary, on a best-effort basis (in some cases,
11760: there may be some loss).
11761:
11762: As an example, consider the POSIX function @code{lseek()}:
11763:
11764: @example
11765: off_t lseek(int fd, off_t offset, int whence);
11766: @end example
11767:
11768: This function takes three integer arguments, and returns an integer
11769: argument, so a Forth call for setting the current file offset to the
11770: start of the file could look like this:
11771:
11772: @example
11773: fd @@ 0 SEEK_SET lseek -1 = if
11774: ... \ error handling
11775: then
11776: @end example
11777:
11778: You might be worried that an @code{off_t} does not fit into a cell, so
11779: you could not pass larger offsets to lseek, and might get only a part
11780: of the return values. In that case, in your declaration of the
11781: function (@pxref{Declaring C Functions}) you should declare it to use
11782: double-cells for the off_t argument and return value, and maybe give
11783: the resulting Forth word a different name, like @code{dlseek}; the
11784: result could be called like this:
11785:
11786: @example
11787: fd @@ 0. SEEK_SET dlseek -1. d= if
11788: ... \ error handling
11789: then
11790: @end example
11791:
11792: Passing and returning structs or unions is currently not supported by
11793: our interface@footnote{If you know the calling convention of your C
11794: compiler, you usually can call such functions in some way, but that
11795: way is usually not portable between platforms, and sometimes not even
11796: between C compilers.}.
11797:
11798: Calling functions with a variable number of arguments (e.g.,
11799: @code{printf()}) is currently only supported by having you declare one
11800: function-calling word for each argument pattern, and calling the
11801: appropriate word for the desired pattern.
11802:
11803:
11804: @node Declaring C Functions, Callbacks, Calling C Functions, C Interface
11805: @subsection Declaring C Functions
11806: @cindex C functions, declarations
11807: @cindex declaring C functions
11808:
11809: Before you can call @code{lseek} or @code{dlseek}, you have to declare
11810: it. You have to look up in your system what the concrete type for the
11811: abstract type @code{off_t} is; let's assume it is @code{long}. Then
11812: the declarations for these words are:
11813:
11814: @example
11815: library libc libc.so.6
11816: libc lseek int long int (long) lseek ( fd noffset whence -- noffset2 )
11817: libc dlseek int dlong int (dlong) lseek ( fd doffset whence -- doffset2 )
11818: @end example
11819:
11820: The first line defines a Forth word @code{libc} for accessing the C
11821: functions in the shared library @file{libc.so.6} (the name of the
11822: shared library depends on the library and the OS; this example is the
11823: standard C library (containing most of the standard C and Unix
11824: functions) for GNU/Linux systems since about 1998).
11825:
11826: The next two lines define two Forth words for the same C function
11827: @code{lseek()}; the middle line defines @code{lseek ( n1 n2 n3 -- n
11828: )}, and the last line defines @code{dlseek ( n1 d2 n3 -- d )}.
11829:
11830: As you can see, the declarations are relatively platform-dependent
11831: (e.g., on one platform @code{off_t} may be a @code{long}, whereas on
11832: another platform it may be a @code{long long}; actually, in this case
11833: you can have this difference even on the same platform), while the
11834: resulting function-calling words are platform-independent, and calls
11835: to them are portable.
11836:
11837: At some point in the future this interface will be superseded by a
11838: more convenient one with fewer portability issues. But the resulting
11839: words for calling the C function will still have the same interface,
11840: so you will not need to change the calls.
11841:
11842: Anyway, here are the words for the current interface:
11843:
11844: doc-library
11845: doc-int
11846: doc-dint
11847: doc-uint
11848: doc-udint
11849: doc-long
11850: doc-dlong
11851: doc-ulong
11852: doc-udlong
11853: doc-longlong
11854: doc-dlonglong
11855: doc-ulonglong
11856: doc-udlonglong
11857: doc-ptr
11858: doc-cfloat
11859: doc-cdouble
11860: doc-clongdouble
11861: doc-(int)
11862: doc-(dint)
11863: doc-(uint)
11864: doc-(udint)
11865: doc-(long)
11866: doc-(dlong)
11867: doc-(ulong)
11868: doc-(udlong)
11869: doc-(longlong)
11870: doc-(dlonglong)
11871: doc-(ulonglong)
11872: doc-(udlonglong)
11873: doc-(ptr)
11874: doc-(cfloat)
11875: doc-(cdouble)
11876: doc-(clongdouble)
11877:
11878:
11879: @node Callbacks, Low-Level C Interface Words, Declaring C Functions, C Interface
11880: @subsection Callbacks
11881: @cindex Callback functions written in Forth
11882: @cindex C function pointers to Forth words
11883:
11884: In some cases you have to pass a function pointer to a C function,
11885: i.e., the library wants to call back to your application (and the
11886: pointed-to function is called a callback function). You can pass the
11887: address of an existing C function (that you get with @code{lib-sym},
11888: @pxref{Low-Level C Interface Words}), but if there is no appropriate C
11889: function, you probably want to define the function as a Forth word.
11890:
11891: !!!
11892: @c I don't understand the existing callback interface from the example - anton
11893:
11894: doc-callback
11895: doc-callback;
11896: doc-fptr
11897:
11898:
11899: @c > > Und dann gibt's noch die fptr-Deklaration, die einem
11900: @c > > C-Funktionspointer entspricht (Deklaration gleich wie bei
11901: @c > > Library-Funktionen, nur ohne den C-Namen, Aufruf mit der
11902: @c > > C-Funktionsadresse auf dem TOS).
11903: @c >
11904: @c > Ja, da bin ich dann ausgestiegen, weil ich aus dem Beispiel nicht
11905: @c > gesehen habe, wozu das gut ist.
11906: @c
11907: @c Irgendwie muss ich den Callback ja testen. Und es soll ja auch
11908: @c vorkommen, dass man von irgendwelchen kranken Interfaces einen
11909: @c Funktionspointer übergeben bekommt, den man dann bei Gelegenheit
11910: @c aufrufen muss. Also kann man den deklarieren, und das damit deklarierte
11911: @c Wort verhält sich dann wie ein EXECUTE für alle C-Funktionen mit
11912: @c demselben Prototyp.
11913:
11914:
11915: @node Low-Level C Interface Words, , Callbacks, C Interface
11916: @subsection Low-Level C Interface Words
11917:
11918: doc-open-lib
11919: doc-lib-sym
11920:
11921: @c -------------------------------------------------------------
11922: @node Assembler and Code Words, Threading Words, C Interface, Words
11923: @section Assembler and Code Words
11924: @cindex assembler
11925: @cindex code words
11926:
11927: @menu
11928: * Code and ;code::
11929: * Common Assembler:: Assembler Syntax
11930: * Common Disassembler::
11931: * 386 Assembler:: Deviations and special cases
11932: * Alpha Assembler:: Deviations and special cases
11933: * MIPS assembler:: Deviations and special cases
11934: * PowerPC assembler:: Deviations and special cases
11935: * Other assemblers:: How to write them
11936: @end menu
11937:
11938: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
11939: @subsection @code{Code} and @code{;code}
11940:
11941: Gforth provides some words for defining primitives (words written in
11942: machine code), and for defining the machine-code equivalent of
11943: @code{DOES>}-based defining words. However, the machine-independent
11944: nature of Gforth poses a few problems: First of all, Gforth runs on
11945: several architectures, so it can provide no standard assembler. What's
11946: worse is that the register allocation not only depends on the processor,
11947: but also on the @code{gcc} version and options used.
11948:
11949: The words that Gforth offers encapsulate some system dependences (e.g.,
11950: the header structure), so a system-independent assembler may be used in
11951: Gforth. If you do not have an assembler, you can compile machine code
11952: directly with @code{,} and @code{c,}@footnote{This isn't portable,
11953: because these words emit stuff in @i{data} space; it works because
11954: Gforth has unified code/data spaces. Assembler isn't likely to be
11955: portable anyway.}.
11956:
11957:
11958: doc-assembler
11959: doc-init-asm
11960: doc-code
11961: doc-end-code
11962: doc-;code
11963: doc-flush-icache
11964:
11965:
11966: If @code{flush-icache} does not work correctly, @code{code} words
11967: etc. will not work (reliably), either.
11968:
11969: The typical usage of these @code{code} words can be shown most easily by
11970: analogy to the equivalent high-level defining words:
11971:
11972: @example
11973: : foo code foo
11974: <high-level Forth words> <assembler>
11975: ; end-code
11976:
11977: : bar : bar
11978: <high-level Forth words> <high-level Forth words>
11979: CREATE CREATE
11980: <high-level Forth words> <high-level Forth words>
11981: DOES> ;code
11982: <high-level Forth words> <assembler>
11983: ; end-code
11984: @end example
11985:
11986: @c anton: the following stuff is also in "Common Assembler", in less detail.
11987:
11988: @cindex registers of the inner interpreter
11989: In the assembly code you will want to refer to the inner interpreter's
11990: registers (e.g., the data stack pointer) and you may want to use other
11991: registers for temporary storage. Unfortunately, the register allocation
11992: is installation-dependent.
11993:
11994: In particular, @code{ip} (Forth instruction pointer) and @code{rp}
11995: (return stack pointer) may be in different places in @code{gforth} and
11996: @code{gforth-fast}, or different installations. This means that you
11997: cannot write a @code{NEXT} routine that works reliably on both versions
11998: or different installations; so for doing @code{NEXT}, I recommend
11999: jumping to @code{' noop >code-address}, which contains nothing but a
12000: @code{NEXT}.
12001:
12002: For general accesses to the inner interpreter's registers, the easiest
12003: solution is to use explicit register declarations (@pxref{Explicit Reg
12004: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) for
12005: all of the inner interpreter's registers: You have to compile Gforth
12006: with @code{-DFORCE_REG} (configure option @code{--enable-force-reg}) and
12007: the appropriate declarations must be present in the @code{machine.h}
12008: file (see @code{mips.h} for an example; you can find a full list of all
12009: declarable register symbols with @code{grep register engine.c}). If you
12010: give explicit registers to all variables that are declared at the
12011: beginning of @code{engine()}, you should be able to use the other
12012: caller-saved registers for temporary storage. Alternatively, you can use
12013: the @code{gcc} option @code{-ffixed-REG} (@pxref{Code Gen Options, ,
12014: Options for Code Generation Conventions, gcc.info, GNU C Manual}) to
12015: reserve a register (however, this restriction on register allocation may
12016: slow Gforth significantly).
12017:
12018: If this solution is not viable (e.g., because @code{gcc} does not allow
12019: you to explicitly declare all the registers you need), you have to find
12020: out by looking at the code where the inner interpreter's registers
12021: reside and which registers can be used for temporary storage. You can
12022: get an assembly listing of the engine's code with @code{make engine.s}.
12023:
12024: In any case, it is good practice to abstract your assembly code from the
12025: actual register allocation. E.g., if the data stack pointer resides in
12026: register @code{$17}, create an alias for this register called @code{sp},
12027: and use that in your assembly code.
12028:
12029: @cindex code words, portable
12030: Another option for implementing normal and defining words efficiently
12031: is to add the desired functionality to the source of Gforth. For normal
12032: words you just have to edit @file{primitives} (@pxref{Automatic
12033: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
12034: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
12035: @file{prims2x.fs}, and possibly @file{cross.fs}.
12036:
12037: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
12038: @subsection Common Assembler
12039:
12040: The assemblers in Gforth generally use a postfix syntax, i.e., the
12041: instruction name follows the operands.
12042:
12043: The operands are passed in the usual order (the same that is used in the
12044: manual of the architecture). Since they all are Forth words, they have
12045: to be separated by spaces; you can also use Forth words to compute the
12046: operands.
12047:
12048: The instruction names usually end with a @code{,}. This makes it easier
12049: to visually separate instructions if you put several of them on one
12050: line; it also avoids shadowing other Forth words (e.g., @code{and}).
12051:
12052: Registers are usually specified by number; e.g., (decimal) @code{11}
12053: specifies registers R11 and F11 on the Alpha architecture (which one,
12054: depends on the instruction). The usual names are also available, e.g.,
12055: @code{s2} for R11 on Alpha.
12056:
12057: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
12058: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
12059: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
12060: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
12061: conditions are specified in a way specific to each assembler.
12062:
12063: Note that the register assignments of the Gforth engine can change
12064: between Gforth versions, or even between different compilations of the
12065: same Gforth version (e.g., if you use a different GCC version). So if
12066: you want to refer to Gforth's registers (e.g., the stack pointer or
12067: TOS), I recommend defining your own words for refering to these
12068: registers, and using them later on; then you can easily adapt to a
12069: changed register assignment. The stability of the register assignment
12070: is usually better if you build Gforth with @code{--enable-force-reg}.
12071:
12072: The most common use of these registers is to dispatch to the next word
12073: (the @code{next} routine). A portable way to do this is to jump to
12074: @code{' noop >code-address} (of course, this is less efficient than
12075: integrating the @code{next} code and scheduling it well).
12076:
12077: Another difference between Gforth version is that the top of stack is
12078: kept in memory in @code{gforth} and, on most platforms, in a register in
12079: @code{gforth-fast}.
12080:
12081: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
12082: @subsection Common Disassembler
12083: @cindex disassembler, general
12084: @cindex gdb disassembler
12085:
12086: You can disassemble a @code{code} word with @code{see}
12087: (@pxref{Debugging}). You can disassemble a section of memory with
12088:
12089: doc-discode
12090:
12091: There are two kinds of disassembler for Gforth: The Forth disassembler
12092: (available on some CPUs) and the gdb disassembler (available on
12093: platforms with @command{gdb} and @command{mktemp}). If both are
12094: available, the Forth disassembler is used by default. If you prefer
12095: the gdb disassembler, say
12096:
12097: @example
12098: ' disasm-gdb is discode
12099: @end example
12100:
12101: If neither is available, @code{discode} performs @code{dump}.
12102:
12103: The Forth disassembler generally produces output that can be fed into the
12104: assembler (i.e., same syntax, etc.). It also includes additional
12105: information in comments. In particular, the address of the instruction
12106: is given in a comment before the instruction.
12107:
12108: The gdb disassembler produces output in the same format as the gdb
12109: @code{disassemble} command (@pxref{Machine Code,,Source and machine
12110: code,gdb,Debugging with GDB}), in the default flavour (AT&T syntax for
12111: the 386 and AMD64 architectures).
12112:
12113: @code{See} may display more or less than the actual code of the word,
12114: because the recognition of the end of the code is unreliable. You can
12115: use @code{discode} if it did not display enough. It may display more, if
12116: the code word is not immediately followed by a named word. If you have
12117: something else there, you can follow the word with @code{align latest ,}
12118: to ensure that the end is recognized.
12119:
12120: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
12121: @subsection 386 Assembler
12122:
12123: The 386 assembler included in Gforth was written by Bernd Paysan, it's
12124: available under GPL, and originally part of bigFORTH.
12125:
12126: The 386 disassembler included in Gforth was written by Andrew McKewan
12127: and is in the public domain.
12128:
12129: The disassembler displays code in an Intel-like prefix syntax.
12130:
12131: The assembler uses a postfix syntax with reversed parameters.
12132:
12133: The assembler includes all instruction of the Athlon, i.e. 486 core
12134: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
12135: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
12136: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
12137:
12138: There are several prefixes to switch between different operation sizes,
12139: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
12140: double-word accesses. Addressing modes can be switched with @code{.wa}
12141: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
12142: need a prefix for byte register names (@code{AL} et al).
12143:
12144: For floating point operations, the prefixes are @code{.fs} (IEEE
12145: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
12146: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
12147:
12148: The MMX opcodes don't have size prefixes, they are spelled out like in
12149: the Intel assembler. Instead of move from and to memory, there are
12150: PLDQ/PLDD and PSTQ/PSTD.
12151:
12152: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
12153: ax. Immediate values are indicated by postfixing them with @code{#},
12154: e.g., @code{3 #}. Here are some examples of addressing modes in various
12155: syntaxes:
12156:
12157: @example
12158: Gforth Intel (NASM) AT&T (gas) Name
12159: .w ax ax %ax register (16 bit)
12160: ax eax %eax register (32 bit)
12161: 3 # offset 3 $3 immediate
12162: 1000 #) byte ptr 1000 1000 displacement
12163: bx ) [ebx] (%ebx) base
12164: 100 di d) 100[edi] 100(%edi) base+displacement
12165: 20 ax *4 i#) 20[eax*4] 20(,%eax,4) (index*scale)+displacement
12166: di ax *4 i) [edi][eax*4] (%edi,%eax,4) base+(index*scale)
12167: 4 bx cx di) 4[ebx][ecx] 4(%ebx,%ecx) base+index+displacement
12168: 12 sp ax *2 di) 12[esp][eax*2] 12(%esp,%eax,2) base+(index*scale)+displacement
12169: @end example
12170:
12171: You can use @code{L)} and @code{LI)} instead of @code{D)} and
12172: @code{DI)} to enforce 32-bit displacement fields (useful for
12173: later patching).
12174:
12175: Some example of instructions are:
12176:
12177: @example
12178: ax bx mov \ move ebx,eax
12179: 3 # ax mov \ mov eax,3
12180: 100 di d) ax mov \ mov eax,100[edi]
12181: 4 bx cx di) ax mov \ mov eax,4[ebx][ecx]
12182: .w ax bx mov \ mov bx,ax
12183: @end example
12184:
12185: The following forms are supported for binary instructions:
12186:
12187: @example
12188: <reg> <reg> <inst>
12189: <n> # <reg> <inst>
12190: <mem> <reg> <inst>
12191: <reg> <mem> <inst>
12192: <n> # <mem> <inst>
12193: @end example
12194:
12195: The shift/rotate syntax is:
12196:
12197: @example
12198: <reg/mem> 1 # shl \ shortens to shift without immediate
12199: <reg/mem> 4 # shl
12200: <reg/mem> cl shl
12201: @end example
12202:
12203: Precede string instructions (@code{movs} etc.) with @code{.b} to get
12204: the byte version.
12205:
12206: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
12207: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
12208: pc < >= <= >}. (Note that most of these words shadow some Forth words
12209: when @code{assembler} is in front of @code{forth} in the search path,
12210: e.g., in @code{code} words). Currently the control structure words use
12211: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
12212: to shuffle them (you can also use @code{swap} etc.).
12213:
12214: Here is an example of a @code{code} word (assumes that the stack pointer
12215: is in esi and the TOS is in ebx):
12216:
12217: @example
12218: code my+ ( n1 n2 -- n )
12219: 4 si D) bx add
12220: 4 # si add
12221: Next
12222: end-code
12223: @end example
12224:
12225:
12226: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
12227: @subsection Alpha Assembler
12228:
12229: The Alpha assembler and disassembler were originally written by Bernd
12230: Thallner.
12231:
12232: The register names @code{a0}--@code{a5} are not available to avoid
12233: shadowing hex numbers.
12234:
12235: Immediate forms of arithmetic instructions are distinguished by a
12236: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
12237: does not count as arithmetic instruction).
12238:
12239: You have to specify all operands to an instruction, even those that
12240: other assemblers consider optional, e.g., the destination register for
12241: @code{br,}, or the destination register and hint for @code{jmp,}.
12242:
12243: You can specify conditions for @code{if,} by removing the first @code{b}
12244: and the trailing @code{,} from a branch with a corresponding name; e.g.,
12245:
12246: @example
12247: 11 fgt if, \ if F11>0e
12248: ...
12249: endif,
12250: @end example
12251:
12252: @code{fbgt,} gives @code{fgt}.
12253:
12254: @node MIPS assembler, PowerPC assembler, Alpha Assembler, Assembler and Code Words
12255: @subsection MIPS assembler
12256:
12257: The MIPS assembler was originally written by Christian Pirker.
12258:
12259: Currently the assembler and disassembler only cover the MIPS-I
12260: architecture (R3000), and don't support FP instructions.
12261:
12262: The register names @code{$a0}--@code{$a3} are not available to avoid
12263: shadowing hex numbers.
12264:
12265: Because there is no way to distinguish registers from immediate values,
12266: you have to explicitly use the immediate forms of instructions, i.e.,
12267: @code{addiu,}, not just @code{addu,} (@command{as} does this
12268: implicitly).
12269:
12270: If the architecture manual specifies several formats for the instruction
12271: (e.g., for @code{jalr,}), you usually have to use the one with more
12272: arguments (i.e., two for @code{jalr,}). When in doubt, see
12273: @code{arch/mips/testasm.fs} for an example of correct use.
12274:
12275: Branches and jumps in the MIPS architecture have a delay slot. You have
12276: to fill it yourself (the simplest way is to use @code{nop,}), the
12277: assembler does not do it for you (unlike @command{as}). Even
12278: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
12279: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
12280: and @code{then,} just specify branch targets, they are not affected.
12281:
12282: Note that you must not put branches, jumps, or @code{li,} into the delay
12283: slot: @code{li,} may expand to several instructions, and control flow
12284: instructions may not be put into the branch delay slot in any case.
12285:
12286: For branches the argument specifying the target is a relative address;
12287: You have to add the address of the delay slot to get the absolute
12288: address.
12289:
12290: The MIPS architecture also has load delay slots and restrictions on
12291: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
12292: yourself to satisfy these restrictions, the assembler does not do it for
12293: you.
12294:
12295: You can specify the conditions for @code{if,} etc. by taking a
12296: conditional branch and leaving away the @code{b} at the start and the
12297: @code{,} at the end. E.g.,
12298:
12299: @example
12300: 4 5 eq if,
12301: ... \ do something if $4 equals $5
12302: then,
12303: @end example
12304:
12305:
12306: @node PowerPC assembler, Other assemblers, MIPS assembler, Assembler and Code Words
12307: @subsection PowerPC assembler
12308:
12309: The PowerPC assembler and disassembler were contributed by Michal
12310: Revucky.
12311:
12312: This assembler does not follow the convention of ending mnemonic names
12313: with a ``,'', so some mnemonic names shadow regular Forth words (in
12314: particular: @code{and or xor fabs}); so if you want to use the Forth
12315: words, you have to make them visible first, e.g., with @code{also
12316: forth}.
12317:
12318: Registers are referred to by their number, e.g., @code{9} means the
12319: integer register 9 or the FP register 9 (depending on the
12320: instruction).
12321:
12322: Because there is no way to distinguish registers from immediate values,
12323: you have to explicitly use the immediate forms of instructions, i.e.,
12324: @code{addi,}, not just @code{add,}.
12325:
12326: The assembler and disassembler usually support the most general form
12327: of an instruction, but usually not the shorter forms (especially for
12328: branches).
12329:
12330:
12331:
12332: @node Other assemblers, , PowerPC assembler, Assembler and Code Words
12333: @subsection Other assemblers
12334:
12335: If you want to contribute another assembler/disassembler, please contact
12336: us (@email{anton@@mips.complang.tuwien.ac.at}) to check if we have such
12337: an assembler already. If you are writing them from scratch, please use
12338: a similar syntax style as the one we use (i.e., postfix, commas at the
12339: end of the instruction names, @pxref{Common Assembler}); make the output
12340: of the disassembler be valid input for the assembler, and keep the style
12341: similar to the style we used.
12342:
12343: Hints on implementation: The most important part is to have a good test
12344: suite that contains all instructions. Once you have that, the rest is
12345: easy. For actual coding you can take a look at
12346: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
12347: the assembler and disassembler, avoiding redundancy and some potential
12348: bugs. You can also look at that file (and @pxref{Advanced does> usage
12349: example}) to get ideas how to factor a disassembler.
12350:
12351: Start with the disassembler, because it's easier to reuse data from the
12352: disassembler for the assembler than the other way round.
12353:
12354: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
12355: how simple it can be.
12356:
12357:
12358:
12359:
12360: @c -------------------------------------------------------------
12361: @node Threading Words, Passing Commands to the OS, Assembler and Code Words, Words
12362: @section Threading Words
12363: @cindex threading words
12364:
12365: @cindex code address
12366: These words provide access to code addresses and other threading stuff
12367: in Gforth (and, possibly, other interpretive Forths). It more or less
12368: abstracts away the differences between direct and indirect threading
12369: (and, for direct threading, the machine dependences). However, at
12370: present this wordset is still incomplete. It is also pretty low-level;
12371: some day it will hopefully be made unnecessary by an internals wordset
12372: that abstracts implementation details away completely.
12373:
12374: The terminology used here stems from indirect threaded Forth systems; in
12375: such a system, the XT of a word is represented by the CFA (code field
12376: address) of a word; the CFA points to a cell that contains the code
12377: address. The code address is the address of some machine code that
12378: performs the run-time action of invoking the word (e.g., the
12379: @code{dovar:} routine pushes the address of the body of the word (a
12380: variable) on the stack
12381: ).
12382:
12383: @cindex code address
12384: @cindex code field address
12385: In an indirect threaded Forth, you can get the code address of @i{name}
12386: with @code{' @i{name} @@}; in Gforth you can get it with @code{' @i{name}
12387: >code-address}, independent of the threading method.
12388:
12389: doc-threading-method
12390: doc->code-address
12391: doc-code-address!
12392:
12393: @cindex @code{does>}-handler
12394: @cindex @code{does>}-code
12395: For a word defined with @code{DOES>}, the code address usually points to
12396: a jump instruction (the @dfn{does-handler}) that jumps to the dodoes
12397: routine (in Gforth on some platforms, it can also point to the dodoes
12398: routine itself). What you are typically interested in, though, is
12399: whether a word is a @code{DOES>}-defined word, and what Forth code it
12400: executes; @code{>does-code} tells you that.
12401:
12402: doc->does-code
12403:
12404: To create a @code{DOES>}-defined word with the following basic words,
12405: you have to set up a @code{DOES>}-handler with @code{does-handler!};
12406: @code{/does-handler} aus behind you have to place your executable Forth
12407: code. Finally you have to create a word and modify its behaviour with
12408: @code{does-handler!}.
12409:
12410: doc-does-code!
12411: doc-does-handler!
12412: doc-/does-handler
12413:
12414: The code addresses produced by various defining words are produced by
12415: the following words:
12416:
12417: doc-docol:
12418: doc-docon:
12419: doc-dovar:
12420: doc-douser:
12421: doc-dodefer:
12422: doc-dofield:
12423:
12424: @cindex definer
12425: The following two words generalize @code{>code-address},
12426: @code{>does-code}, @code{code-address!}, and @code{does-code!}:
12427:
12428: doc->definer
12429: doc-definer!
12430:
12431: @c -------------------------------------------------------------
12432: @node Passing Commands to the OS, Keeping track of Time, Threading Words, Words
12433: @section Passing Commands to the Operating System
12434: @cindex operating system - passing commands
12435: @cindex shell commands
12436:
12437: Gforth allows you to pass an arbitrary string to the host operating
12438: system shell (if such a thing exists) for execution.
12439:
12440: doc-sh
12441: doc-system
12442: doc-$?
12443: doc-getenv
12444:
12445: @c -------------------------------------------------------------
12446: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
12447: @section Keeping track of Time
12448: @cindex time-related words
12449:
12450: doc-ms
12451: doc-time&date
12452: doc-utime
12453: doc-cputime
12454:
12455:
12456: @c -------------------------------------------------------------
12457: @node Miscellaneous Words, , Keeping track of Time, Words
12458: @section Miscellaneous Words
12459: @cindex miscellaneous words
12460:
12461: @comment TODO find homes for these
12462:
12463: These section lists the ANS Forth words that are not documented
12464: elsewhere in this manual. Ultimately, they all need proper homes.
12465:
12466: doc-quit
12467:
12468: The following ANS Forth words are not currently supported by Gforth
12469: (@pxref{ANS conformance}):
12470:
12471: @code{EDITOR}
12472: @code{EMIT?}
12473: @code{FORGET}
12474:
12475: @c ******************************************************************
12476: @node Error messages, Tools, Words, Top
12477: @chapter Error messages
12478: @cindex error messages
12479: @cindex backtrace
12480:
12481: A typical Gforth error message looks like this:
12482:
12483: @example
12484: in file included from \evaluated string/:-1
12485: in file included from ./yyy.fs:1
12486: ./xxx.fs:4: Invalid memory address
12487: >>>bar<<<
12488: Backtrace:
12489: $400E664C @@
12490: $400E6664 foo
12491: @end example
12492:
12493: The message identifying the error is @code{Invalid memory address}. The
12494: error happened when text-interpreting line 4 of the file
12495: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
12496: word on the line where the error happened, is pointed out (with
12497: @code{>>>} and @code{<<<}).
12498:
12499: The file containing the error was included in line 1 of @file{./yyy.fs},
12500: and @file{yyy.fs} was included from a non-file (in this case, by giving
12501: @file{yyy.fs} as command-line parameter to Gforth).
12502:
12503: At the end of the error message you find a return stack dump that can be
12504: interpreted as a backtrace (possibly empty). On top you find the top of
12505: the return stack when the @code{throw} happened, and at the bottom you
12506: find the return stack entry just above the return stack of the topmost
12507: text interpreter.
12508:
12509: To the right of most return stack entries you see a guess for the word
12510: that pushed that return stack entry as its return address. This gives a
12511: backtrace. In our case we see that @code{bar} called @code{foo}, and
12512: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
12513: address} exception).
12514:
12515: Note that the backtrace is not perfect: We don't know which return stack
12516: entries are return addresses (so we may get false positives); and in
12517: some cases (e.g., for @code{abort"}) we cannot determine from the return
12518: address the word that pushed the return address, so for some return
12519: addresses you see no names in the return stack dump.
12520:
12521: @cindex @code{catch} and backtraces
12522: The return stack dump represents the return stack at the time when a
12523: specific @code{throw} was executed. In programs that make use of
12524: @code{catch}, it is not necessarily clear which @code{throw} should be
12525: used for the return stack dump (e.g., consider one @code{throw} that
12526: indicates an error, which is caught, and during recovery another error
12527: happens; which @code{throw} should be used for the stack dump?).
12528: Gforth presents the return stack dump for the first @code{throw} after
12529: the last executed (not returned-to) @code{catch} or @code{nothrow};
12530: this works well in the usual case. To get the right backtrace, you
12531: usually want to insert @code{nothrow} or @code{['] false catch drop}
12532: after a @code{catch} if the error is not rethrown.
12533:
12534: @cindex @code{gforth-fast} and backtraces
12535: @cindex @code{gforth-fast}, difference from @code{gforth}
12536: @cindex backtraces with @code{gforth-fast}
12537: @cindex return stack dump with @code{gforth-fast}
12538: @code{Gforth} is able to do a return stack dump for throws generated
12539: from primitives (e.g., invalid memory address, stack empty etc.);
12540: @code{gforth-fast} is only able to do a return stack dump from a
12541: directly called @code{throw} (including @code{abort} etc.). Given an
12542: exception caused by a primitive in @code{gforth-fast}, you will
12543: typically see no return stack dump at all; however, if the exception is
12544: caught by @code{catch} (e.g., for restoring some state), and then
12545: @code{throw}n again, the return stack dump will be for the first such
12546: @code{throw}.
12547:
12548: @c ******************************************************************
12549: @node Tools, ANS conformance, Error messages, Top
12550: @chapter Tools
12551:
12552: @menu
12553: * ANS Report:: Report the words used, sorted by wordset.
12554: * Stack depth changes:: Where does this stack item come from?
12555: @end menu
12556:
12557: See also @ref{Emacs and Gforth}.
12558:
12559: @node ANS Report, Stack depth changes, Tools, Tools
12560: @section @file{ans-report.fs}: Report the words used, sorted by wordset
12561: @cindex @file{ans-report.fs}
12562: @cindex report the words used in your program
12563: @cindex words used in your program
12564:
12565: If you want to label a Forth program as ANS Forth Program, you must
12566: document which wordsets the program uses; for extension wordsets, it is
12567: helpful to list the words the program requires from these wordsets
12568: (because Forth systems are allowed to provide only some words of them).
12569:
12570: The @file{ans-report.fs} tool makes it easy for you to determine which
12571: words from which wordset and which non-ANS words your application
12572: uses. You simply have to include @file{ans-report.fs} before loading the
12573: program you want to check. After loading your program, you can get the
12574: report with @code{print-ans-report}. A typical use is to run this as
12575: batch job like this:
12576: @example
12577: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
12578: @end example
12579:
12580: The output looks like this (for @file{compat/control.fs}):
12581: @example
12582: The program uses the following words
12583: from CORE :
12584: : POSTPONE THEN ; immediate ?dup IF 0=
12585: from BLOCK-EXT :
12586: \
12587: from FILE :
12588: (
12589: @end example
12590:
12591: @subsection Caveats
12592:
12593: Note that @file{ans-report.fs} just checks which words are used, not whether
12594: they are used in an ANS Forth conforming way!
12595:
12596: Some words are defined in several wordsets in the
12597: standard. @file{ans-report.fs} reports them for only one of the
12598: wordsets, and not necessarily the one you expect. It depends on usage
12599: which wordset is the right one to specify. E.g., if you only use the
12600: compilation semantics of @code{S"}, it is a Core word; if you also use
12601: its interpretation semantics, it is a File word.
12602:
12603:
12604: @node Stack depth changes, , ANS Report, Tools
12605: @section Stack depth changes during interpretation
12606: @cindex @file{depth-changes.fs}
12607: @cindex depth changes during interpretation
12608: @cindex stack depth changes during interpretation
12609: @cindex items on the stack after interpretation
12610:
12611: Sometimes you notice that, after loading a file, there are items left
12612: on the stack. The tool @file{depth-changes.fs} helps you find out
12613: quickly where in the file these stack items are coming from.
12614:
12615: The simplest way of using @file{depth-changes.fs} is to include it
12616: before the file(s) you want to check, e.g.:
12617:
12618: @example
12619: gforth depth-changes.fs my-file.fs
12620: @end example
12621:
12622: This will compare the stack depths of the data and FP stack at every
12623: empty line (in interpretation state) against these depths at the last
12624: empty line (in interpretation state). If the depths are not equal,
12625: the position in the file and the stack contents are printed with
12626: @code{~~} (@pxref{Debugging}). This indicates that a stack depth
12627: change has occured in the paragraph of non-empty lines before the
12628: indicated line. It is a good idea to leave an empty line at the end
12629: of the file, so the last paragraph is checked, too.
12630:
12631: Checking only at empty lines usually works well, but sometimes you
12632: have big blocks of non-empty lines (e.g., when building a big table),
12633: and you want to know where in this block the stack depth changed. You
12634: can check all interpreted lines with
12635:
12636: @example
12637: gforth depth-changes.fs -e "' all-lines is depth-changes-filter" my-file.fs
12638: @end example
12639:
12640: This checks the stack depth at every end-of-line. So the depth change
12641: occured in the line reported by the @code{~~} (not in the line
12642: before).
12643:
12644: Note that, while this offers better accuracy in indicating where the
12645: stack depth changes, it will often report many intentional stack depth
12646: changes (e.g., when an interpreted computation stretches across
12647: several lines). You can suppress the checking of some lines by
12648: putting backslashes at the end of these lines (not followed by white
12649: space), and using
12650:
12651: @example
12652: gforth depth-changes.fs -e "' most-lines is depth-changes-filter" my-file.fs
12653: @end example
12654:
12655: @c ******************************************************************
12656: @node ANS conformance, Standard vs Extensions, Tools, Top
12657: @chapter ANS conformance
12658: @cindex ANS conformance of Gforth
12659:
12660: To the best of our knowledge, Gforth is an
12661:
12662: ANS Forth System
12663: @itemize @bullet
12664: @item providing the Core Extensions word set
12665: @item providing the Block word set
12666: @item providing the Block Extensions word set
12667: @item providing the Double-Number word set
12668: @item providing the Double-Number Extensions word set
12669: @item providing the Exception word set
12670: @item providing the Exception Extensions word set
12671: @item providing the Facility word set
12672: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
12673: @item providing the File Access word set
12674: @item providing the File Access Extensions word set
12675: @item providing the Floating-Point word set
12676: @item providing the Floating-Point Extensions word set
12677: @item providing the Locals word set
12678: @item providing the Locals Extensions word set
12679: @item providing the Memory-Allocation word set
12680: @item providing the Memory-Allocation Extensions word set (that one's easy)
12681: @item providing the Programming-Tools word set
12682: @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
12683: @item providing the Search-Order word set
12684: @item providing the Search-Order Extensions word set
12685: @item providing the String word set
12686: @item providing the String Extensions word set (another easy one)
12687: @end itemize
12688:
12689: Gforth has the following environmental restrictions:
12690:
12691: @cindex environmental restrictions
12692: @itemize @bullet
12693: @item
12694: While processing the OS command line, if an exception is not caught,
12695: Gforth exits with a non-zero exit code instyead of performing QUIT.
12696:
12697: @item
12698: When an @code{throw} is performed after a @code{query}, Gforth does not
12699: allways restore the input source specification in effect at the
12700: corresponding catch.
12701:
12702: @end itemize
12703:
12704:
12705: @cindex system documentation
12706: In addition, ANS Forth systems are required to document certain
12707: implementation choices. This chapter tries to meet these
12708: requirements. In many cases it gives a way to ask the system for the
12709: information instead of providing the information directly, in
12710: particular, if the information depends on the processor, the operating
12711: system or the installation options chosen, or if they are likely to
12712: change during the maintenance of Gforth.
12713:
12714: @comment The framework for the rest has been taken from pfe.
12715:
12716: @menu
12717: * The Core Words::
12718: * The optional Block word set::
12719: * The optional Double Number word set::
12720: * The optional Exception word set::
12721: * The optional Facility word set::
12722: * The optional File-Access word set::
12723: * The optional Floating-Point word set::
12724: * The optional Locals word set::
12725: * The optional Memory-Allocation word set::
12726: * The optional Programming-Tools word set::
12727: * The optional Search-Order word set::
12728: @end menu
12729:
12730:
12731: @c =====================================================================
12732: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
12733: @comment node-name, next, previous, up
12734: @section The Core Words
12735: @c =====================================================================
12736: @cindex core words, system documentation
12737: @cindex system documentation, core words
12738:
12739: @menu
12740: * core-idef:: Implementation Defined Options
12741: * core-ambcond:: Ambiguous Conditions
12742: * core-other:: Other System Documentation
12743: @end menu
12744:
12745: @c ---------------------------------------------------------------------
12746: @node core-idef, core-ambcond, The Core Words, The Core Words
12747: @subsection Implementation Defined Options
12748: @c ---------------------------------------------------------------------
12749: @cindex core words, implementation-defined options
12750: @cindex implementation-defined options, core words
12751:
12752:
12753: @table @i
12754: @item (Cell) aligned addresses:
12755: @cindex cell-aligned addresses
12756: @cindex aligned addresses
12757: processor-dependent. Gforth's alignment words perform natural alignment
12758: (e.g., an address aligned for a datum of size 8 is divisible by
12759: 8). Unaligned accesses usually result in a @code{-23 THROW}.
12760:
12761: @item @code{EMIT} and non-graphic characters:
12762: @cindex @code{EMIT} and non-graphic characters
12763: @cindex non-graphic characters and @code{EMIT}
12764: The character is output using the C library function (actually, macro)
12765: @code{putc}.
12766:
12767: @item character editing of @code{ACCEPT} and @code{EXPECT}:
12768: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
12769: @cindex editing in @code{ACCEPT} and @code{EXPECT}
12770: @cindex @code{ACCEPT}, editing
12771: @cindex @code{EXPECT}, editing
12772: This is modeled on the GNU readline library (@pxref{Readline
12773: Interaction, , Command Line Editing, readline, The GNU Readline
12774: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
12775: producing a full word completion every time you type it (instead of
12776: producing the common prefix of all completions). @xref{Command-line editing}.
12777:
12778: @item character set:
12779: @cindex character set
12780: The character set of your computer and display device. Gforth is
12781: 8-bit-clean (but some other component in your system may make trouble).
12782:
12783: @item Character-aligned address requirements:
12784: @cindex character-aligned address requirements
12785: installation-dependent. Currently a character is represented by a C
12786: @code{unsigned char}; in the future we might switch to @code{wchar_t}
12787: (Comments on that requested).
12788:
12789: @item character-set extensions and matching of names:
12790: @cindex character-set extensions and matching of names
12791: @cindex case-sensitivity for name lookup
12792: @cindex name lookup, case-sensitivity
12793: @cindex locale and case-sensitivity
12794: Any character except the ASCII NUL character can be used in a
12795: name. Matching is case-insensitive (except in @code{TABLE}s). The
12796: matching is performed using the C library function @code{strncasecmp}, whose
12797: function is probably influenced by the locale. E.g., the @code{C} locale
12798: does not know about accents and umlauts, so they are matched
12799: case-sensitively in that locale. For portability reasons it is best to
12800: write programs such that they work in the @code{C} locale. Then one can
12801: use libraries written by a Polish programmer (who might use words
12802: containing ISO Latin-2 encoded characters) and by a French programmer
12803: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
12804: funny results for some of the words (which ones, depends on the font you
12805: are using)). Also, the locale you prefer may not be available in other
12806: operating systems. Hopefully, Unicode will solve these problems one day.
12807:
12808: @item conditions under which control characters match a space delimiter:
12809: @cindex space delimiters
12810: @cindex control characters as delimiters
12811: If @code{word} is called with the space character as a delimiter, all
12812: white-space characters (as identified by the C macro @code{isspace()})
12813: are delimiters. @code{Parse}, on the other hand, treats space like other
12814: delimiters. @code{Parse-name}, which is used by the outer
12815: interpreter (aka text interpreter) by default, treats all white-space
12816: characters as delimiters.
12817:
12818: @item format of the control-flow stack:
12819: @cindex control-flow stack, format
12820: The data stack is used as control-flow stack. The size of a control-flow
12821: stack item in cells is given by the constant @code{cs-item-size}. At the
12822: time of this writing, an item consists of a (pointer to a) locals list
12823: (third), an address in the code (second), and a tag for identifying the
12824: item (TOS). The following tags are used: @code{defstart},
12825: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
12826: @code{scopestart}.
12827:
12828: @item conversion of digits > 35
12829: @cindex digits > 35
12830: The characters @code{[\]^_'} are the digits with the decimal value
12831: 36@minus{}41. There is no way to input many of the larger digits.
12832:
12833: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
12834: @cindex @code{EXPECT}, display after end of input
12835: @cindex @code{ACCEPT}, display after end of input
12836: The cursor is moved to the end of the entered string. If the input is
12837: terminated using the @kbd{Return} key, a space is typed.
12838:
12839: @item exception abort sequence of @code{ABORT"}:
12840: @cindex exception abort sequence of @code{ABORT"}
12841: @cindex @code{ABORT"}, exception abort sequence
12842: The error string is stored into the variable @code{"error} and a
12843: @code{-2 throw} is performed.
12844:
12845: @item input line terminator:
12846: @cindex input line terminator
12847: @cindex line terminator on input
12848: @cindex newline character on input
12849: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
12850: lines. One of these characters is typically produced when you type the
12851: @kbd{Enter} or @kbd{Return} key.
12852:
12853: @item maximum size of a counted string:
12854: @cindex maximum size of a counted string
12855: @cindex counted string, maximum size
12856: @code{s" /counted-string" environment? drop .}. Currently 255 characters
12857: on all platforms, but this may change.
12858:
12859: @item maximum size of a parsed string:
12860: @cindex maximum size of a parsed string
12861: @cindex parsed string, maximum size
12862: Given by the constant @code{/line}. Currently 255 characters.
12863:
12864: @item maximum size of a definition name, in characters:
12865: @cindex maximum size of a definition name, in characters
12866: @cindex name, maximum length
12867: MAXU/8
12868:
12869: @item maximum string length for @code{ENVIRONMENT?}, in characters:
12870: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
12871: @cindex @code{ENVIRONMENT?} string length, maximum
12872: MAXU/8
12873:
12874: @item method of selecting the user input device:
12875: @cindex user input device, method of selecting
12876: The user input device is the standard input. There is currently no way to
12877: change it from within Gforth. However, the input can typically be
12878: redirected in the command line that starts Gforth.
12879:
12880: @item method of selecting the user output device:
12881: @cindex user output device, method of selecting
12882: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
12883: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
12884: output when the user output device is a terminal, otherwise the output
12885: is buffered.
12886:
12887: @item methods of dictionary compilation:
12888: What are we expected to document here?
12889:
12890: @item number of bits in one address unit:
12891: @cindex number of bits in one address unit
12892: @cindex address unit, size in bits
12893: @code{s" address-units-bits" environment? drop .}. 8 in all current
12894: platforms.
12895:
12896: @item number representation and arithmetic:
12897: @cindex number representation and arithmetic
12898: Processor-dependent. Binary two's complement on all current platforms.
12899:
12900: @item ranges for integer types:
12901: @cindex ranges for integer types
12902: @cindex integer types, ranges
12903: Installation-dependent. Make environmental queries for @code{MAX-N},
12904: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
12905: unsigned (and positive) types is 0. The lower bound for signed types on
12906: two's complement and one's complement machines machines can be computed
12907: by adding 1 to the upper bound.
12908:
12909: @item read-only data space regions:
12910: @cindex read-only data space regions
12911: @cindex data-space, read-only regions
12912: The whole Forth data space is writable.
12913:
12914: @item size of buffer at @code{WORD}:
12915: @cindex size of buffer at @code{WORD}
12916: @cindex @code{WORD} buffer size
12917: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12918: shared with the pictured numeric output string. If overwriting
12919: @code{PAD} is acceptable, it is as large as the remaining dictionary
12920: space, although only as much can be sensibly used as fits in a counted
12921: string.
12922:
12923: @item size of one cell in address units:
12924: @cindex cell size
12925: @code{1 cells .}.
12926:
12927: @item size of one character in address units:
12928: @cindex char size
12929: @code{1 chars .}. 1 on all current platforms.
12930:
12931: @item size of the keyboard terminal buffer:
12932: @cindex size of the keyboard terminal buffer
12933: @cindex terminal buffer, size
12934: Varies. You can determine the size at a specific time using @code{lp@@
12935: tib - .}. It is shared with the locals stack and TIBs of files that
12936: include the current file. You can change the amount of space for TIBs
12937: and locals stack at Gforth startup with the command line option
12938: @code{-l}.
12939:
12940: @item size of the pictured numeric output buffer:
12941: @cindex size of the pictured numeric output buffer
12942: @cindex pictured numeric output buffer, size
12943: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12944: shared with @code{WORD}.
12945:
12946: @item size of the scratch area returned by @code{PAD}:
12947: @cindex size of the scratch area returned by @code{PAD}
12948: @cindex @code{PAD} size
12949: The remainder of dictionary space. @code{unused pad here - - .}.
12950:
12951: @item system case-sensitivity characteristics:
12952: @cindex case-sensitivity characteristics
12953: Dictionary searches are case-insensitive (except in
12954: @code{TABLE}s). However, as explained above under @i{character-set
12955: extensions}, the matching for non-ASCII characters is determined by the
12956: locale you are using. In the default @code{C} locale all non-ASCII
12957: characters are matched case-sensitively.
12958:
12959: @item system prompt:
12960: @cindex system prompt
12961: @cindex prompt
12962: @code{ ok} in interpret state, @code{ compiled} in compile state.
12963:
12964: @item division rounding:
12965: @cindex division rounding
12966: The ordinary division words @code{/ mod /mod */ */mod} perform floored
12967: division (with the default installation of Gforth). You can check
12968: this with @code{s" floored" environment? drop .}. If you write
12969: programs that need a specific division rounding, best use
12970: @code{fm/mod} or @code{sm/rem} for portability.
12971:
12972: @item values of @code{STATE} when true:
12973: @cindex @code{STATE} values
12974: -1.
12975:
12976: @item values returned after arithmetic overflow:
12977: On two's complement machines, arithmetic is performed modulo
12978: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12979: arithmetic (with appropriate mapping for signed types). Division by
12980: zero typically results in a @code{-55 throw} (Floating-point
12981: unidentified fault) or @code{-10 throw} (divide by zero). Integer
12982: division overflow can result in these throws, or in @code{-11 throw};
12983: in @code{gforth-fast} division overflow and divide by zero may also
12984: result in returning bogus results without producing an exception.
12985:
12986: @item whether the current definition can be found after @t{DOES>}:
12987: @cindex @t{DOES>}, visibility of current definition
12988: No.
12989:
12990: @end table
12991:
12992: @c ---------------------------------------------------------------------
12993: @node core-ambcond, core-other, core-idef, The Core Words
12994: @subsection Ambiguous conditions
12995: @c ---------------------------------------------------------------------
12996: @cindex core words, ambiguous conditions
12997: @cindex ambiguous conditions, core words
12998:
12999: @table @i
13000:
13001: @item a name is neither a word nor a number:
13002: @cindex name not found
13003: @cindex undefined word
13004: @code{-13 throw} (Undefined word).
13005:
13006: @item a definition name exceeds the maximum length allowed:
13007: @cindex word name too long
13008: @code{-19 throw} (Word name too long)
13009:
13010: @item addressing a region not inside the various data spaces of the forth system:
13011: @cindex Invalid memory address
13012: The stacks, code space and header space are accessible. Machine code space is
13013: typically readable. Accessing other addresses gives results dependent on
13014: the operating system. On decent systems: @code{-9 throw} (Invalid memory
13015: address).
13016:
13017: @item argument type incompatible with parameter:
13018: @cindex argument type mismatch
13019: This is usually not caught. Some words perform checks, e.g., the control
13020: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
13021: mismatch).
13022:
13023: @item attempting to obtain the execution token of a word with undefined execution semantics:
13024: @cindex Interpreting a compile-only word, for @code{'} etc.
13025: @cindex execution token of words with undefined execution semantics
13026: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
13027: get an execution token for @code{compile-only-error} (which performs a
13028: @code{-14 throw} when executed).
13029:
13030: @item dividing by zero:
13031: @cindex dividing by zero
13032: @cindex floating point unidentified fault, integer division
13033: On some platforms, this produces a @code{-10 throw} (Division by
13034: zero); on other systems, this typically results in a @code{-55 throw}
13035: (Floating-point unidentified fault).
13036:
13037: @item insufficient data stack or return stack space:
13038: @cindex insufficient data stack or return stack space
13039: @cindex stack overflow
13040: @cindex address alignment exception, stack overflow
13041: @cindex Invalid memory address, stack overflow
13042: Depending on the operating system, the installation, and the invocation
13043: of Gforth, this is either checked by the memory management hardware, or
13044: it is not checked. If it is checked, you typically get a @code{-3 throw}
13045: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
13046: throw} (Invalid memory address) (depending on the platform and how you
13047: achieved the overflow) as soon as the overflow happens. If it is not
13048: checked, overflows typically result in mysterious illegal memory
13049: accesses, producing @code{-9 throw} (Invalid memory address) or
13050: @code{-23 throw} (Address alignment exception); they might also destroy
13051: the internal data structure of @code{ALLOCATE} and friends, resulting in
13052: various errors in these words.
13053:
13054: @item insufficient space for loop control parameters:
13055: @cindex insufficient space for loop control parameters
13056: Like other return stack overflows.
13057:
13058: @item insufficient space in the dictionary:
13059: @cindex insufficient space in the dictionary
13060: @cindex dictionary overflow
13061: If you try to allot (either directly with @code{allot}, or indirectly
13062: with @code{,}, @code{create} etc.) more memory than available in the
13063: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
13064: to access memory beyond the end of the dictionary, the results are
13065: similar to stack overflows.
13066:
13067: @item interpreting a word with undefined interpretation semantics:
13068: @cindex interpreting a word with undefined interpretation semantics
13069: @cindex Interpreting a compile-only word
13070: For some words, we have defined interpretation semantics. For the
13071: others: @code{-14 throw} (Interpreting a compile-only word).
13072:
13073: @item modifying the contents of the input buffer or a string literal:
13074: @cindex modifying the contents of the input buffer or a string literal
13075: These are located in writable memory and can be modified.
13076:
13077: @item overflow of the pictured numeric output string:
13078: @cindex overflow of the pictured numeric output string
13079: @cindex pictured numeric output string, overflow
13080: @code{-17 throw} (Pictured numeric ouput string overflow).
13081:
13082: @item parsed string overflow:
13083: @cindex parsed string overflow
13084: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
13085:
13086: @item producing a result out of range:
13087: @cindex result out of range
13088: On two's complement machines, arithmetic is performed modulo
13089: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
13090: arithmetic (with appropriate mapping for signed types). Division by
13091: zero typically results in a @code{-10 throw} (divide by zero) or
13092: @code{-55 throw} (floating point unidentified fault). Overflow on
13093: division may result in these errors or in @code{-11 throw} (result out
13094: of range). @code{Gforth-fast} may silently produce bogus results on
13095: division overflow or division by zero. @code{Convert} and
13096: @code{>number} currently overflow silently.
13097:
13098: @item reading from an empty data or return stack:
13099: @cindex stack empty
13100: @cindex stack underflow
13101: @cindex return stack underflow
13102: The data stack is checked by the outer (aka text) interpreter after
13103: every word executed. If it has underflowed, a @code{-4 throw} (Stack
13104: underflow) is performed. Apart from that, stacks may be checked or not,
13105: depending on operating system, installation, and invocation. If they are
13106: caught by a check, they typically result in @code{-4 throw} (Stack
13107: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
13108: (Invalid memory address), depending on the platform and which stack
13109: underflows and by how much. Note that even if the system uses checking
13110: (through the MMU), your program may have to underflow by a significant
13111: number of stack items to trigger the reaction (the reason for this is
13112: that the MMU, and therefore the checking, works with a page-size
13113: granularity). If there is no checking, the symptoms resulting from an
13114: underflow are similar to those from an overflow. Unbalanced return
13115: stack errors can result in a variety of symptoms, including @code{-9 throw}
13116: (Invalid memory address) and Illegal Instruction (typically @code{-260
13117: throw}).
13118:
13119: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
13120: @cindex unexpected end of the input buffer
13121: @cindex zero-length string as a name
13122: @cindex Attempt to use zero-length string as a name
13123: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
13124: use zero-length string as a name). Words like @code{'} probably will not
13125: find what they search. Note that it is possible to create zero-length
13126: names with @code{nextname} (should it not?).
13127:
13128: @item @code{>IN} greater than input buffer:
13129: @cindex @code{>IN} greater than input buffer
13130: The next invocation of a parsing word returns a string with length 0.
13131:
13132: @item @code{RECURSE} appears after @code{DOES>}:
13133: @cindex @code{RECURSE} appears after @code{DOES>}
13134: Compiles a recursive call to the defining word, not to the defined word.
13135:
13136: @item argument input source different than current input source for @code{RESTORE-INPUT}:
13137: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
13138: @cindex argument type mismatch, @code{RESTORE-INPUT}
13139: @cindex @code{RESTORE-INPUT}, Argument type mismatch
13140: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
13141: the end of the file was reached), its source-id may be
13142: reused. Therefore, restoring an input source specification referencing a
13143: closed file may lead to unpredictable results instead of a @code{-12
13144: THROW}.
13145:
13146: In the future, Gforth may be able to restore input source specifications
13147: from other than the current input source.
13148:
13149: @item data space containing definitions gets de-allocated:
13150: @cindex data space containing definitions gets de-allocated
13151: Deallocation with @code{allot} is not checked. This typically results in
13152: memory access faults or execution of illegal instructions.
13153:
13154: @item data space read/write with incorrect alignment:
13155: @cindex data space read/write with incorrect alignment
13156: @cindex alignment faults
13157: @cindex address alignment exception
13158: Processor-dependent. Typically results in a @code{-23 throw} (Address
13159: alignment exception). Under Linux-Intel on a 486 or later processor with
13160: alignment turned on, incorrect alignment results in a @code{-9 throw}
13161: (Invalid memory address). There are reportedly some processors with
13162: alignment restrictions that do not report violations.
13163:
13164: @item data space pointer not properly aligned, @code{,}, @code{C,}:
13165: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
13166: Like other alignment errors.
13167:
13168: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
13169: Like other stack underflows.
13170:
13171: @item loop control parameters not available:
13172: @cindex loop control parameters not available
13173: Not checked. The counted loop words simply assume that the top of return
13174: stack items are loop control parameters and behave accordingly.
13175:
13176: @item most recent definition does not have a name (@code{IMMEDIATE}):
13177: @cindex most recent definition does not have a name (@code{IMMEDIATE})
13178: @cindex last word was headerless
13179: @code{abort" last word was headerless"}.
13180:
13181: @item name not defined by @code{VALUE} used by @code{TO}:
13182: @cindex name not defined by @code{VALUE} used by @code{TO}
13183: @cindex @code{TO} on non-@code{VALUE}s
13184: @cindex Invalid name argument, @code{TO}
13185: @code{-32 throw} (Invalid name argument) (unless name is a local or was
13186: defined by @code{CONSTANT}; in the latter case it just changes the constant).
13187:
13188: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
13189: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
13190: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
13191: @code{-13 throw} (Undefined word)
13192:
13193: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
13194: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
13195: Gforth behaves as if they were of the same type. I.e., you can predict
13196: the behaviour by interpreting all parameters as, e.g., signed.
13197:
13198: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
13199: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
13200: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
13201: compilation semantics of @code{TO}.
13202:
13203: @item String longer than a counted string returned by @code{WORD}:
13204: @cindex string longer than a counted string returned by @code{WORD}
13205: @cindex @code{WORD}, string overflow
13206: Not checked. The string will be ok, but the count will, of course,
13207: contain only the least significant bits of the length.
13208:
13209: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
13210: @cindex @code{LSHIFT}, large shift counts
13211: @cindex @code{RSHIFT}, large shift counts
13212: Processor-dependent. Typical behaviours are returning 0 and using only
13213: the low bits of the shift count.
13214:
13215: @item word not defined via @code{CREATE}:
13216: @cindex @code{>BODY} of non-@code{CREATE}d words
13217: @code{>BODY} produces the PFA of the word no matter how it was defined.
13218:
13219: @cindex @code{DOES>} of non-@code{CREATE}d words
13220: @code{DOES>} changes the execution semantics of the last defined word no
13221: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
13222: @code{CREATE , DOES>}.
13223:
13224: @item words improperly used outside @code{<#} and @code{#>}:
13225: Not checked. As usual, you can expect memory faults.
13226:
13227: @end table
13228:
13229:
13230: @c ---------------------------------------------------------------------
13231: @node core-other, , core-ambcond, The Core Words
13232: @subsection Other system documentation
13233: @c ---------------------------------------------------------------------
13234: @cindex other system documentation, core words
13235: @cindex core words, other system documentation
13236:
13237: @table @i
13238: @item nonstandard words using @code{PAD}:
13239: @cindex @code{PAD} use by nonstandard words
13240: None.
13241:
13242: @item operator's terminal facilities available:
13243: @cindex operator's terminal facilities available
13244: After processing the OS's command line, Gforth goes into interactive mode,
13245: and you can give commands to Gforth interactively. The actual facilities
13246: available depend on how you invoke Gforth.
13247:
13248: @item program data space available:
13249: @cindex program data space available
13250: @cindex data space available
13251: @code{UNUSED .} gives the remaining dictionary space. The total
13252: dictionary space can be specified with the @code{-m} switch
13253: (@pxref{Invoking Gforth}) when Gforth starts up.
13254:
13255: @item return stack space available:
13256: @cindex return stack space available
13257: You can compute the total return stack space in cells with
13258: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
13259: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
13260:
13261: @item stack space available:
13262: @cindex stack space available
13263: You can compute the total data stack space in cells with
13264: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
13265: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
13266:
13267: @item system dictionary space required, in address units:
13268: @cindex system dictionary space required, in address units
13269: Type @code{here forthstart - .} after startup. At the time of this
13270: writing, this gives 80080 (bytes) on a 32-bit system.
13271: @end table
13272:
13273:
13274: @c =====================================================================
13275: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
13276: @section The optional Block word set
13277: @c =====================================================================
13278: @cindex system documentation, block words
13279: @cindex block words, system documentation
13280:
13281: @menu
13282: * block-idef:: Implementation Defined Options
13283: * block-ambcond:: Ambiguous Conditions
13284: * block-other:: Other System Documentation
13285: @end menu
13286:
13287:
13288: @c ---------------------------------------------------------------------
13289: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
13290: @subsection Implementation Defined Options
13291: @c ---------------------------------------------------------------------
13292: @cindex implementation-defined options, block words
13293: @cindex block words, implementation-defined options
13294:
13295: @table @i
13296: @item the format for display by @code{LIST}:
13297: @cindex @code{LIST} display format
13298: First the screen number is displayed, then 16 lines of 64 characters,
13299: each line preceded by the line number.
13300:
13301: @item the length of a line affected by @code{\}:
13302: @cindex length of a line affected by @code{\}
13303: @cindex @code{\}, line length in blocks
13304: 64 characters.
13305: @end table
13306:
13307:
13308: @c ---------------------------------------------------------------------
13309: @node block-ambcond, block-other, block-idef, The optional Block word set
13310: @subsection Ambiguous conditions
13311: @c ---------------------------------------------------------------------
13312: @cindex block words, ambiguous conditions
13313: @cindex ambiguous conditions, block words
13314:
13315: @table @i
13316: @item correct block read was not possible:
13317: @cindex block read not possible
13318: Typically results in a @code{throw} of some OS-derived value (between
13319: -512 and -2048). If the blocks file was just not long enough, blanks are
13320: supplied for the missing portion.
13321:
13322: @item I/O exception in block transfer:
13323: @cindex I/O exception in block transfer
13324: @cindex block transfer, I/O exception
13325: Typically results in a @code{throw} of some OS-derived value (between
13326: -512 and -2048).
13327:
13328: @item invalid block number:
13329: @cindex invalid block number
13330: @cindex block number invalid
13331: @code{-35 throw} (Invalid block number)
13332:
13333: @item a program directly alters the contents of @code{BLK}:
13334: @cindex @code{BLK}, altering @code{BLK}
13335: The input stream is switched to that other block, at the same
13336: position. If the storing to @code{BLK} happens when interpreting
13337: non-block input, the system will get quite confused when the block ends.
13338:
13339: @item no current block buffer for @code{UPDATE}:
13340: @cindex @code{UPDATE}, no current block buffer
13341: @code{UPDATE} has no effect.
13342:
13343: @end table
13344:
13345: @c ---------------------------------------------------------------------
13346: @node block-other, , block-ambcond, The optional Block word set
13347: @subsection Other system documentation
13348: @c ---------------------------------------------------------------------
13349: @cindex other system documentation, block words
13350: @cindex block words, other system documentation
13351:
13352: @table @i
13353: @item any restrictions a multiprogramming system places on the use of buffer addresses:
13354: No restrictions (yet).
13355:
13356: @item the number of blocks available for source and data:
13357: depends on your disk space.
13358:
13359: @end table
13360:
13361:
13362: @c =====================================================================
13363: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
13364: @section The optional Double Number word set
13365: @c =====================================================================
13366: @cindex system documentation, double words
13367: @cindex double words, system documentation
13368:
13369: @menu
13370: * double-ambcond:: Ambiguous Conditions
13371: @end menu
13372:
13373:
13374: @c ---------------------------------------------------------------------
13375: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
13376: @subsection Ambiguous conditions
13377: @c ---------------------------------------------------------------------
13378: @cindex double words, ambiguous conditions
13379: @cindex ambiguous conditions, double words
13380:
13381: @table @i
13382: @item @i{d} outside of range of @i{n} in @code{D>S}:
13383: @cindex @code{D>S}, @i{d} out of range of @i{n}
13384: The least significant cell of @i{d} is produced.
13385:
13386: @end table
13387:
13388:
13389: @c =====================================================================
13390: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
13391: @section The optional Exception word set
13392: @c =====================================================================
13393: @cindex system documentation, exception words
13394: @cindex exception words, system documentation
13395:
13396: @menu
13397: * exception-idef:: Implementation Defined Options
13398: @end menu
13399:
13400:
13401: @c ---------------------------------------------------------------------
13402: @node exception-idef, , The optional Exception word set, The optional Exception word set
13403: @subsection Implementation Defined Options
13404: @c ---------------------------------------------------------------------
13405: @cindex implementation-defined options, exception words
13406: @cindex exception words, implementation-defined options
13407:
13408: @table @i
13409: @item @code{THROW}-codes used in the system:
13410: @cindex @code{THROW}-codes used in the system
13411: The codes -256@minus{}-511 are used for reporting signals. The mapping
13412: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
13413: codes -512@minus{}-2047 are used for OS errors (for file and memory
13414: allocation operations). The mapping from OS error numbers to throw codes
13415: is -512@minus{}@code{errno}. One side effect of this mapping is that
13416: undefined OS errors produce a message with a strange number; e.g.,
13417: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
13418: @end table
13419:
13420: @c =====================================================================
13421: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
13422: @section The optional Facility word set
13423: @c =====================================================================
13424: @cindex system documentation, facility words
13425: @cindex facility words, system documentation
13426:
13427: @menu
13428: * facility-idef:: Implementation Defined Options
13429: * facility-ambcond:: Ambiguous Conditions
13430: @end menu
13431:
13432:
13433: @c ---------------------------------------------------------------------
13434: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
13435: @subsection Implementation Defined Options
13436: @c ---------------------------------------------------------------------
13437: @cindex implementation-defined options, facility words
13438: @cindex facility words, implementation-defined options
13439:
13440: @table @i
13441: @item encoding of keyboard events (@code{EKEY}):
13442: @cindex keyboard events, encoding in @code{EKEY}
13443: @cindex @code{EKEY}, encoding of keyboard events
13444: Keys corresponding to ASCII characters are encoded as ASCII characters.
13445: Other keys are encoded with the constants @code{k-left}, @code{k-right},
13446: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
13447: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
13448: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
13449:
13450:
13451: @item duration of a system clock tick:
13452: @cindex duration of a system clock tick
13453: @cindex clock tick duration
13454: System dependent. With respect to @code{MS}, the time is specified in
13455: microseconds. How well the OS and the hardware implement this, is
13456: another question.
13457:
13458: @item repeatability to be expected from the execution of @code{MS}:
13459: @cindex repeatability to be expected from the execution of @code{MS}
13460: @cindex @code{MS}, repeatability to be expected
13461: System dependent. On Unix, a lot depends on load. If the system is
13462: lightly loaded, and the delay is short enough that Gforth does not get
13463: swapped out, the performance should be acceptable. Under MS-DOS and
13464: other single-tasking systems, it should be good.
13465:
13466: @end table
13467:
13468:
13469: @c ---------------------------------------------------------------------
13470: @node facility-ambcond, , facility-idef, The optional Facility word set
13471: @subsection Ambiguous conditions
13472: @c ---------------------------------------------------------------------
13473: @cindex facility words, ambiguous conditions
13474: @cindex ambiguous conditions, facility words
13475:
13476: @table @i
13477: @item @code{AT-XY} can't be performed on user output device:
13478: @cindex @code{AT-XY} can't be performed on user output device
13479: Largely terminal dependent. No range checks are done on the arguments.
13480: No errors are reported. You may see some garbage appearing, you may see
13481: simply nothing happen.
13482:
13483: @end table
13484:
13485:
13486: @c =====================================================================
13487: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
13488: @section The optional File-Access word set
13489: @c =====================================================================
13490: @cindex system documentation, file words
13491: @cindex file words, system documentation
13492:
13493: @menu
13494: * file-idef:: Implementation Defined Options
13495: * file-ambcond:: Ambiguous Conditions
13496: @end menu
13497:
13498: @c ---------------------------------------------------------------------
13499: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
13500: @subsection Implementation Defined Options
13501: @c ---------------------------------------------------------------------
13502: @cindex implementation-defined options, file words
13503: @cindex file words, implementation-defined options
13504:
13505: @table @i
13506: @item file access methods used:
13507: @cindex file access methods used
13508: @code{R/O}, @code{R/W} and @code{BIN} work as you would
13509: expect. @code{W/O} translates into the C file opening mode @code{w} (or
13510: @code{wb}): The file is cleared, if it exists, and created, if it does
13511: not (with both @code{open-file} and @code{create-file}). Under Unix
13512: @code{create-file} creates a file with 666 permissions modified by your
13513: umask.
13514:
13515: @item file exceptions:
13516: @cindex file exceptions
13517: The file words do not raise exceptions (except, perhaps, memory access
13518: faults when you pass illegal addresses or file-ids).
13519:
13520: @item file line terminator:
13521: @cindex file line terminator
13522: System-dependent. Gforth uses C's newline character as line
13523: terminator. What the actual character code(s) of this are is
13524: system-dependent.
13525:
13526: @item file name format:
13527: @cindex file name format
13528: System dependent. Gforth just uses the file name format of your OS.
13529:
13530: @item information returned by @code{FILE-STATUS}:
13531: @cindex @code{FILE-STATUS}, returned information
13532: @code{FILE-STATUS} returns the most powerful file access mode allowed
13533: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
13534: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
13535: along with the returned mode.
13536:
13537: @item input file state after an exception when including source:
13538: @cindex exception when including source
13539: All files that are left via the exception are closed.
13540:
13541: @item @i{ior} values and meaning:
13542: @cindex @i{ior} values and meaning
13543: @cindex @i{wior} values and meaning
13544: The @i{ior}s returned by the file and memory allocation words are
13545: intended as throw codes. They typically are in the range
13546: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
13547: @i{ior}s is -512@minus{}@i{errno}.
13548:
13549: @item maximum depth of file input nesting:
13550: @cindex maximum depth of file input nesting
13551: @cindex file input nesting, maximum depth
13552: limited by the amount of return stack, locals/TIB stack, and the number
13553: of open files available. This should not give you troubles.
13554:
13555: @item maximum size of input line:
13556: @cindex maximum size of input line
13557: @cindex input line size, maximum
13558: @code{/line}. Currently 255.
13559:
13560: @item methods of mapping block ranges to files:
13561: @cindex mapping block ranges to files
13562: @cindex files containing blocks
13563: @cindex blocks in files
13564: By default, blocks are accessed in the file @file{blocks.fb} in the
13565: current working directory. The file can be switched with @code{USE}.
13566:
13567: @item number of string buffers provided by @code{S"}:
13568: @cindex @code{S"}, number of string buffers
13569: 1
13570:
13571: @item size of string buffer used by @code{S"}:
13572: @cindex @code{S"}, size of string buffer
13573: @code{/line}. currently 255.
13574:
13575: @end table
13576:
13577: @c ---------------------------------------------------------------------
13578: @node file-ambcond, , file-idef, The optional File-Access word set
13579: @subsection Ambiguous conditions
13580: @c ---------------------------------------------------------------------
13581: @cindex file words, ambiguous conditions
13582: @cindex ambiguous conditions, file words
13583:
13584: @table @i
13585: @item attempting to position a file outside its boundaries:
13586: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
13587: @code{REPOSITION-FILE} is performed as usual: Afterwards,
13588: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
13589:
13590: @item attempting to read from file positions not yet written:
13591: @cindex reading from file positions not yet written
13592: End-of-file, i.e., zero characters are read and no error is reported.
13593:
13594: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
13595: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
13596: An appropriate exception may be thrown, but a memory fault or other
13597: problem is more probable.
13598:
13599: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
13600: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
13601: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
13602: The @i{ior} produced by the operation, that discovered the problem, is
13603: thrown.
13604:
13605: @item named file cannot be opened (@code{INCLUDED}):
13606: @cindex @code{INCLUDED}, named file cannot be opened
13607: The @i{ior} produced by @code{open-file} is thrown.
13608:
13609: @item requesting an unmapped block number:
13610: @cindex unmapped block numbers
13611: There are no unmapped legal block numbers. On some operating systems,
13612: writing a block with a large number may overflow the file system and
13613: have an error message as consequence.
13614:
13615: @item using @code{source-id} when @code{blk} is non-zero:
13616: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
13617: @code{source-id} performs its function. Typically it will give the id of
13618: the source which loaded the block. (Better ideas?)
13619:
13620: @end table
13621:
13622:
13623: @c =====================================================================
13624: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
13625: @section The optional Floating-Point word set
13626: @c =====================================================================
13627: @cindex system documentation, floating-point words
13628: @cindex floating-point words, system documentation
13629:
13630: @menu
13631: * floating-idef:: Implementation Defined Options
13632: * floating-ambcond:: Ambiguous Conditions
13633: @end menu
13634:
13635:
13636: @c ---------------------------------------------------------------------
13637: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
13638: @subsection Implementation Defined Options
13639: @c ---------------------------------------------------------------------
13640: @cindex implementation-defined options, floating-point words
13641: @cindex floating-point words, implementation-defined options
13642:
13643: @table @i
13644: @item format and range of floating point numbers:
13645: @cindex format and range of floating point numbers
13646: @cindex floating point numbers, format and range
13647: System-dependent; the @code{double} type of C.
13648:
13649: @item results of @code{REPRESENT} when @i{float} is out of range:
13650: @cindex @code{REPRESENT}, results when @i{float} is out of range
13651: System dependent; @code{REPRESENT} is implemented using the C library
13652: function @code{ecvt()} and inherits its behaviour in this respect.
13653:
13654: @item rounding or truncation of floating-point numbers:
13655: @cindex rounding of floating-point numbers
13656: @cindex truncation of floating-point numbers
13657: @cindex floating-point numbers, rounding or truncation
13658: System dependent; the rounding behaviour is inherited from the hosting C
13659: compiler. IEEE-FP-based (i.e., most) systems by default round to
13660: nearest, and break ties by rounding to even (i.e., such that the last
13661: bit of the mantissa is 0).
13662:
13663: @item size of floating-point stack:
13664: @cindex floating-point stack size
13665: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
13666: the floating-point stack (in floats). You can specify this on startup
13667: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
13668:
13669: @item width of floating-point stack:
13670: @cindex floating-point stack width
13671: @code{1 floats}.
13672:
13673: @end table
13674:
13675:
13676: @c ---------------------------------------------------------------------
13677: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
13678: @subsection Ambiguous conditions
13679: @c ---------------------------------------------------------------------
13680: @cindex floating-point words, ambiguous conditions
13681: @cindex ambiguous conditions, floating-point words
13682:
13683: @table @i
13684: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
13685: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
13686: System-dependent. Typically results in a @code{-23 THROW} like other
13687: alignment violations.
13688:
13689: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
13690: @cindex @code{f@@} used with an address that is not float aligned
13691: @cindex @code{f!} used with an address that is not float aligned
13692: System-dependent. Typically results in a @code{-23 THROW} like other
13693: alignment violations.
13694:
13695: @item floating-point result out of range:
13696: @cindex floating-point result out of range
13697: System-dependent. Can result in a @code{-43 throw} (floating point
13698: overflow), @code{-54 throw} (floating point underflow), @code{-41 throw}
13699: (floating point inexact result), @code{-55 THROW} (Floating-point
13700: unidentified fault), or can produce a special value representing, e.g.,
13701: Infinity.
13702:
13703: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
13704: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
13705: System-dependent. Typically results in an alignment fault like other
13706: alignment violations.
13707:
13708: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
13709: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
13710: The floating-point number is converted into decimal nonetheless.
13711:
13712: @item Both arguments are equal to zero (@code{FATAN2}):
13713: @cindex @code{FATAN2}, both arguments are equal to zero
13714: System-dependent. @code{FATAN2} is implemented using the C library
13715: function @code{atan2()}.
13716:
13717: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
13718: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
13719: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
13720: because of small errors and the tan will be a very large (or very small)
13721: but finite number.
13722:
13723: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
13724: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
13725: The result is rounded to the nearest float.
13726:
13727: @item dividing by zero:
13728: @cindex dividing by zero, floating-point
13729: @cindex floating-point dividing by zero
13730: @cindex floating-point unidentified fault, FP divide-by-zero
13731: Platform-dependent; can produce an Infinity, NaN, @code{-42 throw}
13732: (floating point divide by zero) or @code{-55 throw} (Floating-point
13733: unidentified fault).
13734:
13735: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
13736: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
13737: System dependent. On IEEE-FP based systems the number is converted into
13738: an infinity.
13739:
13740: @item @i{float}<1 (@code{FACOSH}):
13741: @cindex @code{FACOSH}, @i{float}<1
13742: @cindex floating-point unidentified fault, @code{FACOSH}
13743: Platform-dependent; on IEEE-FP systems typically produces a NaN.
13744:
13745: @item @i{float}=<-1 (@code{FLNP1}):
13746: @cindex @code{FLNP1}, @i{float}=<-1
13747: @cindex floating-point unidentified fault, @code{FLNP1}
13748: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13749: negative infinity for @i{float}=-1).
13750:
13751: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
13752: @cindex @code{FLN}, @i{float}=<0
13753: @cindex @code{FLOG}, @i{float}=<0
13754: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
13755: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13756: negative infinity for @i{float}=0).
13757:
13758: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
13759: @cindex @code{FASINH}, @i{float}<0
13760: @cindex @code{FSQRT}, @i{float}<0
13761: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
13762: Platform-dependent; for @code{fsqrt} this typically gives a NaN, for
13763: @code{fasinh} some platforms produce a NaN, others a number (bug in the
13764: C library?).
13765:
13766: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
13767: @cindex @code{FACOS}, |@i{float}|>1
13768: @cindex @code{FASIN}, |@i{float}|>1
13769: @cindex @code{FATANH}, |@i{float}|>1
13770: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
13771: Platform-dependent; IEEE-FP systems typically produce a NaN.
13772:
13773: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
13774: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
13775: @cindex floating-point unidentified fault, @code{F>D}
13776: Platform-dependent; typically, some double number is produced and no
13777: error is reported.
13778:
13779: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
13780: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
13781: @code{Precision} characters of the numeric output area are used. If
13782: @code{precision} is too high, these words will smash the data or code
13783: close to @code{here}.
13784: @end table
13785:
13786: @c =====================================================================
13787: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
13788: @section The optional Locals word set
13789: @c =====================================================================
13790: @cindex system documentation, locals words
13791: @cindex locals words, system documentation
13792:
13793: @menu
13794: * locals-idef:: Implementation Defined Options
13795: * locals-ambcond:: Ambiguous Conditions
13796: @end menu
13797:
13798:
13799: @c ---------------------------------------------------------------------
13800: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
13801: @subsection Implementation Defined Options
13802: @c ---------------------------------------------------------------------
13803: @cindex implementation-defined options, locals words
13804: @cindex locals words, implementation-defined options
13805:
13806: @table @i
13807: @item maximum number of locals in a definition:
13808: @cindex maximum number of locals in a definition
13809: @cindex locals, maximum number in a definition
13810: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
13811: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
13812: characters. The number of locals in a definition is bounded by the size
13813: of locals-buffer, which contains the names of the locals.
13814:
13815: @end table
13816:
13817:
13818: @c ---------------------------------------------------------------------
13819: @node locals-ambcond, , locals-idef, The optional Locals word set
13820: @subsection Ambiguous conditions
13821: @c ---------------------------------------------------------------------
13822: @cindex locals words, ambiguous conditions
13823: @cindex ambiguous conditions, locals words
13824:
13825: @table @i
13826: @item executing a named local in interpretation state:
13827: @cindex local in interpretation state
13828: @cindex Interpreting a compile-only word, for a local
13829: Locals have no interpretation semantics. If you try to perform the
13830: interpretation semantics, you will get a @code{-14 throw} somewhere
13831: (Interpreting a compile-only word). If you perform the compilation
13832: semantics, the locals access will be compiled (irrespective of state).
13833:
13834: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
13835: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
13836: @cindex @code{TO} on non-@code{VALUE}s and non-locals
13837: @cindex Invalid name argument, @code{TO}
13838: @code{-32 throw} (Invalid name argument)
13839:
13840: @end table
13841:
13842:
13843: @c =====================================================================
13844: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
13845: @section The optional Memory-Allocation word set
13846: @c =====================================================================
13847: @cindex system documentation, memory-allocation words
13848: @cindex memory-allocation words, system documentation
13849:
13850: @menu
13851: * memory-idef:: Implementation Defined Options
13852: @end menu
13853:
13854:
13855: @c ---------------------------------------------------------------------
13856: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
13857: @subsection Implementation Defined Options
13858: @c ---------------------------------------------------------------------
13859: @cindex implementation-defined options, memory-allocation words
13860: @cindex memory-allocation words, implementation-defined options
13861:
13862: @table @i
13863: @item values and meaning of @i{ior}:
13864: @cindex @i{ior} values and meaning
13865: The @i{ior}s returned by the file and memory allocation words are
13866: intended as throw codes. They typically are in the range
13867: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
13868: @i{ior}s is -512@minus{}@i{errno}.
13869:
13870: @end table
13871:
13872: @c =====================================================================
13873: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
13874: @section The optional Programming-Tools word set
13875: @c =====================================================================
13876: @cindex system documentation, programming-tools words
13877: @cindex programming-tools words, system documentation
13878:
13879: @menu
13880: * programming-idef:: Implementation Defined Options
13881: * programming-ambcond:: Ambiguous Conditions
13882: @end menu
13883:
13884:
13885: @c ---------------------------------------------------------------------
13886: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
13887: @subsection Implementation Defined Options
13888: @c ---------------------------------------------------------------------
13889: @cindex implementation-defined options, programming-tools words
13890: @cindex programming-tools words, implementation-defined options
13891:
13892: @table @i
13893: @item ending sequence for input following @code{;CODE} and @code{CODE}:
13894: @cindex @code{;CODE} ending sequence
13895: @cindex @code{CODE} ending sequence
13896: @code{END-CODE}
13897:
13898: @item manner of processing input following @code{;CODE} and @code{CODE}:
13899: @cindex @code{;CODE}, processing input
13900: @cindex @code{CODE}, processing input
13901: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
13902: the input is processed by the text interpreter, (starting) in interpret
13903: state.
13904:
13905: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
13906: @cindex @code{ASSEMBLER}, search order capability
13907: The ANS Forth search order word set.
13908:
13909: @item source and format of display by @code{SEE}:
13910: @cindex @code{SEE}, source and format of output
13911: The source for @code{see} is the executable code used by the inner
13912: interpreter. The current @code{see} tries to output Forth source code
13913: (and on some platforms, assembly code for primitives) as well as
13914: possible.
13915:
13916: @end table
13917:
13918: @c ---------------------------------------------------------------------
13919: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
13920: @subsection Ambiguous conditions
13921: @c ---------------------------------------------------------------------
13922: @cindex programming-tools words, ambiguous conditions
13923: @cindex ambiguous conditions, programming-tools words
13924:
13925: @table @i
13926:
13927: @item deleting the compilation word list (@code{FORGET}):
13928: @cindex @code{FORGET}, deleting the compilation word list
13929: Not implemented (yet).
13930:
13931: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
13932: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
13933: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
13934: @cindex control-flow stack underflow
13935: This typically results in an @code{abort"} with a descriptive error
13936: message (may change into a @code{-22 throw} (Control structure mismatch)
13937: in the future). You may also get a memory access error. If you are
13938: unlucky, this ambiguous condition is not caught.
13939:
13940: @item @i{name} can't be found (@code{FORGET}):
13941: @cindex @code{FORGET}, @i{name} can't be found
13942: Not implemented (yet).
13943:
13944: @item @i{name} not defined via @code{CREATE}:
13945: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
13946: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
13947: the execution semantics of the last defined word no matter how it was
13948: defined.
13949:
13950: @item @code{POSTPONE} applied to @code{[IF]}:
13951: @cindex @code{POSTPONE} applied to @code{[IF]}
13952: @cindex @code{[IF]} and @code{POSTPONE}
13953: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
13954: equivalent to @code{[IF]}.
13955:
13956: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
13957: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
13958: Continue in the same state of conditional compilation in the next outer
13959: input source. Currently there is no warning to the user about this.
13960:
13961: @item removing a needed definition (@code{FORGET}):
13962: @cindex @code{FORGET}, removing a needed definition
13963: Not implemented (yet).
13964:
13965: @end table
13966:
13967:
13968: @c =====================================================================
13969: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
13970: @section The optional Search-Order word set
13971: @c =====================================================================
13972: @cindex system documentation, search-order words
13973: @cindex search-order words, system documentation
13974:
13975: @menu
13976: * search-idef:: Implementation Defined Options
13977: * search-ambcond:: Ambiguous Conditions
13978: @end menu
13979:
13980:
13981: @c ---------------------------------------------------------------------
13982: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
13983: @subsection Implementation Defined Options
13984: @c ---------------------------------------------------------------------
13985: @cindex implementation-defined options, search-order words
13986: @cindex search-order words, implementation-defined options
13987:
13988: @table @i
13989: @item maximum number of word lists in search order:
13990: @cindex maximum number of word lists in search order
13991: @cindex search order, maximum depth
13992: @code{s" wordlists" environment? drop .}. Currently 16.
13993:
13994: @item minimum search order:
13995: @cindex minimum search order
13996: @cindex search order, minimum
13997: @code{root root}.
13998:
13999: @end table
14000:
14001: @c ---------------------------------------------------------------------
14002: @node search-ambcond, , search-idef, The optional Search-Order word set
14003: @subsection Ambiguous conditions
14004: @c ---------------------------------------------------------------------
14005: @cindex search-order words, ambiguous conditions
14006: @cindex ambiguous conditions, search-order words
14007:
14008: @table @i
14009: @item changing the compilation word list (during compilation):
14010: @cindex changing the compilation word list (during compilation)
14011: @cindex compilation word list, change before definition ends
14012: The word is entered into the word list that was the compilation word list
14013: at the start of the definition. Any changes to the name field (e.g.,
14014: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
14015: are applied to the latest defined word (as reported by @code{latest} or
14016: @code{latestxt}), if possible, irrespective of the compilation word list.
14017:
14018: @item search order empty (@code{previous}):
14019: @cindex @code{previous}, search order empty
14020: @cindex vocstack empty, @code{previous}
14021: @code{abort" Vocstack empty"}.
14022:
14023: @item too many word lists in search order (@code{also}):
14024: @cindex @code{also}, too many word lists in search order
14025: @cindex vocstack full, @code{also}
14026: @code{abort" Vocstack full"}.
14027:
14028: @end table
14029:
14030: @c ***************************************************************
14031: @node Standard vs Extensions, Model, ANS conformance, Top
14032: @chapter Should I use Gforth extensions?
14033: @cindex Gforth extensions
14034:
14035: As you read through the rest of this manual, you will see documentation
14036: for @i{Standard} words, and documentation for some appealing Gforth
14037: @i{extensions}. You might ask yourself the question: @i{``Should I
14038: restrict myself to the standard, or should I use the extensions?''}
14039:
14040: The answer depends on the goals you have for the program you are working
14041: on:
14042:
14043: @itemize @bullet
14044:
14045: @item Is it just for yourself or do you want to share it with others?
14046:
14047: @item
14048: If you want to share it, do the others all use Gforth?
14049:
14050: @item
14051: If it is just for yourself, do you want to restrict yourself to Gforth?
14052:
14053: @end itemize
14054:
14055: If restricting the program to Gforth is ok, then there is no reason not
14056: to use extensions. It is still a good idea to keep to the standard
14057: where it is easy, in case you want to reuse these parts in another
14058: program that you want to be portable.
14059:
14060: If you want to be able to port the program to other Forth systems, there
14061: are the following points to consider:
14062:
14063: @itemize @bullet
14064:
14065: @item
14066: Most Forth systems that are being maintained support the ANS Forth
14067: standard. So if your program complies with the standard, it will be
14068: portable among many systems.
14069:
14070: @item
14071: A number of the Gforth extensions can be implemented in ANS Forth using
14072: public-domain files provided in the @file{compat/} directory. These are
14073: mentioned in the text in passing. There is no reason not to use these
14074: extensions, your program will still be ANS Forth compliant; just include
14075: the appropriate compat files with your program.
14076:
14077: @item
14078: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
14079: analyse your program and determine what non-Standard words it relies
14080: upon. However, it does not check whether you use standard words in a
14081: non-standard way.
14082:
14083: @item
14084: Some techniques are not standardized by ANS Forth, and are hard or
14085: impossible to implement in a standard way, but can be implemented in
14086: most Forth systems easily, and usually in similar ways (e.g., accessing
14087: word headers). Forth has a rich historical precedent for programmers
14088: taking advantage of implementation-dependent features of their tools
14089: (for example, relying on a knowledge of the dictionary
14090: structure). Sometimes these techniques are necessary to extract every
14091: last bit of performance from the hardware, sometimes they are just a
14092: programming shorthand.
14093:
14094: @item
14095: Does using a Gforth extension save more work than the porting this part
14096: to other Forth systems (if any) will cost?
14097:
14098: @item
14099: Is the additional functionality worth the reduction in portability and
14100: the additional porting problems?
14101:
14102: @end itemize
14103:
14104: In order to perform these consideratios, you need to know what's
14105: standard and what's not. This manual generally states if something is
14106: non-standard, but the authoritative source is the
14107: @uref{http://www.taygeta.com/forth/dpans.html,standard document}.
14108: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
14109: into the thought processes of the technical committee.
14110:
14111: Note also that portability between Forth systems is not the only
14112: portability issue; there is also the issue of portability between
14113: different platforms (processor/OS combinations).
14114:
14115: @c ***************************************************************
14116: @node Model, Integrating Gforth, Standard vs Extensions, Top
14117: @chapter Model
14118:
14119: This chapter has yet to be written. It will contain information, on
14120: which internal structures you can rely.
14121:
14122: @c ***************************************************************
14123: @node Integrating Gforth, Emacs and Gforth, Model, Top
14124: @chapter Integrating Gforth into C programs
14125:
14126: This is not yet implemented.
14127:
14128: Several people like to use Forth as scripting language for applications
14129: that are otherwise written in C, C++, or some other language.
14130:
14131: The Forth system ATLAST provides facilities for embedding it into
14132: applications; unfortunately it has several disadvantages: most
14133: importantly, it is not based on ANS Forth, and it is apparently dead
14134: (i.e., not developed further and not supported). The facilities
14135: provided by Gforth in this area are inspired by ATLAST's facilities, so
14136: making the switch should not be hard.
14137:
14138: We also tried to design the interface such that it can easily be
14139: implemented by other Forth systems, so that we may one day arrive at a
14140: standardized interface. Such a standard interface would allow you to
14141: replace the Forth system without having to rewrite C code.
14142:
14143: You embed the Gforth interpreter by linking with the library
14144: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
14145: global symbols in this library that belong to the interface, have the
14146: prefix @code{forth_}. (Global symbols that are used internally have the
14147: prefix @code{gforth_}).
14148:
14149: You can include the declarations of Forth types and the functions and
14150: variables of the interface with @code{#include <forth.h>}.
14151:
14152: Types.
14153:
14154: Variables.
14155:
14156: Data and FP Stack pointer. Area sizes.
14157:
14158: functions.
14159:
14160: forth_init(imagefile)
14161: forth_evaluate(string) exceptions?
14162: forth_goto(address) (or forth_execute(xt)?)
14163: forth_continue() (a corountining mechanism)
14164:
14165: Adding primitives.
14166:
14167: No checking.
14168:
14169: Signals?
14170:
14171: Accessing the Stacks
14172:
14173: @c ******************************************************************
14174: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
14175: @chapter Emacs and Gforth
14176: @cindex Emacs and Gforth
14177:
14178: @cindex @file{gforth.el}
14179: @cindex @file{forth.el}
14180: @cindex Rydqvist, Goran
14181: @cindex Kuehling, David
14182: @cindex comment editing commands
14183: @cindex @code{\}, editing with Emacs
14184: @cindex debug tracer editing commands
14185: @cindex @code{~~}, removal with Emacs
14186: @cindex Forth mode in Emacs
14187:
14188: Gforth comes with @file{gforth.el}, an improved version of
14189: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
14190: improvements are:
14191:
14192: @itemize @bullet
14193: @item
14194: A better handling of indentation.
14195: @item
14196: A custom hilighting engine for Forth-code.
14197: @item
14198: Comment paragraph filling (@kbd{M-q})
14199: @item
14200: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
14201: @item
14202: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
14203: @item
14204: Support of the @code{info-lookup} feature for looking up the
14205: documentation of a word.
14206: @item
14207: Support for reading and writing blocks files.
14208: @end itemize
14209:
14210: To get a basic description of these features, enter Forth mode and
14211: type @kbd{C-h m}.
14212:
14213: @cindex source location of error or debugging output in Emacs
14214: @cindex error output, finding the source location in Emacs
14215: @cindex debugging output, finding the source location in Emacs
14216: In addition, Gforth supports Emacs quite well: The source code locations
14217: given in error messages, debugging output (from @code{~~}) and failed
14218: assertion messages are in the right format for Emacs' compilation mode
14219: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
14220: Manual}) so the source location corresponding to an error or other
14221: message is only a few keystrokes away (@kbd{C-x `} for the next error,
14222: @kbd{C-c C-c} for the error under the cursor).
14223:
14224: @cindex viewing the documentation of a word in Emacs
14225: @cindex context-sensitive help
14226: Moreover, for words documented in this manual, you can look up the
14227: glossary entry quickly by using @kbd{C-h TAB}
14228: (@code{info-lookup-symbol}, @pxref{Documentation, ,Documentation
14229: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
14230: later and does not work for words containing @code{:}.
14231:
14232: @menu
14233: * Installing gforth.el:: Making Emacs aware of Forth.
14234: * Emacs Tags:: Viewing the source of a word in Emacs.
14235: * Hilighting:: Making Forth code look prettier.
14236: * Auto-Indentation:: Customizing auto-indentation.
14237: * Blocks Files:: Reading and writing blocks files.
14238: @end menu
14239:
14240: @c ----------------------------------
14241: @node Installing gforth.el, Emacs Tags, Emacs and Gforth, Emacs and Gforth
14242: @section Installing gforth.el
14243: @cindex @file{.emacs}
14244: @cindex @file{gforth.el}, installation
14245: To make the features from @file{gforth.el} available in Emacs, add
14246: the following lines to your @file{.emacs} file:
14247:
14248: @example
14249: (autoload 'forth-mode "gforth.el")
14250: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode)
14251: auto-mode-alist))
14252: (autoload 'forth-block-mode "gforth.el")
14253: (setq auto-mode-alist (cons '("\\.fb\\'" . forth-block-mode)
14254: auto-mode-alist))
14255: (add-hook 'forth-mode-hook (function (lambda ()
14256: ;; customize variables here:
14257: (setq forth-indent-level 4)
14258: (setq forth-minor-indent-level 2)
14259: (setq forth-hilight-level 3)
14260: ;;; ...
14261: )))
14262: @end example
14263:
14264: @c ----------------------------------
14265: @node Emacs Tags, Hilighting, Installing gforth.el, Emacs and Gforth
14266: @section Emacs Tags
14267: @cindex @file{TAGS} file
14268: @cindex @file{etags.fs}
14269: @cindex viewing the source of a word in Emacs
14270: @cindex @code{require}, placement in files
14271: @cindex @code{include}, placement in files
14272: If you @code{require} @file{etags.fs}, a new @file{TAGS} file will be
14273: produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
14274: contains the definitions of all words defined afterwards. You can then
14275: find the source for a word using @kbd{M-.}. Note that Emacs can use
14276: several tags files at the same time (e.g., one for the Gforth sources
14277: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
14278: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
14279: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
14280: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
14281: with @file{etags.fs}, you should avoid putting definitions both before
14282: and after @code{require} etc., otherwise you will see the same file
14283: visited several times by commands like @code{tags-search}.
14284:
14285: @c ----------------------------------
14286: @node Hilighting, Auto-Indentation, Emacs Tags, Emacs and Gforth
14287: @section Hilighting
14288: @cindex hilighting Forth code in Emacs
14289: @cindex highlighting Forth code in Emacs
14290: @file{gforth.el} comes with a custom source hilighting engine. When
14291: you open a file in @code{forth-mode}, it will be completely parsed,
14292: assigning faces to keywords, comments, strings etc. While you edit
14293: the file, modified regions get parsed and updated on-the-fly.
14294:
14295: Use the variable `forth-hilight-level' to change the level of
14296: decoration from 0 (no hilighting at all) to 3 (the default). Even if
14297: you set the hilighting level to 0, the parser will still work in the
14298: background, collecting information about whether regions of text are
14299: ``compiled'' or ``interpreted''. Those information are required for
14300: auto-indentation to work properly. Set `forth-disable-parser' to
14301: non-nil if your computer is too slow to handle parsing. This will
14302: have an impact on the smartness of the auto-indentation engine,
14303: though.
14304:
14305: Sometimes Forth sources define new features that should be hilighted,
14306: new control structures, defining-words etc. You can use the variable
14307: `forth-custom-words' to make @code{forth-mode} hilight additional
14308: words and constructs. See the docstring of `forth-words' for details
14309: (in Emacs, type @kbd{C-h v forth-words}).
14310:
14311: `forth-custom-words' is meant to be customized in your
14312: @file{.emacs} file. To customize hilighing in a file-specific manner,
14313: set `forth-local-words' in a local-variables section at the end of
14314: your source file (@pxref{Local Variables in Files,, Variables, emacs, Emacs Manual}).
14315:
14316: Example:
14317: @example
14318: 0 [IF]
14319: Local Variables:
14320: forth-local-words:
14321: ((("t:") definition-starter (font-lock-keyword-face . 1)
14322: "[ \t\n]" t name (font-lock-function-name-face . 3))
14323: ((";t") definition-ender (font-lock-keyword-face . 1)))
14324: End:
14325: [THEN]
14326: @end example
14327:
14328: @c ----------------------------------
14329: @node Auto-Indentation, Blocks Files, Hilighting, Emacs and Gforth
14330: @section Auto-Indentation
14331: @cindex auto-indentation of Forth code in Emacs
14332: @cindex indentation of Forth code in Emacs
14333: @code{forth-mode} automatically tries to indent lines in a smart way,
14334: whenever you type @key{TAB} or break a line with @kbd{C-m}.
14335:
14336: Simple customization can be achieved by setting
14337: `forth-indent-level' and `forth-minor-indent-level' in your
14338: @file{.emacs} file. For historical reasons @file{gforth.el} indents
14339: per default by multiples of 4 columns. To use the more traditional
14340: 3-column indentation, add the following lines to your @file{.emacs}:
14341:
14342: @example
14343: (add-hook 'forth-mode-hook (function (lambda ()
14344: ;; customize variables here:
14345: (setq forth-indent-level 3)
14346: (setq forth-minor-indent-level 1)
14347: )))
14348: @end example
14349:
14350: If you want indentation to recognize non-default words, customize it
14351: by setting `forth-custom-indent-words' in your @file{.emacs}. See the
14352: docstring of `forth-indent-words' for details (in Emacs, type @kbd{C-h
14353: v forth-indent-words}).
14354:
14355: To customize indentation in a file-specific manner, set
14356: `forth-local-indent-words' in a local-variables section at the end of
14357: your source file (@pxref{Local Variables in Files, Variables,,emacs,
14358: Emacs Manual}).
14359:
14360: Example:
14361: @example
14362: 0 [IF]
14363: Local Variables:
14364: forth-local-indent-words:
14365: ((("t:") (0 . 2) (0 . 2))
14366: ((";t") (-2 . 0) (0 . -2)))
14367: End:
14368: [THEN]
14369: @end example
14370:
14371: @c ----------------------------------
14372: @node Blocks Files, , Auto-Indentation, Emacs and Gforth
14373: @section Blocks Files
14374: @cindex blocks files, use with Emacs
14375: @code{forth-mode} Autodetects blocks files by checking whether the
14376: length of the first line exceeds 1023 characters. It then tries to
14377: convert the file into normal text format. When you save the file, it
14378: will be written to disk as normal stream-source file.
14379:
14380: If you want to write blocks files, use @code{forth-blocks-mode}. It
14381: inherits all the features from @code{forth-mode}, plus some additions:
14382:
14383: @itemize @bullet
14384: @item
14385: Files are written to disk in blocks file format.
14386: @item
14387: Screen numbers are displayed in the mode line (enumerated beginning
14388: with the value of `forth-block-base')
14389: @item
14390: Warnings are displayed when lines exceed 64 characters.
14391: @item
14392: The beginning of the currently edited block is marked with an
14393: overlay-arrow.
14394: @end itemize
14395:
14396: There are some restrictions you should be aware of. When you open a
14397: blocks file that contains tabulator or newline characters, these
14398: characters will be translated into spaces when the file is written
14399: back to disk. If tabs or newlines are encountered during blocks file
14400: reading, an error is output to the echo area. So have a look at the
14401: `*Messages*' buffer, when Emacs' bell rings during reading.
14402:
14403: Please consult the docstring of @code{forth-blocks-mode} for more
14404: information by typing @kbd{C-h v forth-blocks-mode}).
14405:
14406: @c ******************************************************************
14407: @node Image Files, Engine, Emacs and Gforth, Top
14408: @chapter Image Files
14409: @cindex image file
14410: @cindex @file{.fi} files
14411: @cindex precompiled Forth code
14412: @cindex dictionary in persistent form
14413: @cindex persistent form of dictionary
14414:
14415: An image file is a file containing an image of the Forth dictionary,
14416: i.e., compiled Forth code and data residing in the dictionary. By
14417: convention, we use the extension @code{.fi} for image files.
14418:
14419: @menu
14420: * Image Licensing Issues:: Distribution terms for images.
14421: * Image File Background:: Why have image files?
14422: * Non-Relocatable Image Files:: don't always work.
14423: * Data-Relocatable Image Files:: are better.
14424: * Fully Relocatable Image Files:: better yet.
14425: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
14426: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
14427: * Modifying the Startup Sequence:: and turnkey applications.
14428: @end menu
14429:
14430: @node Image Licensing Issues, Image File Background, Image Files, Image Files
14431: @section Image Licensing Issues
14432: @cindex license for images
14433: @cindex image license
14434:
14435: An image created with @code{gforthmi} (@pxref{gforthmi}) or
14436: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
14437: original image; i.e., according to copyright law it is a derived work of
14438: the original image.
14439:
14440: Since Gforth is distributed under the GNU GPL, the newly created image
14441: falls under the GNU GPL, too. In particular, this means that if you
14442: distribute the image, you have to make all of the sources for the image
14443: available, including those you wrote. For details see @ref{Copying, ,
14444: GNU General Public License (Section 3)}.
14445:
14446: If you create an image with @code{cross} (@pxref{cross.fs}), the image
14447: contains only code compiled from the sources you gave it; if none of
14448: these sources is under the GPL, the terms discussed above do not apply
14449: to the image. However, if your image needs an engine (a gforth binary)
14450: that is under the GPL, you should make sure that you distribute both in
14451: a way that is at most a @emph{mere aggregation}, if you don't want the
14452: terms of the GPL to apply to the image.
14453:
14454: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
14455: @section Image File Background
14456: @cindex image file background
14457:
14458: Gforth consists not only of primitives (in the engine), but also of
14459: definitions written in Forth. Since the Forth compiler itself belongs to
14460: those definitions, it is not possible to start the system with the
14461: engine and the Forth source alone. Therefore we provide the Forth
14462: code as an image file in nearly executable form. When Gforth starts up,
14463: a C routine loads the image file into memory, optionally relocates the
14464: addresses, then sets up the memory (stacks etc.) according to
14465: information in the image file, and (finally) starts executing Forth
14466: code.
14467:
14468: The image file variants represent different compromises between the
14469: goals of making it easy to generate image files and making them
14470: portable.
14471:
14472: @cindex relocation at run-time
14473: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
14474: run-time. This avoids many of the complications discussed below (image
14475: files are data relocatable without further ado), but costs performance
14476: (one addition per memory access).
14477:
14478: @cindex relocation at load-time
14479: By contrast, the Gforth loader performs relocation at image load time. The
14480: loader also has to replace tokens that represent primitive calls with the
14481: appropriate code-field addresses (or code addresses in the case of
14482: direct threading).
14483:
14484: There are three kinds of image files, with different degrees of
14485: relocatability: non-relocatable, data-relocatable, and fully relocatable
14486: image files.
14487:
14488: @cindex image file loader
14489: @cindex relocating loader
14490: @cindex loader for image files
14491: These image file variants have several restrictions in common; they are
14492: caused by the design of the image file loader:
14493:
14494: @itemize @bullet
14495: @item
14496: There is only one segment; in particular, this means, that an image file
14497: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
14498: them). The contents of the stacks are not represented, either.
14499:
14500: @item
14501: The only kinds of relocation supported are: adding the same offset to
14502: all cells that represent data addresses; and replacing special tokens
14503: with code addresses or with pieces of machine code.
14504:
14505: If any complex computations involving addresses are performed, the
14506: results cannot be represented in the image file. Several applications that
14507: use such computations come to mind:
14508: @itemize @minus
14509: @item
14510: Hashing addresses (or data structures which contain addresses) for table
14511: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
14512: purpose, you will have no problem, because the hash tables are
14513: recomputed automatically when the system is started. If you use your own
14514: hash tables, you will have to do something similar.
14515:
14516: @item
14517: There's a cute implementation of doubly-linked lists that uses
14518: @code{XOR}ed addresses. You could represent such lists as singly-linked
14519: in the image file, and restore the doubly-linked representation on
14520: startup.@footnote{In my opinion, though, you should think thrice before
14521: using a doubly-linked list (whatever implementation).}
14522:
14523: @item
14524: The code addresses of run-time routines like @code{docol:} cannot be
14525: represented in the image file (because their tokens would be replaced by
14526: machine code in direct threaded implementations). As a workaround,
14527: compute these addresses at run-time with @code{>code-address} from the
14528: executions tokens of appropriate words (see the definitions of
14529: @code{docol:} and friends in @file{kernel/getdoers.fs}).
14530:
14531: @item
14532: On many architectures addresses are represented in machine code in some
14533: shifted or mangled form. You cannot put @code{CODE} words that contain
14534: absolute addresses in this form in a relocatable image file. Workarounds
14535: are representing the address in some relative form (e.g., relative to
14536: the CFA, which is present in some register), or loading the address from
14537: a place where it is stored in a non-mangled form.
14538: @end itemize
14539: @end itemize
14540:
14541: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
14542: @section Non-Relocatable Image Files
14543: @cindex non-relocatable image files
14544: @cindex image file, non-relocatable
14545:
14546: These files are simple memory dumps of the dictionary. They are specific
14547: to the executable (i.e., @file{gforth} file) they were created
14548: with. What's worse, they are specific to the place on which the
14549: dictionary resided when the image was created. Now, there is no
14550: guarantee that the dictionary will reside at the same place the next
14551: time you start Gforth, so there's no guarantee that a non-relocatable
14552: image will work the next time (Gforth will complain instead of crashing,
14553: though).
14554:
14555: You can create a non-relocatable image file with
14556:
14557:
14558: doc-savesystem
14559:
14560:
14561: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
14562: @section Data-Relocatable Image Files
14563: @cindex data-relocatable image files
14564: @cindex image file, data-relocatable
14565:
14566: These files contain relocatable data addresses, but fixed code addresses
14567: (instead of tokens). They are specific to the executable (i.e.,
14568: @file{gforth} file) they were created with. For direct threading on some
14569: architectures (e.g., the i386), data-relocatable images do not work. You
14570: get a data-relocatable image, if you use @file{gforthmi} with a
14571: Gforth binary that is not doubly indirect threaded (@pxref{Fully
14572: Relocatable Image Files}).
14573:
14574: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
14575: @section Fully Relocatable Image Files
14576: @cindex fully relocatable image files
14577: @cindex image file, fully relocatable
14578:
14579: @cindex @file{kern*.fi}, relocatability
14580: @cindex @file{gforth.fi}, relocatability
14581: These image files have relocatable data addresses, and tokens for code
14582: addresses. They can be used with different binaries (e.g., with and
14583: without debugging) on the same machine, and even across machines with
14584: the same data formats (byte order, cell size, floating point
14585: format). However, they are usually specific to the version of Gforth
14586: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
14587: are fully relocatable.
14588:
14589: There are two ways to create a fully relocatable image file:
14590:
14591: @menu
14592: * gforthmi:: The normal way
14593: * cross.fs:: The hard way
14594: @end menu
14595:
14596: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
14597: @subsection @file{gforthmi}
14598: @cindex @file{comp-i.fs}
14599: @cindex @file{gforthmi}
14600:
14601: You will usually use @file{gforthmi}. If you want to create an
14602: image @i{file} that contains everything you would load by invoking
14603: Gforth with @code{gforth @i{options}}, you simply say:
14604: @example
14605: gforthmi @i{file} @i{options}
14606: @end example
14607:
14608: E.g., if you want to create an image @file{asm.fi} that has the file
14609: @file{asm.fs} loaded in addition to the usual stuff, you could do it
14610: like this:
14611:
14612: @example
14613: gforthmi asm.fi asm.fs
14614: @end example
14615:
14616: @file{gforthmi} is implemented as a sh script and works like this: It
14617: produces two non-relocatable images for different addresses and then
14618: compares them. Its output reflects this: first you see the output (if
14619: any) of the two Gforth invocations that produce the non-relocatable image
14620: files, then you see the output of the comparing program: It displays the
14621: offset used for data addresses and the offset used for code addresses;
14622: moreover, for each cell that cannot be represented correctly in the
14623: image files, it displays a line like this:
14624:
14625: @example
14626: 78DC BFFFFA50 BFFFFA40
14627: @end example
14628:
14629: This means that at offset $78dc from @code{forthstart}, one input image
14630: contains $bffffa50, and the other contains $bffffa40. Since these cells
14631: cannot be represented correctly in the output image, you should examine
14632: these places in the dictionary and verify that these cells are dead
14633: (i.e., not read before they are written).
14634:
14635: @cindex --application, @code{gforthmi} option
14636: If you insert the option @code{--application} in front of the image file
14637: name, you will get an image that uses the @code{--appl-image} option
14638: instead of the @code{--image-file} option (@pxref{Invoking
14639: Gforth}). When you execute such an image on Unix (by typing the image
14640: name as command), the Gforth engine will pass all options to the image
14641: instead of trying to interpret them as engine options.
14642:
14643: If you type @file{gforthmi} with no arguments, it prints some usage
14644: instructions.
14645:
14646: @cindex @code{savesystem} during @file{gforthmi}
14647: @cindex @code{bye} during @file{gforthmi}
14648: @cindex doubly indirect threaded code
14649: @cindex environment variables
14650: @cindex @code{GFORTHD} -- environment variable
14651: @cindex @code{GFORTH} -- environment variable
14652: @cindex @code{gforth-ditc}
14653: There are a few wrinkles: After processing the passed @i{options}, the
14654: words @code{savesystem} and @code{bye} must be visible. A special doubly
14655: indirect threaded version of the @file{gforth} executable is used for
14656: creating the non-relocatable images; you can pass the exact filename of
14657: this executable through the environment variable @code{GFORTHD}
14658: (default: @file{gforth-ditc}); if you pass a version that is not doubly
14659: indirect threaded, you will not get a fully relocatable image, but a
14660: data-relocatable image (because there is no code address offset). The
14661: normal @file{gforth} executable is used for creating the relocatable
14662: image; you can pass the exact filename of this executable through the
14663: environment variable @code{GFORTH}.
14664:
14665: @node cross.fs, , gforthmi, Fully Relocatable Image Files
14666: @subsection @file{cross.fs}
14667: @cindex @file{cross.fs}
14668: @cindex cross-compiler
14669: @cindex metacompiler
14670: @cindex target compiler
14671:
14672: You can also use @code{cross}, a batch compiler that accepts a Forth-like
14673: programming language (@pxref{Cross Compiler}).
14674:
14675: @code{cross} allows you to create image files for machines with
14676: different data sizes and data formats than the one used for generating
14677: the image file. You can also use it to create an application image that
14678: does not contain a Forth compiler. These features are bought with
14679: restrictions and inconveniences in programming. E.g., addresses have to
14680: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
14681: order to make the code relocatable.
14682:
14683:
14684: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
14685: @section Stack and Dictionary Sizes
14686: @cindex image file, stack and dictionary sizes
14687: @cindex dictionary size default
14688: @cindex stack size default
14689:
14690: If you invoke Gforth with a command line flag for the size
14691: (@pxref{Invoking Gforth}), the size you specify is stored in the
14692: dictionary. If you save the dictionary with @code{savesystem} or create
14693: an image with @file{gforthmi}, this size will become the default
14694: for the resulting image file. E.g., the following will create a
14695: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
14696:
14697: @example
14698: gforthmi gforth.fi -m 1M
14699: @end example
14700:
14701: In other words, if you want to set the default size for the dictionary
14702: and the stacks of an image, just invoke @file{gforthmi} with the
14703: appropriate options when creating the image.
14704:
14705: @cindex stack size, cache-friendly
14706: Note: For cache-friendly behaviour (i.e., good performance), you should
14707: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
14708: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
14709: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
14710:
14711: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
14712: @section Running Image Files
14713: @cindex running image files
14714: @cindex invoking image files
14715: @cindex image file invocation
14716:
14717: @cindex -i, invoke image file
14718: @cindex --image file, invoke image file
14719: You can invoke Gforth with an image file @i{image} instead of the
14720: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
14721: @example
14722: gforth -i @i{image}
14723: @end example
14724:
14725: @cindex executable image file
14726: @cindex image file, executable
14727: If your operating system supports starting scripts with a line of the
14728: form @code{#! ...}, you just have to type the image file name to start
14729: Gforth with this image file (note that the file extension @code{.fi} is
14730: just a convention). I.e., to run Gforth with the image file @i{image},
14731: you can just type @i{image} instead of @code{gforth -i @i{image}}.
14732: This works because every @code{.fi} file starts with a line of this
14733: format:
14734:
14735: @example
14736: #! /usr/local/bin/gforth-0.4.0 -i
14737: @end example
14738:
14739: The file and pathname for the Gforth engine specified on this line is
14740: the specific Gforth executable that it was built against; i.e. the value
14741: of the environment variable @code{GFORTH} at the time that
14742: @file{gforthmi} was executed.
14743:
14744: You can make use of the same shell capability to make a Forth source
14745: file into an executable. For example, if you place this text in a file:
14746:
14747: @example
14748: #! /usr/local/bin/gforth
14749:
14750: ." Hello, world" CR
14751: bye
14752: @end example
14753:
14754: @noindent
14755: and then make the file executable (chmod +x in Unix), you can run it
14756: directly from the command line. The sequence @code{#!} is used in two
14757: ways; firstly, it is recognised as a ``magic sequence'' by the operating
14758: system@footnote{The Unix kernel actually recognises two types of files:
14759: executable files and files of data, where the data is processed by an
14760: interpreter that is specified on the ``interpreter line'' -- the first
14761: line of the file, starting with the sequence #!. There may be a small
14762: limit (e.g., 32) on the number of characters that may be specified on
14763: the interpreter line.} secondly it is treated as a comment character by
14764: Gforth. Because of the second usage, a space is required between
14765: @code{#!} and the path to the executable (moreover, some Unixes
14766: require the sequence @code{#! /}).
14767:
14768: The disadvantage of this latter technique, compared with using
14769: @file{gforthmi}, is that it is slightly slower; the Forth source code is
14770: compiled on-the-fly, each time the program is invoked.
14771:
14772: doc-#!
14773:
14774:
14775: @node Modifying the Startup Sequence, , Running Image Files, Image Files
14776: @section Modifying the Startup Sequence
14777: @cindex startup sequence for image file
14778: @cindex image file initialization sequence
14779: @cindex initialization sequence of image file
14780:
14781: You can add your own initialization to the startup sequence of an image
14782: through the deferred word @code{'cold}. @code{'cold} is invoked just
14783: before the image-specific command line processing (i.e., loading files
14784: and evaluating (@code{-e}) strings) starts.
14785:
14786: A sequence for adding your initialization usually looks like this:
14787:
14788: @example
14789: :noname
14790: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
14791: ... \ your stuff
14792: ; IS 'cold
14793: @end example
14794:
14795: After @code{'cold}, Gforth processes the image options
14796: (@pxref{Invoking Gforth}), and then it performs @code{bootmessage},
14797: another deferred word. This normally prints Gforth's startup message
14798: and does nothing else.
14799:
14800: @cindex turnkey image files
14801: @cindex image file, turnkey applications
14802: So, if you want to make a turnkey image (i.e., an image for an
14803: application instead of an extended Forth system), you can do this in
14804: two ways:
14805:
14806: @itemize @bullet
14807:
14808: @item
14809: If you want to do your interpretation of the OS command-line
14810: arguments, hook into @code{'cold}. In that case you probably also
14811: want to build the image with @code{gforthmi --application}
14812: (@pxref{gforthmi}) to keep the engine from processing OS command line
14813: options. You can then do your own command-line processing with
14814: @code{next-arg}
14815:
14816: @item
14817: If you want to have the normal Gforth processing of OS command-line
14818: arguments, hook into @code{bootmessage}.
14819:
14820: @end itemize
14821:
14822: In either case, you probably do not want the word that you execute in
14823: these hooks to exit normally, but use @code{bye} or @code{throw}.
14824: Otherwise the Gforth startup process would continue and eventually
14825: present the Forth command line to the user.
14826:
14827: doc-'cold
14828: doc-bootmessage
14829:
14830: @c ******************************************************************
14831: @node Engine, Cross Compiler, Image Files, Top
14832: @chapter Engine
14833: @cindex engine
14834: @cindex virtual machine
14835:
14836: Reading this chapter is not necessary for programming with Gforth. It
14837: may be helpful for finding your way in the Gforth sources.
14838:
14839: The ideas in this section have also been published in the following
14840: papers: Bernd Paysan, @cite{ANS fig/GNU/??? Forth} (in German),
14841: Forth-Tagung '93; M. Anton Ertl,
14842: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
14843: Portable Forth Engine}}, EuroForth '93; M. Anton Ertl,
14844: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl02.ps.gz,
14845: Threaded code variations and optimizations (extended version)}},
14846: Forth-Tagung '02.
14847:
14848: @menu
14849: * Portability::
14850: * Threading::
14851: * Primitives::
14852: * Performance::
14853: @end menu
14854:
14855: @node Portability, Threading, Engine, Engine
14856: @section Portability
14857: @cindex engine portability
14858:
14859: An important goal of the Gforth Project is availability across a wide
14860: range of personal machines. fig-Forth, and, to a lesser extent, F83,
14861: achieved this goal by manually coding the engine in assembly language
14862: for several then-popular processors. This approach is very
14863: labor-intensive and the results are short-lived due to progress in
14864: computer architecture.
14865:
14866: @cindex C, using C for the engine
14867: Others have avoided this problem by coding in C, e.g., Mitch Bradley
14868: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
14869: particularly popular for UNIX-based Forths due to the large variety of
14870: architectures of UNIX machines. Unfortunately an implementation in C
14871: does not mix well with the goals of efficiency and with using
14872: traditional techniques: Indirect or direct threading cannot be expressed
14873: in C, and switch threading, the fastest technique available in C, is
14874: significantly slower. Another problem with C is that it is very
14875: cumbersome to express double integer arithmetic.
14876:
14877: @cindex GNU C for the engine
14878: @cindex long long
14879: Fortunately, there is a portable language that does not have these
14880: limitations: GNU C, the version of C processed by the GNU C compiler
14881: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
14882: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
14883: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
14884: threading possible, its @code{long long} type (@pxref{Long Long, ,
14885: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
14886: double numbers on many systems. GNU C is freely available on all
14887: important (and many unimportant) UNIX machines, VMS, 80386s running
14888: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
14889: on all these machines.
14890:
14891: Writing in a portable language has the reputation of producing code that
14892: is slower than assembly. For our Forth engine we repeatedly looked at
14893: the code produced by the compiler and eliminated most compiler-induced
14894: inefficiencies by appropriate changes in the source code.
14895:
14896: @cindex explicit register declarations
14897: @cindex --enable-force-reg, configuration flag
14898: @cindex -DFORCE_REG
14899: However, register allocation cannot be portably influenced by the
14900: programmer, leading to some inefficiencies on register-starved
14901: machines. We use explicit register declarations (@pxref{Explicit Reg
14902: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
14903: improve the speed on some machines. They are turned on by using the
14904: configuration flag @code{--enable-force-reg} (@code{gcc} switch
14905: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
14906: machine, but also on the compiler version: On some machines some
14907: compiler versions produce incorrect code when certain explicit register
14908: declarations are used. So by default @code{-DFORCE_REG} is not used.
14909:
14910: @node Threading, Primitives, Portability, Engine
14911: @section Threading
14912: @cindex inner interpreter implementation
14913: @cindex threaded code implementation
14914:
14915: @cindex labels as values
14916: GNU C's labels as values extension (available since @code{gcc-2.0},
14917: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
14918: makes it possible to take the address of @i{label} by writing
14919: @code{&&@i{label}}. This address can then be used in a statement like
14920: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
14921: @code{goto x}.
14922:
14923: @cindex @code{NEXT}, indirect threaded
14924: @cindex indirect threaded inner interpreter
14925: @cindex inner interpreter, indirect threaded
14926: With this feature an indirect threaded @code{NEXT} looks like:
14927: @example
14928: cfa = *ip++;
14929: ca = *cfa;
14930: goto *ca;
14931: @end example
14932: @cindex instruction pointer
14933: For those unfamiliar with the names: @code{ip} is the Forth instruction
14934: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
14935: execution token and points to the code field of the next word to be
14936: executed; The @code{ca} (code address) fetched from there points to some
14937: executable code, e.g., a primitive or the colon definition handler
14938: @code{docol}.
14939:
14940: @cindex @code{NEXT}, direct threaded
14941: @cindex direct threaded inner interpreter
14942: @cindex inner interpreter, direct threaded
14943: Direct threading is even simpler:
14944: @example
14945: ca = *ip++;
14946: goto *ca;
14947: @end example
14948:
14949: Of course we have packaged the whole thing neatly in macros called
14950: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
14951:
14952: @menu
14953: * Scheduling::
14954: * Direct or Indirect Threaded?::
14955: * Dynamic Superinstructions::
14956: * DOES>::
14957: @end menu
14958:
14959: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
14960: @subsection Scheduling
14961: @cindex inner interpreter optimization
14962:
14963: There is a little complication: Pipelined and superscalar processors,
14964: i.e., RISC and some modern CISC machines can process independent
14965: instructions while waiting for the results of an instruction. The
14966: compiler usually reorders (schedules) the instructions in a way that
14967: achieves good usage of these delay slots. However, on our first tries
14968: the compiler did not do well on scheduling primitives. E.g., for
14969: @code{+} implemented as
14970: @example
14971: n=sp[0]+sp[1];
14972: sp++;
14973: sp[0]=n;
14974: NEXT;
14975: @end example
14976: the @code{NEXT} comes strictly after the other code, i.e., there is
14977: nearly no scheduling. After a little thought the problem becomes clear:
14978: The compiler cannot know that @code{sp} and @code{ip} point to different
14979: addresses (and the version of @code{gcc} we used would not know it even
14980: if it was possible), so it could not move the load of the cfa above the
14981: store to the TOS. Indeed the pointers could be the same, if code on or
14982: very near the top of stack were executed. In the interest of speed we
14983: chose to forbid this probably unused ``feature'' and helped the compiler
14984: in scheduling: @code{NEXT} is divided into several parts:
14985: @code{NEXT_P0}, @code{NEXT_P1} and @code{NEXT_P2}). @code{+} now looks
14986: like:
14987: @example
14988: NEXT_P0;
14989: n=sp[0]+sp[1];
14990: sp++;
14991: NEXT_P1;
14992: sp[0]=n;
14993: NEXT_P2;
14994: @end example
14995:
14996: There are various schemes that distribute the different operations of
14997: NEXT between these parts in several ways; in general, different schemes
14998: perform best on different processors. We use a scheme for most
14999: architectures that performs well for most processors of this
15000: architecture; in the future we may switch to benchmarking and chosing
15001: the scheme on installation time.
15002:
15003:
15004: @node Direct or Indirect Threaded?, Dynamic Superinstructions, Scheduling, Threading
15005: @subsection Direct or Indirect Threaded?
15006: @cindex threading, direct or indirect?
15007:
15008: Threaded forth code consists of references to primitives (simple machine
15009: code routines like @code{+}) and to non-primitives (e.g., colon
15010: definitions, variables, constants); for a specific class of
15011: non-primitives (e.g., variables) there is one code routine (e.g.,
15012: @code{dovar}), but each variable needs a separate reference to its data.
15013:
15014: Traditionally Forth has been implemented as indirect threaded code,
15015: because this allows to use only one cell to reference a non-primitive
15016: (basically you point to the data, and find the code address there).
15017:
15018: @cindex primitive-centric threaded code
15019: However, threaded code in Gforth (since 0.6.0) uses two cells for
15020: non-primitives, one for the code address, and one for the data address;
15021: the data pointer is an immediate argument for the virtual machine
15022: instruction represented by the code address. We call this
15023: @emph{primitive-centric} threaded code, because all code addresses point
15024: to simple primitives. E.g., for a variable, the code address is for
15025: @code{lit} (also used for integer literals like @code{99}).
15026:
15027: Primitive-centric threaded code allows us to use (faster) direct
15028: threading as dispatch method, completely portably (direct threaded code
15029: in Gforth before 0.6.0 required architecture-specific code). It also
15030: eliminates the performance problems related to I-cache consistency that
15031: 386 implementations have with direct threaded code, and allows
15032: additional optimizations.
15033:
15034: @cindex hybrid direct/indirect threaded code
15035: There is a catch, however: the @var{xt} parameter of @code{execute} can
15036: occupy only one cell, so how do we pass non-primitives with their code
15037: @emph{and} data addresses to them? Our answer is to use indirect
15038: threaded dispatch for @code{execute} and other words that use a
15039: single-cell xt. So, normal threaded code in colon definitions uses
15040: direct threading, and @code{execute} and similar words, which dispatch
15041: to xts on the data stack, use indirect threaded code. We call this
15042: @emph{hybrid direct/indirect} threaded code.
15043:
15044: @cindex engines, gforth vs. gforth-fast vs. gforth-itc
15045: @cindex gforth engine
15046: @cindex gforth-fast engine
15047: The engines @command{gforth} and @command{gforth-fast} use hybrid
15048: direct/indirect threaded code. This means that with these engines you
15049: cannot use @code{,} to compile an xt. Instead, you have to use
15050: @code{compile,}.
15051:
15052: @cindex gforth-itc engine
15053: If you want to compile xts with @code{,}, use @command{gforth-itc}.
15054: This engine uses plain old indirect threaded code. It still compiles in
15055: a primitive-centric style, so you cannot use @code{compile,} instead of
15056: @code{,} (e.g., for producing tables of xts with @code{] word1 word2
15057: ... [}). If you want to do that, you have to use @command{gforth-itc}
15058: and execute @code{' , is compile,}. Your program can check if it is
15059: running on a hybrid direct/indirect threaded engine or a pure indirect
15060: threaded engine with @code{threading-method} (@pxref{Threading Words}).
15061:
15062:
15063: @node Dynamic Superinstructions, DOES>, Direct or Indirect Threaded?, Threading
15064: @subsection Dynamic Superinstructions
15065: @cindex Dynamic superinstructions with replication
15066: @cindex Superinstructions
15067: @cindex Replication
15068:
15069: The engines @command{gforth} and @command{gforth-fast} use another
15070: optimization: Dynamic superinstructions with replication. As an
15071: example, consider the following colon definition:
15072:
15073: @example
15074: : squared ( n1 -- n2 )
15075: dup * ;
15076: @end example
15077:
15078: Gforth compiles this into the threaded code sequence
15079:
15080: @example
15081: dup
15082: *
15083: ;s
15084: @end example
15085:
15086: In normal direct threaded code there is a code address occupying one
15087: cell for each of these primitives. Each code address points to a
15088: machine code routine, and the interpreter jumps to this machine code in
15089: order to execute the primitive. The routines for these three
15090: primitives are (in @command{gforth-fast} on the 386):
15091:
15092: @example
15093: Code dup
15094: ( $804B950 ) add esi , # -4 \ $83 $C6 $FC
15095: ( $804B953 ) add ebx , # 4 \ $83 $C3 $4
15096: ( $804B956 ) mov dword ptr 4 [esi] , ecx \ $89 $4E $4
15097: ( $804B959 ) jmp dword ptr FC [ebx] \ $FF $63 $FC
15098: end-code
15099: Code *
15100: ( $804ACC4 ) mov eax , dword ptr 4 [esi] \ $8B $46 $4
15101: ( $804ACC7 ) add esi , # 4 \ $83 $C6 $4
15102: ( $804ACCA ) add ebx , # 4 \ $83 $C3 $4
15103: ( $804ACCD ) imul ecx , eax \ $F $AF $C8
15104: ( $804ACD0 ) jmp dword ptr FC [ebx] \ $FF $63 $FC
15105: end-code
15106: Code ;s
15107: ( $804A693 ) mov eax , dword ptr [edi] \ $8B $7
15108: ( $804A695 ) add edi , # 4 \ $83 $C7 $4
15109: ( $804A698 ) lea ebx , dword ptr 4 [eax] \ $8D $58 $4
15110: ( $804A69B ) jmp dword ptr FC [ebx] \ $FF $63 $FC
15111: end-code
15112: @end example
15113:
15114: With dynamic superinstructions and replication the compiler does not
15115: just lay down the threaded code, but also copies the machine code
15116: fragments, usually without the jump at the end.
15117:
15118: @example
15119: ( $4057D27D ) add esi , # -4 \ $83 $C6 $FC
15120: ( $4057D280 ) add ebx , # 4 \ $83 $C3 $4
15121: ( $4057D283 ) mov dword ptr 4 [esi] , ecx \ $89 $4E $4
15122: ( $4057D286 ) mov eax , dword ptr 4 [esi] \ $8B $46 $4
15123: ( $4057D289 ) add esi , # 4 \ $83 $C6 $4
15124: ( $4057D28C ) add ebx , # 4 \ $83 $C3 $4
15125: ( $4057D28F ) imul ecx , eax \ $F $AF $C8
15126: ( $4057D292 ) mov eax , dword ptr [edi] \ $8B $7
15127: ( $4057D294 ) add edi , # 4 \ $83 $C7 $4
15128: ( $4057D297 ) lea ebx , dword ptr 4 [eax] \ $8D $58 $4
15129: ( $4057D29A ) jmp dword ptr FC [ebx] \ $FF $63 $FC
15130: @end example
15131:
15132: Only when a threaded-code control-flow change happens (e.g., in
15133: @code{;s}), the jump is appended. This optimization eliminates many of
15134: these jumps and makes the rest much more predictable. The speedup
15135: depends on the processor and the application; on the Athlon and Pentium
15136: III this optimization typically produces a speedup by a factor of 2.
15137:
15138: The code addresses in the direct-threaded code are set to point to the
15139: appropriate points in the copied machine code, in this example like
15140: this:
15141:
15142: @example
15143: primitive code address
15144: dup $4057D27D
15145: * $4057D286
15146: ;s $4057D292
15147: @end example
15148:
15149: Thus there can be threaded-code jumps to any place in this piece of
15150: code. This also simplifies decompilation quite a bit.
15151:
15152: @cindex --no-dynamic command-line option
15153: @cindex --no-super command-line option
15154: You can disable this optimization with @option{--no-dynamic}. You can
15155: use the copying without eliminating the jumps (i.e., dynamic
15156: replication, but without superinstructions) with @option{--no-super};
15157: this gives the branch prediction benefit alone; the effect on
15158: performance depends on the CPU; on the Athlon and Pentium III the
15159: speedup is a little less than for dynamic superinstructions with
15160: replication.
15161:
15162: @cindex patching threaded code
15163: One use of these options is if you want to patch the threaded code.
15164: With superinstructions, many of the dispatch jumps are eliminated, so
15165: patching often has no effect. These options preserve all the dispatch
15166: jumps.
15167:
15168: @cindex --dynamic command-line option
15169: On some machines dynamic superinstructions are disabled by default,
15170: because it is unsafe on these machines. However, if you feel
15171: adventurous, you can enable it with @option{--dynamic}.
15172:
15173: @node DOES>, , Dynamic Superinstructions, Threading
15174: @subsection DOES>
15175: @cindex @code{DOES>} implementation
15176:
15177: @cindex @code{dodoes} routine
15178: @cindex @code{DOES>}-code
15179: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
15180: the chunk of code executed by every word defined by a
15181: @code{CREATE}...@code{DOES>} pair; actually with primitive-centric code,
15182: this is only needed if the xt of the word is @code{execute}d. The main
15183: problem here is: How to find the Forth code to be executed, i.e. the
15184: code after the @code{DOES>} (the @code{DOES>}-code)? There are two
15185: solutions:
15186:
15187: In fig-Forth the code field points directly to the @code{dodoes} and the
15188: @code{DOES>}-code address is stored in the cell after the code address
15189: (i.e. at @code{@i{CFA} cell+}). It may seem that this solution is
15190: illegal in the Forth-79 and all later standards, because in fig-Forth
15191: this address lies in the body (which is illegal in these
15192: standards). However, by making the code field larger for all words this
15193: solution becomes legal again. We use this approach. Leaving a cell
15194: unused in most words is a bit wasteful, but on the machines we are
15195: targeting this is hardly a problem.
15196:
15197:
15198: @node Primitives, Performance, Threading, Engine
15199: @section Primitives
15200: @cindex primitives, implementation
15201: @cindex virtual machine instructions, implementation
15202:
15203: @menu
15204: * Automatic Generation::
15205: * TOS Optimization::
15206: * Produced code::
15207: @end menu
15208:
15209: @node Automatic Generation, TOS Optimization, Primitives, Primitives
15210: @subsection Automatic Generation
15211: @cindex primitives, automatic generation
15212:
15213: @cindex @file{prims2x.fs}
15214:
15215: Since the primitives are implemented in a portable language, there is no
15216: longer any need to minimize the number of primitives. On the contrary,
15217: having many primitives has an advantage: speed. In order to reduce the
15218: number of errors in primitives and to make programming them easier, we
15219: provide a tool, the primitive generator (@file{prims2x.fs} aka Vmgen,
15220: @pxref{Top, Vmgen, Introduction, vmgen, Vmgen}), that automatically
15221: generates most (and sometimes all) of the C code for a primitive from
15222: the stack effect notation. The source for a primitive has the following
15223: form:
15224:
15225: @cindex primitive source format
15226: @format
15227: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
15228: [@code{""}@i{glossary entry}@code{""}]
15229: @i{C code}
15230: [@code{:}
15231: @i{Forth code}]
15232: @end format
15233:
15234: The items in brackets are optional. The category and glossary fields
15235: are there for generating the documentation, the Forth code is there
15236: for manual implementations on machines without GNU C. E.g., the source
15237: for the primitive @code{+} is:
15238: @example
15239: + ( n1 n2 -- n ) core plus
15240: n = n1+n2;
15241: @end example
15242:
15243: This looks like a specification, but in fact @code{n = n1+n2} is C
15244: code. Our primitive generation tool extracts a lot of information from
15245: the stack effect notations@footnote{We use a one-stack notation, even
15246: though we have separate data and floating-point stacks; The separate
15247: notation can be generated easily from the unified notation.}: The number
15248: of items popped from and pushed on the stack, their type, and by what
15249: name they are referred to in the C code. It then generates a C code
15250: prelude and postlude for each primitive. The final C code for @code{+}
15251: looks like this:
15252:
15253: @example
15254: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
15255: /* */ /* documentation */
15256: NAME("+") /* debugging output (with -DDEBUG) */
15257: @{
15258: DEF_CA /* definition of variable ca (indirect threading) */
15259: Cell n1; /* definitions of variables */
15260: Cell n2;
15261: Cell n;
15262: NEXT_P0; /* NEXT part 0 */
15263: n1 = (Cell) sp[1]; /* input */
15264: n2 = (Cell) TOS;
15265: sp += 1; /* stack adjustment */
15266: @{
15267: n = n1+n2; /* C code taken from the source */
15268: @}
15269: NEXT_P1; /* NEXT part 1 */
15270: TOS = (Cell)n; /* output */
15271: NEXT_P2; /* NEXT part 2 */
15272: @}
15273: @end example
15274:
15275: This looks long and inefficient, but the GNU C compiler optimizes quite
15276: well and produces optimal code for @code{+} on, e.g., the R3000 and the
15277: HP RISC machines: Defining the @code{n}s does not produce any code, and
15278: using them as intermediate storage also adds no cost.
15279:
15280: There are also other optimizations that are not illustrated by this
15281: example: assignments between simple variables are usually for free (copy
15282: propagation). If one of the stack items is not used by the primitive
15283: (e.g. in @code{drop}), the compiler eliminates the load from the stack
15284: (dead code elimination). On the other hand, there are some things that
15285: the compiler does not do, therefore they are performed by
15286: @file{prims2x.fs}: The compiler does not optimize code away that stores
15287: a stack item to the place where it just came from (e.g., @code{over}).
15288:
15289: While programming a primitive is usually easy, there are a few cases
15290: where the programmer has to take the actions of the generator into
15291: account, most notably @code{?dup}, but also words that do not (always)
15292: fall through to @code{NEXT}.
15293:
15294: For more information
15295:
15296: @node TOS Optimization, Produced code, Automatic Generation, Primitives
15297: @subsection TOS Optimization
15298: @cindex TOS optimization for primitives
15299: @cindex primitives, keeping the TOS in a register
15300:
15301: An important optimization for stack machine emulators, e.g., Forth
15302: engines, is keeping one or more of the top stack items in
15303: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
15304: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
15305: @itemize @bullet
15306: @item
15307: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
15308: due to fewer loads from and stores to the stack.
15309: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
15310: @i{y<n}, due to additional moves between registers.
15311: @end itemize
15312:
15313: @cindex -DUSE_TOS
15314: @cindex -DUSE_NO_TOS
15315: In particular, keeping one item in a register is never a disadvantage,
15316: if there are enough registers. Keeping two items in registers is a
15317: disadvantage for frequent words like @code{?branch}, constants,
15318: variables, literals and @code{i}. Therefore our generator only produces
15319: code that keeps zero or one items in registers. The generated C code
15320: covers both cases; the selection between these alternatives is made at
15321: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
15322: code for @code{+} is just a simple variable name in the one-item case,
15323: otherwise it is a macro that expands into @code{sp[0]}. Note that the
15324: GNU C compiler tries to keep simple variables like @code{TOS} in
15325: registers, and it usually succeeds, if there are enough registers.
15326:
15327: @cindex -DUSE_FTOS
15328: @cindex -DUSE_NO_FTOS
15329: The primitive generator performs the TOS optimization for the
15330: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
15331: operations the benefit of this optimization is even larger:
15332: floating-point operations take quite long on most processors, but can be
15333: performed in parallel with other operations as long as their results are
15334: not used. If the FP-TOS is kept in a register, this works. If
15335: it is kept on the stack, i.e., in memory, the store into memory has to
15336: wait for the result of the floating-point operation, lengthening the
15337: execution time of the primitive considerably.
15338:
15339: The TOS optimization makes the automatic generation of primitives a
15340: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
15341: @code{TOS} is not sufficient. There are some special cases to
15342: consider:
15343: @itemize @bullet
15344: @item In the case of @code{dup ( w -- w w )} the generator must not
15345: eliminate the store to the original location of the item on the stack,
15346: if the TOS optimization is turned on.
15347: @item Primitives with stack effects of the form @code{--}
15348: @i{out1}...@i{outy} must store the TOS to the stack at the start.
15349: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
15350: must load the TOS from the stack at the end. But for the null stack
15351: effect @code{--} no stores or loads should be generated.
15352: @end itemize
15353:
15354: @node Produced code, , TOS Optimization, Primitives
15355: @subsection Produced code
15356: @cindex primitives, assembly code listing
15357:
15358: @cindex @file{engine.s}
15359: To see what assembly code is produced for the primitives on your machine
15360: with your compiler and your flag settings, type @code{make engine.s} and
15361: look at the resulting file @file{engine.s}. Alternatively, you can also
15362: disassemble the code of primitives with @code{see} on some architectures.
15363:
15364: @node Performance, , Primitives, Engine
15365: @section Performance
15366: @cindex performance of some Forth interpreters
15367: @cindex engine performance
15368: @cindex benchmarking Forth systems
15369: @cindex Gforth performance
15370:
15371: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
15372: impossible to write a significantly faster threaded-code engine.
15373:
15374: On register-starved machines like the 386 architecture processors
15375: improvements are possible, because @code{gcc} does not utilize the
15376: registers as well as a human, even with explicit register declarations;
15377: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
15378: and hand-tuned it for the 486; this system is 1.19 times faster on the
15379: Sieve benchmark on a 486DX2/66 than Gforth compiled with
15380: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
15381: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
15382: registers fit in real registers (and we can even afford to use the TOS
15383: optimization), resulting in a speedup of 1.14 on the sieve over the
15384: earlier results. And dynamic superinstructions provide another speedup
15385: (but only around a factor 1.2 on the 486).
15386:
15387: @cindex Win32Forth performance
15388: @cindex NT Forth performance
15389: @cindex eforth performance
15390: @cindex ThisForth performance
15391: @cindex PFE performance
15392: @cindex TILE performance
15393: The potential advantage of assembly language implementations is not
15394: necessarily realized in complete Forth systems: We compared Gforth-0.5.9
15395: (direct threaded, compiled with @code{gcc-2.95.1} and
15396: @code{-DFORCE_REG}) with Win32Forth 1.2093 (newer versions are
15397: reportedly much faster), LMI's NT Forth (Beta, May 1994) and Eforth
15398: (with and without peephole (aka pinhole) optimization of the threaded
15399: code); all these systems were written in assembly language. We also
15400: compared Gforth with three systems written in C: PFE-0.9.14 (compiled
15401: with @code{gcc-2.6.3} with the default configuration for Linux:
15402: @code{-O2 -fomit-frame-pointer -DUSE_REGS -DUNROLL_NEXT}), ThisForth
15403: Beta (compiled with @code{gcc-2.6.3 -O3 -fomit-frame-pointer}; ThisForth
15404: employs peephole optimization of the threaded code) and TILE (compiled
15405: with @code{make opt}). We benchmarked Gforth, PFE, ThisForth and TILE on
15406: a 486DX2/66 under Linux. Kenneth O'Heskin kindly provided the results
15407: for Win32Forth and NT Forth on a 486DX2/66 with similar memory
15408: performance under Windows NT. Marcel Hendrix ported Eforth to Linux,
15409: then extended it to run the benchmarks, added the peephole optimizer,
15410: ran the benchmarks and reported the results.
15411:
15412: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
15413: matrix multiplication come from the Stanford integer benchmarks and have
15414: been translated into Forth by Martin Fraeman; we used the versions
15415: included in the TILE Forth package, but with bigger data set sizes; and
15416: a recursive Fibonacci number computation for benchmarking calling
15417: performance. The following table shows the time taken for the benchmarks
15418: scaled by the time taken by Gforth (in other words, it shows the speedup
15419: factor that Gforth achieved over the other systems).
15420:
15421: @example
15422: relative Win32- NT eforth This-
15423: time Gforth Forth Forth eforth +opt PFE Forth TILE
15424: sieve 1.00 2.16 1.78 2.16 1.32 2.46 4.96 13.37
15425: bubble 1.00 1.93 2.07 2.18 1.29 2.21 5.70
15426: matmul 1.00 1.92 1.76 1.90 0.96 2.06 5.32
15427: fib 1.00 2.32 2.03 1.86 1.31 2.64 4.55 6.54
15428: @end example
15429:
15430: You may be quite surprised by the good performance of Gforth when
15431: compared with systems written in assembly language. One important reason
15432: for the disappointing performance of these other systems is probably
15433: that they are not written optimally for the 486 (e.g., they use the
15434: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
15435: but costly method for relocating the Forth image: like @code{cforth}, it
15436: computes the actual addresses at run time, resulting in two address
15437: computations per @code{NEXT} (@pxref{Image File Background}).
15438:
15439: The speedup of Gforth over PFE, ThisForth and TILE can be easily
15440: explained with the self-imposed restriction of the latter systems to
15441: standard C, which makes efficient threading impossible (however, the
15442: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
15443: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
15444: Moreover, current C compilers have a hard time optimizing other aspects
15445: of the ThisForth and the TILE source.
15446:
15447: The performance of Gforth on 386 architecture processors varies widely
15448: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
15449: allocate any of the virtual machine registers into real machine
15450: registers by itself and would not work correctly with explicit register
15451: declarations, giving a significantly slower engine (on a 486DX2/66
15452: running the Sieve) than the one measured above.
15453:
15454: Note that there have been several releases of Win32Forth since the
15455: release presented here, so the results presented above may have little
15456: predictive value for the performance of Win32Forth today (results for
15457: the current release on an i486DX2/66 are welcome).
15458:
15459: @cindex @file{Benchres}
15460: In
15461: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
15462: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
15463: Maierhofer (presented at EuroForth '95), an indirect threaded version of
15464: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
15465: several native code systems; that version of Gforth is slower on a 486
15466: than the version used here. You can find a newer version of these
15467: measurements at
15468: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
15469: find numbers for Gforth on various machines in @file{Benchres}.
15470:
15471: @c ******************************************************************
15472: @c @node Binding to System Library, Cross Compiler, Engine, Top
15473: @c @chapter Binding to System Library
15474:
15475: @c ****************************************************************
15476: @node Cross Compiler, Bugs, Engine, Top
15477: @chapter Cross Compiler
15478: @cindex @file{cross.fs}
15479: @cindex cross-compiler
15480: @cindex metacompiler
15481: @cindex target compiler
15482:
15483: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
15484: mostly written in Forth, including crucial parts like the outer
15485: interpreter and compiler, it needs compiled Forth code to get
15486: started. The cross compiler allows to create new images for other
15487: architectures, even running under another Forth system.
15488:
15489: @menu
15490: * Using the Cross Compiler::
15491: * How the Cross Compiler Works::
15492: @end menu
15493:
15494: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
15495: @section Using the Cross Compiler
15496:
15497: The cross compiler uses a language that resembles Forth, but isn't. The
15498: main difference is that you can execute Forth code after definition,
15499: while you usually can't execute the code compiled by cross, because the
15500: code you are compiling is typically for a different computer than the
15501: one you are compiling on.
15502:
15503: @c anton: This chapter is somewhat different from waht I would expect: I
15504: @c would expect an explanation of the cross language and how to create an
15505: @c application image with it. The section explains some aspects of
15506: @c creating a Gforth kernel.
15507:
15508: The Makefile is already set up to allow you to create kernels for new
15509: architectures with a simple make command. The generic kernels using the
15510: GCC compiled virtual machine are created in the normal build process
15511: with @code{make}. To create a embedded Gforth executable for e.g. the
15512: 8086 processor (running on a DOS machine), type
15513:
15514: @example
15515: make kernl-8086.fi
15516: @end example
15517:
15518: This will use the machine description from the @file{arch/8086}
15519: directory to create a new kernel. A machine file may look like that:
15520:
15521: @example
15522: \ Parameter for target systems 06oct92py
15523:
15524: 4 Constant cell \ cell size in bytes
15525: 2 Constant cell<< \ cell shift to bytes
15526: 5 Constant cell>bit \ cell shift to bits
15527: 8 Constant bits/char \ bits per character
15528: 8 Constant bits/byte \ bits per byte [default: 8]
15529: 8 Constant float \ bytes per float
15530: 8 Constant /maxalign \ maximum alignment in bytes
15531: false Constant bigendian \ byte order
15532: ( true=big, false=little )
15533:
15534: include machpc.fs \ feature list
15535: @end example
15536:
15537: This part is obligatory for the cross compiler itself, the feature list
15538: is used by the kernel to conditionally compile some features in and out,
15539: depending on whether the target supports these features.
15540:
15541: There are some optional features, if you define your own primitives,
15542: have an assembler, or need special, nonstandard preparation to make the
15543: boot process work. @code{asm-include} includes an assembler,
15544: @code{prims-include} includes primitives, and @code{>boot} prepares for
15545: booting.
15546:
15547: @example
15548: : asm-include ." Include assembler" cr
15549: s" arch/8086/asm.fs" included ;
15550:
15551: : prims-include ." Include primitives" cr
15552: s" arch/8086/prim.fs" included ;
15553:
15554: : >boot ." Prepare booting" cr
15555: s" ' boot >body into-forth 1+ !" evaluate ;
15556: @end example
15557:
15558: These words are used as sort of macro during the cross compilation in
15559: the file @file{kernel/main.fs}. Instead of using these macros, it would
15560: be possible --- but more complicated --- to write a new kernel project
15561: file, too.
15562:
15563: @file{kernel/main.fs} expects the machine description file name on the
15564: stack; the cross compiler itself (@file{cross.fs}) assumes that either
15565: @code{mach-file} leaves a counted string on the stack, or
15566: @code{machine-file} leaves an address, count pair of the filename on the
15567: stack.
15568:
15569: The feature list is typically controlled using @code{SetValue}, generic
15570: files that are used by several projects can use @code{DefaultValue}
15571: instead. Both functions work like @code{Value}, when the value isn't
15572: defined, but @code{SetValue} works like @code{to} if the value is
15573: defined, and @code{DefaultValue} doesn't set anything, if the value is
15574: defined.
15575:
15576: @example
15577: \ generic mach file for pc gforth 03sep97jaw
15578:
15579: true DefaultValue NIL \ relocating
15580:
15581: >ENVIRON
15582:
15583: true DefaultValue file \ controls the presence of the
15584: \ file access wordset
15585: true DefaultValue OS \ flag to indicate a operating system
15586:
15587: true DefaultValue prims \ true: primitives are c-code
15588:
15589: true DefaultValue floating \ floating point wordset is present
15590:
15591: true DefaultValue glocals \ gforth locals are present
15592: \ will be loaded
15593: true DefaultValue dcomps \ double number comparisons
15594:
15595: true DefaultValue hash \ hashing primitives are loaded/present
15596:
15597: true DefaultValue xconds \ used together with glocals,
15598: \ special conditionals supporting gforths'
15599: \ local variables
15600: true DefaultValue header \ save a header information
15601:
15602: true DefaultValue backtrace \ enables backtrace code
15603:
15604: false DefaultValue ec
15605: false DefaultValue crlf
15606:
15607: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
15608:
15609: &16 KB DefaultValue stack-size
15610: &15 KB &512 + DefaultValue fstack-size
15611: &15 KB DefaultValue rstack-size
15612: &14 KB &512 + DefaultValue lstack-size
15613: @end example
15614:
15615: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
15616: @section How the Cross Compiler Works
15617:
15618: @node Bugs, Origin, Cross Compiler, Top
15619: @appendix Bugs
15620: @cindex bug reporting
15621:
15622: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
15623:
15624: If you find a bug, please submit a bug report through
15625: @uref{https://savannah.gnu.org/bugs/?func=addbug&group=gforth}.
15626:
15627: @itemize @bullet
15628: @item
15629: A program (or a sequence of keyboard commands) that reproduces the bug.
15630: @item
15631: A description of what you think constitutes the buggy behaviour.
15632: @item
15633: The Gforth version used (it is announced at the start of an
15634: interactive Gforth session).
15635: @item
15636: The machine and operating system (on Unix
15637: systems @code{uname -a} will report this information).
15638: @item
15639: The installation options (you can find the configure options at the
15640: start of @file{config.status}) and configuration (@code{configure}
15641: output or @file{config.cache}).
15642: @item
15643: A complete list of changes (if any) you (or your installer) have made to the
15644: Gforth sources.
15645: @end itemize
15646:
15647: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
15648: to Report Bugs, gcc.info, GNU C Manual}.
15649:
15650:
15651: @node Origin, Forth-related information, Bugs, Top
15652: @appendix Authors and Ancestors of Gforth
15653:
15654: @section Authors and Contributors
15655: @cindex authors of Gforth
15656: @cindex contributors to Gforth
15657:
15658: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
15659: Ertl. The third major author was Jens Wilke. Neal Crook contributed a
15660: lot to the manual. Assemblers and disassemblers were contributed by
15661: Andrew McKewan, Christian Pirker, Bernd Thallner, and Michal Revucky.
15662: Lennart Benschop (who was one of Gforth's first users, in mid-1993)
15663: and Stuart Ramsden inspired us with their continuous feedback. Lennart
15664: Benshop contributed @file{glosgen.fs}, while Stuart Ramsden has been
15665: working on automatic support for calling C libraries. Helpful comments
15666: also came from Paul Kleinrubatscher, Christian Pirker, Dirk Zoller,
15667: Marcel Hendrix, John Wavrik, Barrie Stott, Marc de Groot, Jorge
15668: Acerada, Bruce Hoyt, Robert Epprecht, Dennis Ruffer and David
15669: N. Williams. Since the release of Gforth-0.2.1 there were also helpful
15670: comments from many others; thank you all, sorry for not listing you
15671: here (but digging through my mailbox to extract your names is on my
15672: to-do list).
15673:
15674: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
15675: and autoconf, among others), and to the creators of the Internet: Gforth
15676: was developed across the Internet, and its authors did not meet
15677: physically for the first 4 years of development.
15678:
15679: @section Pedigree
15680: @cindex pedigree of Gforth
15681:
15682: Gforth descends from bigFORTH (1993) and fig-Forth. Of course, a
15683: significant part of the design of Gforth was prescribed by ANS Forth.
15684:
15685: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
15686: 32 bit native code version of VolksForth for the Atari ST, written
15687: mostly by Dietrich Weineck.
15688:
15689: VolksForth was written by Klaus Schleisiek, Bernd Pennemann, Georg
15690: Rehfeld and Dietrich Weineck for the C64 (called UltraForth there) in
15691: the mid-80s and ported to the Atari ST in 1986. It descends from fig-Forth.
15692:
15693: @c Henry Laxen and Mike Perry wrote F83 as a model implementation of the
15694: @c Forth-83 standard. !! Pedigree? When?
15695:
15696: A team led by Bill Ragsdale implemented fig-Forth on many processors in
15697: 1979. Robert Selzer and Bill Ragsdale developed the original
15698: implementation of fig-Forth for the 6502 based on microForth.
15699:
15700: The principal architect of microForth was Dean Sanderson. microForth was
15701: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
15702: the 1802, and subsequently implemented on the 8080, the 6800 and the
15703: Z80.
15704:
15705: All earlier Forth systems were custom-made, usually by Charles Moore,
15706: who discovered (as he puts it) Forth during the late 60s. The first full
15707: Forth existed in 1971.
15708:
15709: A part of the information in this section comes from
15710: @cite{@uref{http://www.forth.com/Content/History/History1.htm,The
15711: Evolution of Forth}} by Elizabeth D. Rather, Donald R. Colburn and
15712: Charles H. Moore, presented at the HOPL-II conference and preprinted
15713: in SIGPLAN Notices 28(3), 1993. You can find more historical and
15714: genealogical information about Forth there. For a more general (and
15715: graphical) Forth family tree look see
15716: @cite{@uref{http://www.complang.tuwien.ac.at/forth/family-tree/},
15717: Forth Family Tree and Timeline}.
15718:
15719: @c ------------------------------------------------------------------
15720: @node Forth-related information, Licenses, Origin, Top
15721: @appendix Other Forth-related information
15722: @cindex Forth-related information
15723:
15724: @c anton: I threw most of this stuff out, because it can be found through
15725: @c the FAQ and the FAQ is more likely to be up-to-date.
15726:
15727: @cindex comp.lang.forth
15728: @cindex frequently asked questions
15729: There is an active news group (comp.lang.forth) discussing Forth
15730: (including Gforth) and Forth-related issues. Its
15731: @uref{http://www.complang.tuwien.ac.at/forth/faq/faq-general-2.html,FAQs}
15732: (frequently asked questions and their answers) contains a lot of
15733: information on Forth. You should read it before posting to
15734: comp.lang.forth.
15735:
15736: The ANS Forth standard is most usable in its
15737: @uref{http://www.taygeta.com/forth/dpans.html, HTML form}.
15738:
15739: @c ---------------------------------------------------
15740: @node Licenses, Word Index, Forth-related information, Top
15741: @appendix Licenses
15742:
15743: @menu
15744: * GNU Free Documentation License:: License for copying this manual.
15745: * Copying:: GPL (for copying this software).
15746: @end menu
15747:
15748: @include fdl.texi
15749:
15750: @include gpl.texi
15751:
15752:
15753:
15754: @c ------------------------------------------------------------------
15755: @node Word Index, Concept Index, Licenses, Top
15756: @unnumbered Word Index
15757:
15758: This index is a list of Forth words that have ``glossary'' entries
15759: within this manual. Each word is listed with its stack effect and
15760: wordset.
15761:
15762: @printindex fn
15763:
15764: @c anton: the name index seems superfluous given the word and concept indices.
15765:
15766: @c @node Name Index, Concept Index, Word Index, Top
15767: @c @unnumbered Name Index
15768:
15769: @c This index is a list of Forth words that have ``glossary'' entries
15770: @c within this manual.
15771:
15772: @c @printindex ky
15773:
15774: @c -------------------------------------------------------
15775: @node Concept Index, , Word Index, Top
15776: @unnumbered Concept and Word Index
15777:
15778: Not all entries listed in this index are present verbatim in the
15779: text. This index also duplicates, in abbreviated form, all of the words
15780: listed in the Word Index (only the names are listed for the words here).
15781:
15782: @printindex cp
15783:
15784: @bye
15785:
15786:
15787:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to gforth.ds CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [gforth] / gforth / doc