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: @comment .. would be useful to have a word that identified all deferred words
15: @comment should semantics stuff in intro be moved to another section
16:
17:
18: @comment %**start of header (This is for running Texinfo on a region.)
19: @setfilename gforth.info
20: @settitle Gforth Manual
21: @dircategory GNU programming tools
22: @direntry
23: * Gforth: (gforth). A fast interpreter for the Forth language.
24: @end direntry
25: @c The Texinfo manual also recommends doing this, but for Gforth it may
26: @c not make much sense
27: @c @dircategory Individual utilities
28: @c @direntry
29: @c * Gforth: (gforth)Invoking Gforth. gforth, gforth-fast, gforthmi
30: @c @end direntry
31:
32: @comment @setchapternewpage odd
33: @comment TODO this gets left in by HTML converter
34: @macro progstyle {}
35: Programming style note:
36: @end macro
37:
38: @macro assignment {}
39: @table @i
40: @item Assignment:
41: @end macro
42: @macro endassignment {}
43: @end table
44: @end macro
45:
46: @comment %**end of header (This is for running Texinfo on a region.)
47:
48:
49: @comment ----------------------------------------------------------
50: @comment macros for beautifying glossary entries
51: @comment if these are used, need to strip them out for HTML converter
52: @comment else they get repeated verbatim in HTML output.
53: @comment .. not working yet.
54:
55: @macro GLOSS-START {}
56: @iftex
57: @ninerm
58: @end iftex
59: @end macro
60:
61: @macro GLOSS-END {}
62: @iftex
63: @rm
64: @end iftex
65: @end macro
66:
67: @comment ----------------------------------------------------------
68:
69:
70: @include version.texi
71:
72: @ifnottex
73: This file documents Gforth @value{VERSION}
74:
75: Copyright @copyright{} 1995-1999 Free Software Foundation, Inc.
76:
77: Permission is granted to make and distribute verbatim copies of
78: this manual provided the copyright notice and this permission notice
79: are preserved on all copies.
80:
81: @ignore
82: Permission is granted to process this file through TeX and print the
83: results, provided the printed document carries a copying permission
84: notice identical to this one except for the removal of this paragraph
85: (this paragraph not being relevant to the printed manual).
86:
87: @end ignore
88: Permission is granted to copy and distribute modified versions of this
89: manual under the conditions for verbatim copying, provided also that the
90: sections entitled "Distribution" and "General Public License" are
91: included exactly as in the original, and provided that the entire
92: resulting derived work is distributed under the terms of a permission
93: notice identical to this one.
94:
95: Permission is granted to copy and distribute translations of this manual
96: into another language, under the above conditions for modified versions,
97: except that the sections entitled "Distribution" and "General Public
98: License" may be included in a translation approved by the author instead
99: of in the original English.
100: @end ifnottex
101:
102: @finalout
103: @titlepage
104: @sp 10
105: @center @titlefont{Gforth Manual}
106: @sp 2
107: @center for version @value{VERSION}
108: @sp 2
109: @center Neal Crook
110: @center Anton Ertl
111: @center Bernd Paysan
112: @center Jens Wilke
113: @sp 3
114: @center This manual is permanently under construction and was last updated on 15-Mar-2000
115:
116: @comment The following two commands start the copyright page.
117: @page
118: @vskip 0pt plus 1filll
119: Copyright @copyright{} 1995--1999 Free Software Foundation, Inc.
120:
121: @comment !! Published by ... or You can get a copy of this manual ...
122:
123: Permission is granted to make and distribute verbatim copies of
124: this manual provided the copyright notice and this permission notice
125: are preserved on all copies.
126:
127: Permission is granted to copy and distribute modified versions of this
128: manual under the conditions for verbatim copying, provided also that the
129: sections entitled "Distribution" and "General Public License" are
130: included exactly as in the original, and provided that the entire
131: resulting derived work is distributed under the terms of a permission
132: notice identical to this one.
133:
134: Permission is granted to copy and distribute translations of this manual
135: into another language, under the above conditions for modified versions,
136: except that the sections entitled "Distribution" and "General Public
137: License" may be included in a translation approved by the author instead
138: of in the original English.
139: @end titlepage
140:
141: @node Top, License, (dir), (dir)
142: @ifnottex
143: Gforth is a free implementation of ANS Forth available on many
144: personal machines. This manual corresponds to version @value{VERSION}.
145: @end ifnottex
146:
147: @menu
148: * License:: The GPL
149: * Goals:: About the Gforth Project
150: * Gforth Environment:: Starting (and exiting) Gforth
151: * Tutorial:: Hands-on Forth Tutorial
152: * Introduction:: An introduction to ANS Forth
153: * Words:: Forth words available in Gforth
154: * Error messages:: How to interpret them
155: * Tools:: Programming tools
156: * ANS conformance:: Implementation-defined options etc.
157: * Model:: The abstract machine of Gforth
158: * Integrating Gforth:: Forth as scripting language for applications
159: * Emacs and Gforth:: The Gforth Mode
160: * Image Files:: @code{.fi} files contain compiled code
161: * Engine:: The inner interpreter and the primitives
162: * Binding to System Library::
163: * Cross Compiler:: The Cross Compiler
164: * Bugs:: How to report them
165: * Origin:: Authors and ancestors of Gforth
166: * Forth-related information:: Books and places to look on the WWW
167: * Word Index:: An item for each Forth word
168: * Name Index:: Forth words, only names listed
169: * Concept Index:: A menu covering many topics
170:
171: @detailmenu --- The Detailed Node Listing ---
172:
173: Goals of Gforth
174:
175: * Gforth Extensions Sinful?::
176:
177: Gforth Environment
178:
179: * Invoking Gforth:: Getting in
180: * Leaving Gforth:: Getting out
181: * Command-line editing::
182: * Upper and lower case::
183: * Environment variables:: that affect how Gforth starts up
184: * Gforth Files:: What gets installed and where
185: * Startup speed:: When 35ms is not fast enough ...
186:
187: Forth Tutorial
188:
189: * Starting Gforth Tutorial::
190: * Syntax Tutorial::
191: * Crash Course Tutorial::
192: * Stack Tutorial::
193: * Arithmetics Tutorial::
194: * Stack Manipulation Tutorial::
195: * Using files for Forth code Tutorial::
196: * Comments Tutorial::
197: * Colon Definitions Tutorial::
198: * Decompilation Tutorial::
199: * Stack-Effect Comments Tutorial::
200: * Types Tutorial::
201: * Factoring Tutorial::
202: * Designing the stack effect Tutorial::
203: * Local Variables Tutorial::
204: * Conditional execution Tutorial::
205: * Flags and Comparisons Tutorial::
206: * General Loops Tutorial::
207: * Counted loops Tutorial::
208: * Recursion Tutorial::
209: * Leaving definitions or loops Tutorial::
210: * Return Stack Tutorial::
211: * Memory Tutorial::
212: * Characters and Strings Tutorial::
213: * Alignment Tutorial::
214: * Interpretation and Compilation Semantics and Immediacy Tutorial::
215: * Execution Tokens Tutorial::
216: * Exceptions Tutorial::
217: * Defining Words Tutorial::
218: * Arrays and Records Tutorial::
219: * POSTPONE Tutorial::
220: * Literal Tutorial::
221: * Advanced macros Tutorial::
222: * Compilation Tokens Tutorial::
223: * Wordlists and Search Order Tutorial::
224:
225: An Introduction to ANS Forth
226:
227: * Introducing the Text Interpreter::
228: * Stacks and Postfix notation::
229: * Your first definition::
230: * How does that work?::
231: * Forth is written in Forth::
232: * Review - elements of a Forth system::
233: * Where to go next::
234: * Exercises::
235:
236: Forth Words
237:
238: * Notation::
239: * Comments::
240: * Boolean Flags::
241: * Arithmetic::
242: * Stack Manipulation::
243: * Memory::
244: * Control Structures::
245: * Defining Words::
246: * Interpretation and Compilation Semantics::
247: * Tokens for Words::
248: * The Text Interpreter::
249: * Word Lists::
250: * Environmental Queries::
251: * Files::
252: * Blocks::
253: * Other I/O::
254: * Programming Tools::
255: * Assembler and Code Words::
256: * Threading Words::
257: * Locals::
258: * Structures::
259: * Object-oriented Forth::
260: * Passing Commands to the OS::
261: * Keeping track of Time::
262: * Miscellaneous Words::
263:
264: Arithmetic
265:
266: * Single precision::
267: * Bitwise operations::
268: * Double precision:: Double-cell integer arithmetic
269: * Numeric comparison::
270: * Mixed precision:: Operations with single and double-cell integers
271: * Floating Point::
272:
273: Stack Manipulation
274:
275: * Data stack::
276: * Floating point stack::
277: * Return stack::
278: * Locals stack::
279: * Stack pointer manipulation::
280:
281: Memory
282:
283: * Memory model::
284: * Dictionary allocation::
285: * Heap Allocation::
286: * Memory Access::
287: * Address arithmetic::
288: * Memory Blocks::
289:
290: Control Structures
291:
292: * Selection:: IF ... ELSE ... ENDIF
293: * Simple Loops:: BEGIN ...
294: * Counted Loops:: DO
295: * Arbitrary control structures::
296: * Calls and returns::
297: * Exception Handling::
298:
299: Defining Words
300:
301: * CREATE::
302: * Variables:: Variables and user variables
303: * Constants::
304: * Values:: Initialised variables
305: * Colon Definitions::
306: * Anonymous Definitions:: Definitions without names
307: * User-defined Defining Words::
308: * Deferred words:: Allow forward references
309: * Aliases::
310: * Supplying names::
311:
312: Interpretation and Compilation Semantics
313:
314: * Combined words::
315:
316: The Text Interpreter
317:
318: * Input Sources::
319: * Number Conversion::
320: * Interpret/Compile states::
321: * Literals::
322: * Interpreter Directives::
323:
324: Word Lists
325:
326: * Why use word lists?::
327: * Word list examples::
328:
329: Files
330:
331: * Forth source files::
332: * General files::
333: * Search Paths::
334:
335: Search Paths
336:
337: * Forth Search Paths::
338: * General Search Paths::
339:
340: Other I/O
341:
342: * Simple numeric output:: Predefined formats
343: * Formatted numeric output:: Formatted (pictured) output
344: * String Formats:: How Forth stores strings in memory
345: * Displaying characters and strings:: Other stuff
346: * Input:: Input
347:
348: Programming Tools
349:
350: * Debugging:: Simple and quick.
351: * Assertions:: Making your programs self-checking.
352: * Singlestep Debugger:: Executing your program word by word.
353:
354: Locals
355:
356: * Gforth locals::
357: * ANS Forth locals::
358:
359: Gforth locals
360:
361: * Where are locals visible by name?::
362: * How long do locals live?::
363: * Programming Style::
364: * Implementation::
365:
366: Structures
367:
368: * Why explicit structure support?::
369: * Structure Usage::
370: * Structure Naming Convention::
371: * Structure Implementation::
372: * Structure Glossary::
373:
374: Object-oriented Forth
375:
376: * Why object-oriented programming?::
377: * Object-Oriented Terminology::
378: * Objects::
379: * OOF::
380: * Mini-OOF::
381: * Comparison with other object models::
382:
383: The @file{objects.fs} model
384:
385: * Properties of the Objects model::
386: * Basic Objects Usage::
387: * The Objects base class::
388: * Creating objects::
389: * Object-Oriented Programming Style::
390: * Class Binding::
391: * Method conveniences::
392: * Classes and Scoping::
393: * Dividing classes::
394: * Object Interfaces::
395: * Objects Implementation::
396: * Objects Glossary::
397:
398: The @file{oof.fs} model
399:
400: * Properties of the OOF model::
401: * Basic OOF Usage::
402: * The OOF base class::
403: * Class Declaration::
404: * Class Implementation::
405:
406: The @file{mini-oof.fs} model
407:
408: * Basic Mini-OOF Usage::
409: * Mini-OOF Example::
410: * Mini-OOF Implementation::
411: * Comparison with other object models::
412:
413: Tools
414:
415: * ANS Report:: Report the words used, sorted by wordset.
416:
417: ANS conformance
418:
419: * The Core Words::
420: * The optional Block word set::
421: * The optional Double Number word set::
422: * The optional Exception word set::
423: * The optional Facility word set::
424: * The optional File-Access word set::
425: * The optional Floating-Point word set::
426: * The optional Locals word set::
427: * The optional Memory-Allocation word set::
428: * The optional Programming-Tools word set::
429: * The optional Search-Order word set::
430:
431: The Core Words
432:
433: * core-idef:: Implementation Defined Options
434: * core-ambcond:: Ambiguous Conditions
435: * core-other:: Other System Documentation
436:
437: The optional Block word set
438:
439: * block-idef:: Implementation Defined Options
440: * block-ambcond:: Ambiguous Conditions
441: * block-other:: Other System Documentation
442:
443: The optional Double Number word set
444:
445: * double-ambcond:: Ambiguous Conditions
446:
447: The optional Exception word set
448:
449: * exception-idef:: Implementation Defined Options
450:
451: The optional Facility word set
452:
453: * facility-idef:: Implementation Defined Options
454: * facility-ambcond:: Ambiguous Conditions
455:
456: The optional File-Access word set
457:
458: * file-idef:: Implementation Defined Options
459: * file-ambcond:: Ambiguous Conditions
460:
461: The optional Floating-Point word set
462:
463: * floating-idef:: Implementation Defined Options
464: * floating-ambcond:: Ambiguous Conditions
465:
466: The optional Locals word set
467:
468: * locals-idef:: Implementation Defined Options
469: * locals-ambcond:: Ambiguous Conditions
470:
471: The optional Memory-Allocation word set
472:
473: * memory-idef:: Implementation Defined Options
474:
475: The optional Programming-Tools word set
476:
477: * programming-idef:: Implementation Defined Options
478: * programming-ambcond:: Ambiguous Conditions
479:
480: The optional Search-Order word set
481:
482: * search-idef:: Implementation Defined Options
483: * search-ambcond:: Ambiguous Conditions
484:
485: Image Files
486:
487: * Image Licensing Issues:: Distribution terms for images.
488: * Image File Background:: Why have image files?
489: * Non-Relocatable Image Files:: don't always work.
490: * Data-Relocatable Image Files:: are better.
491: * Fully Relocatable Image Files:: better yet.
492: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
493: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
494: * Modifying the Startup Sequence:: and turnkey applications.
495:
496: Fully Relocatable Image Files
497:
498: * gforthmi:: The normal way
499: * cross.fs:: The hard way
500:
501: Engine
502:
503: * Portability::
504: * Threading::
505: * Primitives::
506: * Performance::
507:
508: Threading
509:
510: * Scheduling::
511: * Direct or Indirect Threaded?::
512: * DOES>::
513:
514: Primitives
515:
516: * Automatic Generation::
517: * TOS Optimization::
518: * Produced code::
519:
520: Cross Compiler
521:
522: * Using the Cross Compiler::
523: * How the Cross Compiler Works::
524:
525: Other Forth-related information
526:
527: * Internet resources::
528: * Books::
529: * The Forth Interest Group::
530: * Conferences::
531:
532: @end detailmenu
533: @end menu
534:
535: @node License, Goals, Top, Top
536: @unnumbered GNU GENERAL PUBLIC LICENSE
537: @center Version 2, June 1991
538:
539: @display
540: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
541: 675 Mass Ave, Cambridge, MA 02139, USA
542:
543: Everyone is permitted to copy and distribute verbatim copies
544: of this license document, but changing it is not allowed.
545: @end display
546:
547: @unnumberedsec Preamble
548:
549: The licenses for most software are designed to take away your
550: freedom to share and change it. By contrast, the GNU General Public
551: License is intended to guarantee your freedom to share and change free
552: software---to make sure the software is free for all its users. This
553: General Public License applies to most of the Free Software
554: Foundation's software and to any other program whose authors commit to
555: using it. (Some other Free Software Foundation software is covered by
556: the GNU Library General Public License instead.) You can apply it to
557: your programs, too.
558:
559: When we speak of free software, we are referring to freedom, not
560: price. Our General Public Licenses are designed to make sure that you
561: have the freedom to distribute copies of free software (and charge for
562: this service if you wish), that you receive source code or can get it
563: if you want it, that you can change the software or use pieces of it
564: in new free programs; and that you know you can do these things.
565:
566: To protect your rights, we need to make restrictions that forbid
567: anyone to deny you these rights or to ask you to surrender the rights.
568: These restrictions translate to certain responsibilities for you if you
569: distribute copies of the software, or if you modify it.
570:
571: For example, if you distribute copies of such a program, whether
572: gratis or for a fee, you must give the recipients all the rights that
573: you have. You must make sure that they, too, receive or can get the
574: source code. And you must show them these terms so they know their
575: rights.
576:
577: We protect your rights with two steps: (1) copyright the software, and
578: (2) offer you this license which gives you legal permission to copy,
579: distribute and/or modify the software.
580:
581: Also, for each author's protection and ours, we want to make certain
582: that everyone understands that there is no warranty for this free
583: software. If the software is modified by someone else and passed on, we
584: want its recipients to know that what they have is not the original, so
585: that any problems introduced by others will not reflect on the original
586: authors' reputations.
587:
588: Finally, any free program is threatened constantly by software
589: patents. We wish to avoid the danger that redistributors of a free
590: program will individually obtain patent licenses, in effect making the
591: program proprietary. To prevent this, we have made it clear that any
592: patent must be licensed for everyone's free use or not licensed at all.
593:
594: The precise terms and conditions for copying, distribution and
595: modification follow.
596:
597: @iftex
598: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
599: @end iftex
600: @ifnottex
601: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
602: @end ifnottex
603:
604: @enumerate 0
605: @item
606: This License applies to any program or other work which contains
607: a notice placed by the copyright holder saying it may be distributed
608: under the terms of this General Public License. The ``Program'', below,
609: refers to any such program or work, and a ``work based on the Program''
610: means either the Program or any derivative work under copyright law:
611: that is to say, a work containing the Program or a portion of it,
612: either verbatim or with modifications and/or translated into another
613: language. (Hereinafter, translation is included without limitation in
614: the term ``modification''.) Each licensee is addressed as ``you''.
615:
616: Activities other than copying, distribution and modification are not
617: covered by this License; they are outside its scope. The act of
618: running the Program is not restricted, and the output from the Program
619: is covered only if its contents constitute a work based on the
620: Program (independent of having been made by running the Program).
621: Whether that is true depends on what the Program does.
622:
623: @item
624: You may copy and distribute verbatim copies of the Program's
625: source code as you receive it, in any medium, provided that you
626: conspicuously and appropriately publish on each copy an appropriate
627: copyright notice and disclaimer of warranty; keep intact all the
628: notices that refer to this License and to the absence of any warranty;
629: and give any other recipients of the Program a copy of this License
630: along with the Program.
631:
632: You may charge a fee for the physical act of transferring a copy, and
633: you may at your option offer warranty protection in exchange for a fee.
634:
635: @item
636: You may modify your copy or copies of the Program or any portion
637: of it, thus forming a work based on the Program, and copy and
638: distribute such modifications or work under the terms of Section 1
639: above, provided that you also meet all of these conditions:
640:
641: @enumerate a
642: @item
643: You must cause the modified files to carry prominent notices
644: stating that you changed the files and the date of any change.
645:
646: @item
647: You must cause any work that you distribute or publish, that in
648: whole or in part contains or is derived from the Program or any
649: part thereof, to be licensed as a whole at no charge to all third
650: parties under the terms of this License.
651:
652: @item
653: If the modified program normally reads commands interactively
654: when run, you must cause it, when started running for such
655: interactive use in the most ordinary way, to print or display an
656: announcement including an appropriate copyright notice and a
657: notice that there is no warranty (or else, saying that you provide
658: a warranty) and that users may redistribute the program under
659: these conditions, and telling the user how to view a copy of this
660: License. (Exception: if the Program itself is interactive but
661: does not normally print such an announcement, your work based on
662: the Program is not required to print an announcement.)
663: @end enumerate
664:
665: These requirements apply to the modified work as a whole. If
666: identifiable sections of that work are not derived from the Program,
667: and can be reasonably considered independent and separate works in
668: themselves, then this License, and its terms, do not apply to those
669: sections when you distribute them as separate works. But when you
670: distribute the same sections as part of a whole which is a work based
671: on the Program, the distribution of the whole must be on the terms of
672: this License, whose permissions for other licensees extend to the
673: entire whole, and thus to each and every part regardless of who wrote it.
674:
675: Thus, it is not the intent of this section to claim rights or contest
676: your rights to work written entirely by you; rather, the intent is to
677: exercise the right to control the distribution of derivative or
678: collective works based on the Program.
679:
680: In addition, mere aggregation of another work not based on the Program
681: with the Program (or with a work based on the Program) on a volume of
682: a storage or distribution medium does not bring the other work under
683: the scope of this License.
684:
685: @item
686: You may copy and distribute the Program (or a work based on it,
687: under Section 2) in object code or executable form under the terms of
688: Sections 1 and 2 above provided that you also do one of the following:
689:
690: @enumerate a
691: @item
692: Accompany it with the complete corresponding machine-readable
693: source code, which must be distributed under the terms of Sections
694: 1 and 2 above on a medium customarily used for software interchange; or,
695:
696: @item
697: Accompany it with a written offer, valid for at least three
698: years, to give any third party, for a charge no more than your
699: cost of physically performing source distribution, a complete
700: machine-readable copy of the corresponding source code, to be
701: distributed under the terms of Sections 1 and 2 above on a medium
702: customarily used for software interchange; or,
703:
704: @item
705: Accompany it with the information you received as to the offer
706: to distribute corresponding source code. (This alternative is
707: allowed only for noncommercial distribution and only if you
708: received the program in object code or executable form with such
709: an offer, in accord with Subsection b above.)
710: @end enumerate
711:
712: The source code for a work means the preferred form of the work for
713: making modifications to it. For an executable work, complete source
714: code means all the source code for all modules it contains, plus any
715: associated interface definition files, plus the scripts used to
716: control compilation and installation of the executable. However, as a
717: special exception, the source code distributed need not include
718: anything that is normally distributed (in either source or binary
719: form) with the major components (compiler, kernel, and so on) of the
720: operating system on which the executable runs, unless that component
721: itself accompanies the executable.
722:
723: If distribution of executable or object code is made by offering
724: access to copy from a designated place, then offering equivalent
725: access to copy the source code from the same place counts as
726: distribution of the source code, even though third parties are not
727: compelled to copy the source along with the object code.
728:
729: @item
730: You may not copy, modify, sublicense, or distribute the Program
731: except as expressly provided under this License. Any attempt
732: otherwise to copy, modify, sublicense or distribute the Program is
733: void, and will automatically terminate your rights under this License.
734: However, parties who have received copies, or rights, from you under
735: this License will not have their licenses terminated so long as such
736: parties remain in full compliance.
737:
738: @item
739: You are not required to accept this License, since you have not
740: signed it. However, nothing else grants you permission to modify or
741: distribute the Program or its derivative works. These actions are
742: prohibited by law if you do not accept this License. Therefore, by
743: modifying or distributing the Program (or any work based on the
744: Program), you indicate your acceptance of this License to do so, and
745: all its terms and conditions for copying, distributing or modifying
746: the Program or works based on it.
747:
748: @item
749: Each time you redistribute the Program (or any work based on the
750: Program), the recipient automatically receives a license from the
751: original licensor to copy, distribute or modify the Program subject to
752: these terms and conditions. You may not impose any further
753: restrictions on the recipients' exercise of the rights granted herein.
754: You are not responsible for enforcing compliance by third parties to
755: this License.
756:
757: @item
758: If, as a consequence of a court judgment or allegation of patent
759: infringement or for any other reason (not limited to patent issues),
760: conditions are imposed on you (whether by court order, agreement or
761: otherwise) that contradict the conditions of this License, they do not
762: excuse you from the conditions of this License. If you cannot
763: distribute so as to satisfy simultaneously your obligations under this
764: License and any other pertinent obligations, then as a consequence you
765: may not distribute the Program at all. For example, if a patent
766: license would not permit royalty-free redistribution of the Program by
767: all those who receive copies directly or indirectly through you, then
768: the only way you could satisfy both it and this License would be to
769: refrain entirely from distribution of the Program.
770:
771: If any portion of this section is held invalid or unenforceable under
772: any particular circumstance, the balance of the section is intended to
773: apply and the section as a whole is intended to apply in other
774: circumstances.
775:
776: It is not the purpose of this section to induce you to infringe any
777: patents or other property right claims or to contest validity of any
778: such claims; this section has the sole purpose of protecting the
779: integrity of the free software distribution system, which is
780: implemented by public license practices. Many people have made
781: generous contributions to the wide range of software distributed
782: through that system in reliance on consistent application of that
783: system; it is up to the author/donor to decide if he or she is willing
784: to distribute software through any other system and a licensee cannot
785: impose that choice.
786:
787: This section is intended to make thoroughly clear what is believed to
788: be a consequence of the rest of this License.
789:
790: @item
791: If the distribution and/or use of the Program is restricted in
792: certain countries either by patents or by copyrighted interfaces, the
793: original copyright holder who places the Program under this License
794: may add an explicit geographical distribution limitation excluding
795: those countries, so that distribution is permitted only in or among
796: countries not thus excluded. In such case, this License incorporates
797: the limitation as if written in the body of this License.
798:
799: @item
800: The Free Software Foundation may publish revised and/or new versions
801: of the General Public License from time to time. Such new versions will
802: be similar in spirit to the present version, but may differ in detail to
803: address new problems or concerns.
804:
805: Each version is given a distinguishing version number. If the Program
806: specifies a version number of this License which applies to it and ``any
807: later version'', you have the option of following the terms and conditions
808: either of that version or of any later version published by the Free
809: Software Foundation. If the Program does not specify a version number of
810: this License, you may choose any version ever published by the Free Software
811: Foundation.
812:
813: @item
814: If you wish to incorporate parts of the Program into other free
815: programs whose distribution conditions are different, write to the author
816: to ask for permission. For software which is copyrighted by the Free
817: Software Foundation, write to the Free Software Foundation; we sometimes
818: make exceptions for this. Our decision will be guided by the two goals
819: of preserving the free status of all derivatives of our free software and
820: of promoting the sharing and reuse of software generally.
821:
822: @iftex
823: @heading NO WARRANTY
824: @end iftex
825: @ifnottex
826: @center NO WARRANTY
827: @end ifnottex
828:
829: @item
830: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
831: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
832: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
833: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
834: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
835: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
836: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
837: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
838: REPAIR OR CORRECTION.
839:
840: @item
841: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
842: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
843: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
844: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
845: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
846: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
847: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
848: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
849: POSSIBILITY OF SUCH DAMAGES.
850: @end enumerate
851:
852: @iftex
853: @heading END OF TERMS AND CONDITIONS
854: @end iftex
855: @ifnottex
856: @center END OF TERMS AND CONDITIONS
857: @end ifnottex
858:
859: @page
860: @unnumberedsec How to Apply These Terms to Your New Programs
861:
862: If you develop a new program, and you want it to be of the greatest
863: possible use to the public, the best way to achieve this is to make it
864: free software which everyone can redistribute and change under these terms.
865:
866: To do so, attach the following notices to the program. It is safest
867: to attach them to the start of each source file to most effectively
868: convey the exclusion of warranty; and each file should have at least
869: the ``copyright'' line and a pointer to where the full notice is found.
870:
871: @smallexample
872: @var{one line to give the program's name and a brief idea of what it does.}
873: Copyright (C) 19@var{yy} @var{name of author}
874:
875: This program is free software; you can redistribute it and/or modify
876: it under the terms of the GNU General Public License as published by
877: the Free Software Foundation; either version 2 of the License, or
878: (at your option) any later version.
879:
880: This program is distributed in the hope that it will be useful,
881: but WITHOUT ANY WARRANTY; without even the implied warranty of
882: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
883: GNU General Public License for more details.
884:
885: You should have received a copy of the GNU General Public License
886: along with this program; if not, write to the Free Software
887: Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
888: @end smallexample
889:
890: Also add information on how to contact you by electronic and paper mail.
891:
892: If the program is interactive, make it output a short notice like this
893: when it starts in an interactive mode:
894:
895: @smallexample
896: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
897: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
898: type `show w'.
899: This is free software, and you are welcome to redistribute it
900: under certain conditions; type `show c' for details.
901: @end smallexample
902:
903: The hypothetical commands @samp{show w} and @samp{show c} should show
904: the appropriate parts of the General Public License. Of course, the
905: commands you use may be called something other than @samp{show w} and
906: @samp{show c}; they could even be mouse-clicks or menu items---whatever
907: suits your program.
908:
909: You should also get your employer (if you work as a programmer) or your
910: school, if any, to sign a ``copyright disclaimer'' for the program, if
911: necessary. Here is a sample; alter the names:
912:
913: @smallexample
914: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
915: `Gnomovision' (which makes passes at compilers) written by James Hacker.
916:
917: @var{signature of Ty Coon}, 1 April 1989
918: Ty Coon, President of Vice
919: @end smallexample
920:
921: This General Public License does not permit incorporating your program into
922: proprietary programs. If your program is a subroutine library, you may
923: consider it more useful to permit linking proprietary applications with the
924: library. If this is what you want to do, use the GNU Library General
925: Public License instead of this License.
926:
927: @iftex
928: @unnumbered Preface
929: @cindex Preface
930: This manual documents Gforth. Some introductory material is provided for
931: readers who are unfamiliar with Forth or who are migrating to Gforth
932: from other Forth compilers. However, this manual is primarily a
933: reference manual.
934: @end iftex
935:
936: @comment TODO much more blurb here.
937:
938: @c ******************************************************************
939: @node Goals, Gforth Environment, License, Top
940: @comment node-name, next, previous, up
941: @chapter Goals of Gforth
942: @cindex goals of the Gforth project
943: The goal of the Gforth Project is to develop a standard model for
944: ANS Forth. This can be split into several subgoals:
945:
946: @itemize @bullet
947: @item
948: Gforth should conform to the ANS Forth Standard.
949: @item
950: It should be a model, i.e. it should define all the
951: implementation-dependent things.
952: @item
953: It should become standard, i.e. widely accepted and used. This goal
954: is the most difficult one.
955: @end itemize
956:
957: To achieve these goals Gforth should be
958: @itemize @bullet
959: @item
960: Similar to previous models (fig-Forth, F83)
961: @item
962: Powerful. It should provide for all the things that are considered
963: necessary today and even some that are not yet considered necessary.
964: @item
965: Efficient. It should not get the reputation of being exceptionally
966: slow.
967: @item
968: Free.
969: @item
970: Available on many machines/easy to port.
971: @end itemize
972:
973: Have we achieved these goals? Gforth conforms to the ANS Forth
974: standard. It may be considered a model, but we have not yet documented
975: which parts of the model are stable and which parts we are likely to
976: change. It certainly has not yet become a de facto standard, but it
977: appears to be quite popular. It has some similarities to and some
978: differences from previous models. It has some powerful features, but not
979: yet everything that we envisioned. We certainly have achieved our
980: execution speed goals (@pxref{Performance}). It is free and available
981: on many machines.
982:
983: @menu
984: * Gforth Extensions Sinful?::
985: @end menu
986:
987: @node Gforth Extensions Sinful?, , Goals, Goals
988: @comment node-name, next, previous, up
989: @section Is it a Sin to use Gforth Extensions?
990: @cindex Gforth extensions
991:
992: If you've been paying attention, you will have realised that there is an
993: ANS (American National Standard) for Forth. As you read through the rest
994: of this manual, you will see documentation for @i{Standard} words, and
995: documentation for some appealing Gforth @i{extensions}. You might ask
996: yourself the question: @i{``Given that there is a standard, would I be
997: committing a sin if I use (non-Standard) Gforth extensions?''}
998:
999: The answer to that question is somewhat pragmatic and somewhat
1000: philosophical. Consider these points:
1001:
1002: @itemize @bullet
1003: @item
1004: A number of the Gforth extensions can be implemented in ANS Forth using
1005: files provided in the @file{compat/} directory. These are mentioned in
1006: the text in passing.
1007: @item
1008: Forth has a rich historical precedent for programmers taking advantage
1009: of implementation-dependent features of their tools (for example,
1010: relying on a knowledge of the dictionary structure). Sometimes these
1011: techniques are necessary to extract every last bit of performance from
1012: the hardware, sometimes they are just a programming shorthand.
1013: @item
1014: The best way to break the rules is to know what the rules are. To learn
1015: the rules, there is no substitute for studying the text of the Standard
1016: itself. In particular, Appendix A of the Standard (@var{Rationale})
1017: provides a valuable insight into the thought processes of the technical
1018: committee.
1019: @item
1020: The best reason to break a rule is because you have to; because it's
1021: more productive to do that, because it makes your code run fast enough
1022: or because you can see no Standard way to achieve what you want to
1023: achieve.
1024: @end itemize
1025:
1026: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
1027: analyse your program and determine what non-Standard definitions it
1028: relies upon.
1029:
1030:
1031: @c ******************************************************************
1032: @node Gforth Environment, Tutorial, Goals, Top
1033: @chapter Gforth Environment
1034: @cindex Gforth environment
1035:
1036: Note: ultimately, the Gforth man page will be auto-generated from the
1037: material in this chapter.
1038:
1039: @menu
1040: * Invoking Gforth:: Getting in
1041: * Leaving Gforth:: Getting out
1042: * Command-line editing::
1043: * Upper and lower case::
1044: * Environment variables:: that affect how Gforth starts up
1045: * Gforth Files:: What gets installed and where
1046: * Startup speed:: When 35ms is not fast enough ...
1047: @end menu
1048:
1049: For related information about the creation of images see @ref{Image Files}.
1050:
1051: @comment ----------------------------------------------
1052: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1053: @section Invoking Gforth
1054: @cindex invoking Gforth
1055: @cindex running Gforth
1056: @cindex command-line options
1057: @cindex options on the command line
1058: @cindex flags on the command line
1059:
1060: Gforth is made up of two parts; an executable ``engine'' (named
1061: @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
1062: will usually just say @code{gforth} -- this automatically loads the
1063: default image file @file{gforth.fi}. In many other cases the default
1064: Gforth image will be invoked like this:
1065: @example
1066: gforth [file | -e forth-code] ...
1067: @end example
1068: @noindent
1069: This interprets the contents of the files and the Forth code in the order they
1070: are given.
1071:
1072: In addition to the @file{gforth} engine, there is also an engine called
1073: @file{gforth-fast}, which is faster, but gives less informative error
1074: messages (@pxref{Error messages}).
1075:
1076: In general, the command line looks like this:
1077:
1078: @example
1079: gforth[-fast] [engine options] [image options]
1080: @end example
1081:
1082: The engine options must come before the rest of the command
1083: line. They are:
1084:
1085: @table @code
1086: @cindex -i, command-line option
1087: @cindex --image-file, command-line option
1088: @item --image-file @i{file}
1089: @itemx -i @i{file}
1090: Loads the Forth image @i{file} instead of the default
1091: @file{gforth.fi} (@pxref{Image Files}).
1092:
1093: @cindex --appl-image, command-line option
1094: @item --appl-image @i{file}
1095: Loads the image @i{file} and leaves all further command-line arguments
1096: to the image (instead of processing them as options). This is useful
1097: for building executable application images on Unix, built with
1098: @code{gforthmi --application ...}.
1099:
1100: @cindex --path, command-line option
1101: @cindex -p, command-line option
1102: @item --path @i{path}
1103: @itemx -p @i{path}
1104: Uses @i{path} for searching the image file and Forth source code files
1105: instead of the default in the environment variable @code{GFORTHPATH} or
1106: the path specified at installation time (e.g.,
1107: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1108: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1109:
1110: @cindex --dictionary-size, command-line option
1111: @cindex -m, command-line option
1112: @cindex @i{size} parameters for command-line options
1113: @cindex size of the dictionary and the stacks
1114: @item --dictionary-size @i{size}
1115: @itemx -m @i{size}
1116: Allocate @i{size} space for the Forth dictionary space instead of
1117: using the default specified in the image (typically 256K). The
1118: @i{size} specification for this and subsequent options consists of
1119: an integer and a unit (e.g.,
1120: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1121: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1122: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1123: @code{e} is used.
1124:
1125: @cindex --data-stack-size, command-line option
1126: @cindex -d, command-line option
1127: @item --data-stack-size @i{size}
1128: @itemx -d @i{size}
1129: Allocate @i{size} space for the data stack instead of using the
1130: default specified in the image (typically 16K).
1131:
1132: @cindex --return-stack-size, command-line option
1133: @cindex -r, command-line option
1134: @item --return-stack-size @i{size}
1135: @itemx -r @i{size}
1136: Allocate @i{size} space for the return stack instead of using the
1137: default specified in the image (typically 15K).
1138:
1139: @cindex --fp-stack-size, command-line option
1140: @cindex -f, command-line option
1141: @item --fp-stack-size @i{size}
1142: @itemx -f @i{size}
1143: Allocate @i{size} space for the floating point stack instead of
1144: using the default specified in the image (typically 15.5K). In this case
1145: the unit specifier @code{e} refers to floating point numbers.
1146:
1147: @cindex --locals-stack-size, command-line option
1148: @cindex -l, command-line option
1149: @item --locals-stack-size @i{size}
1150: @itemx -l @i{size}
1151: Allocate @i{size} space for the locals stack instead of using the
1152: default specified in the image (typically 14.5K).
1153:
1154: @cindex -h, command-line option
1155: @cindex --help, command-line option
1156: @item --help
1157: @itemx -h
1158: Print a message about the command-line options
1159:
1160: @cindex -v, command-line option
1161: @cindex --version, command-line option
1162: @item --version
1163: @itemx -v
1164: Print version and exit
1165:
1166: @cindex --debug, command-line option
1167: @item --debug
1168: Print some information useful for debugging on startup.
1169:
1170: @cindex --offset-image, command-line option
1171: @item --offset-image
1172: Start the dictionary at a slightly different position than would be used
1173: otherwise (useful for creating data-relocatable images,
1174: @pxref{Data-Relocatable Image Files}).
1175:
1176: @cindex --no-offset-im, command-line option
1177: @item --no-offset-im
1178: Start the dictionary at the normal position.
1179:
1180: @cindex --clear-dictionary, command-line option
1181: @item --clear-dictionary
1182: Initialize all bytes in the dictionary to 0 before loading the image
1183: (@pxref{Data-Relocatable Image Files}).
1184:
1185: @cindex --die-on-signal, command-line-option
1186: @item --die-on-signal
1187: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1188: or the segmentation violation SIGSEGV) by translating it into a Forth
1189: @code{THROW}. With this option, Gforth exits if it receives such a
1190: signal. This option is useful when the engine and/or the image might be
1191: severely broken (such that it causes another signal before recovering
1192: from the first); this option avoids endless loops in such cases.
1193: @end table
1194:
1195: @cindex loading files at startup
1196: @cindex executing code on startup
1197: @cindex batch processing with Gforth
1198: As explained above, the image-specific command-line arguments for the
1199: default image @file{gforth.fi} consist of a sequence of filenames and
1200: @code{-e @var{forth-code}} options that are interpreted in the sequence
1201: in which they are given. The @code{-e @var{forth-code}} or
1202: @code{--evaluate @var{forth-code}} option evaluates the Forth
1203: code. This option takes only one argument; if you want to evaluate more
1204: Forth words, you have to quote them or use @code{-e} several times. To exit
1205: after processing the command line (instead of entering interactive mode)
1206: append @code{-e bye} to the command line.
1207:
1208: @cindex versions, invoking other versions of Gforth
1209: If you have several versions of Gforth installed, @code{gforth} will
1210: invoke the version that was installed last. @code{gforth-@i{version}}
1211: invokes a specific version. If your environment contains the variable
1212: @code{GFORTHPATH}, you may want to override it by using the
1213: @code{--path} option.
1214:
1215: Not yet implemented:
1216: On startup the system first executes the system initialization file
1217: (unless the option @code{--no-init-file} is given; note that the system
1218: resulting from using this option may not be ANS Forth conformant). Then
1219: the user initialization file @file{.gforth.fs} is executed, unless the
1220: option @code{--no-rc} is given; this file is first searched in @file{.},
1221: then in @file{~}, then in the normal path (see above).
1222:
1223:
1224:
1225: @comment ----------------------------------------------
1226: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1227: @section Leaving Gforth
1228: @cindex Gforth - leaving
1229: @cindex leaving Gforth
1230:
1231: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1232: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1233: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1234: data are discarded. For ways of saving the state of the system before
1235: leaving Gforth see @ref{Image Files}.
1236:
1237: doc-bye
1238:
1239:
1240: @comment ----------------------------------------------
1241: @node Command-line editing, Upper and lower case, Leaving Gforth, Gforth Environment
1242: @section Command-line editing
1243: @cindex command-line editing
1244:
1245: Gforth maintains a history file that records every line that you type to
1246: the text interpreter. This file is preserved between sessions, and is
1247: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1248: repeatedly you can recall successively older commands from this (or
1249: previous) session(s). The full list of command-line editing facilities is:
1250:
1251: @itemize @bullet
1252: @item
1253: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1254: commands from the history buffer.
1255: @item
1256: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1257: from the history buffer.
1258: @item
1259: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1260: @item
1261: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1262: @item
1263: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1264: closing up the line.
1265: @item
1266: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1267: @item
1268: @kbd{Ctrl-a} to move the cursor to the start of the line.
1269: @item
1270: @kbd{Ctrl-e} to move the cursor to the end of the line.
1271: @item
1272: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1273: line.
1274: @item
1275: @key{TAB} to step through all possible full-word completions of the word
1276: currently being typed.
1277: @item
1278: @kbd{Ctrl-d} at the start of the line to terminate Gforth (gracefully,
1279: using @code{bye}).
1280: @end itemize
1281:
1282: When editing, displayable characters are inserted to the left of the
1283: cursor position; the line is always in ``insert'' (as opposed to
1284: ``overstrike'') mode.
1285:
1286: @cindex history file
1287: @cindex @file{.gforth-history}
1288: On Unix systems, the history file is @file{~/.gforth-history} by
1289: default@footnote{i.e. it is stored in the user's home directory.}. You
1290: can find out the name and location of your history file using:
1291:
1292: @example
1293: history-file type \ Unix-class systems
1294:
1295: history-file type \ Other systems
1296: history-dir type
1297: @end example
1298:
1299: If you enter long definitions by hand, you can use a text editor to
1300: paste them out of the history file into a Forth source file for reuse at
1301: a later time.
1302:
1303: Gforth never trims the size of the history file, so you should do this
1304: periodically, if necessary.
1305:
1306: @comment this is all defined in history.fs
1307: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1308: @comment chosen?
1309:
1310:
1311:
1312: @comment ----------------------------------------------
1313: @node Upper and lower case, Environment variables, Command-line editing, Gforth Environment
1314: @section Upper and lower case
1315: @cindex case-sensitivity
1316: @cindex upper and lower case
1317:
1318: Gforth is case-insensitive; you can enter definitions and invoke
1319: Standard words using upper, lower or mixed case (however,
1320: @pxref{core-idef, Implementation-defined options, Implementation-defined
1321: options}).
1322:
1323: ANS Forth only @i{requires} implementations to recognise Standard words
1324: when they are typed entirely in upper case. Therefore, a Standard
1325: program must use upper case for all Standard words. You can use whatever
1326: case you like for words that you define, but in a standard program you
1327: have to use the words in the same case that you defined them.
1328:
1329: Gforth supports case sensitivity through @code{table}s (case-sensitive
1330: wordlists, @pxref{Word Lists}).
1331:
1332: Two people have asked how to convert Gforth to case sensitivity; while
1333: we think this is a bad idea, you can change all wordlists into tables
1334: like this:
1335:
1336: @example
1337: ' table-find forth-wordlist wordlist-map @ !
1338: @end example
1339:
1340: Note that you now have to type the predefined words in the same case
1341: that we defined them, which are varying. You may want to convert them
1342: to your favourite case before doing this operation (I won't explain how,
1343: because if you are even contemplating to do this, you'd better have
1344: enough knowledge of Forth systems to know this already).
1345:
1346: @comment ----------------------------------------------
1347: @node Environment variables, Gforth Files, Upper and lower case, Gforth Environment
1348: @section Environment variables
1349: @cindex environment variables
1350:
1351: Gforth uses these environment variables:
1352:
1353: @itemize @bullet
1354: @item
1355: @cindex @code{GFORTHHIST} -- environment variable
1356: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1357: open/create the history file, @file{.gforth-history}. Default:
1358: @code{$HOME}.
1359:
1360: @item
1361: @cindex @code{GFORTHPATH} -- environment variable
1362: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1363: for Forth source-code files.
1364:
1365: @item
1366: @cindex @code{GFORTH} -- environment variable
1367: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1368:
1369: @item
1370: @cindex @code{GFORTHD} -- environment variable
1371: @code{GFORTHD} -- used by @file{gforthmi} @xref{gforthmi}.
1372:
1373: @item
1374: @cindex @code{TMP}, @code{TEMP} - environment variable
1375: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1376: location for the history file.
1377: @end itemize
1378:
1379: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1380: @comment mentioning these.
1381:
1382: All the Gforth environment variables default to sensible values if they
1383: are not set.
1384:
1385:
1386: @comment ----------------------------------------------
1387: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1388: @section Gforth files
1389: @cindex Gforth files
1390:
1391: When you install Gforth on a Unix system, it installs files in these
1392: locations by default:
1393:
1394: @itemize @bullet
1395: @item
1396: @file{/usr/local/bin/gforth}
1397: @item
1398: @file{/usr/local/bin/gforthmi}
1399: @item
1400: @file{/usr/local/man/man1/gforth.1} - man page.
1401: @item
1402: @file{/usr/local/info} - the Info version of this manual.
1403: @item
1404: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1405: @item
1406: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1407: @item
1408: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1409: @item
1410: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1411: @end itemize
1412:
1413: You can select different places for installation by using
1414: @code{configure} options (listed with @code{configure --help}).
1415:
1416: @comment ----------------------------------------------
1417: @node Startup speed, , Gforth Files, Gforth Environment
1418: @section Startup speed
1419: @cindex Startup speed
1420: @cindex speed, startup
1421:
1422: If Gforth is used for CGI scripts or in shell scripts, its startup
1423: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1424: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1425: system time.
1426:
1427: If startup speed is a problem, you may consider the following ways to
1428: improve it; or you may consider ways to reduce the number of startups
1429: (e.g., Fast-CGI).
1430:
1431: The first step to improve startup speed is to statically link Gforth, by
1432: building it with @code{XLDFLAGS=-static}. This requires more memory for
1433: the code and will therefore slow down the first invocation, but
1434: subsequent invocations avoid the dynamic linking overhead. Another
1435: disadvantage is that Gforth won't profit from library upgrades. As a
1436: result, @code{gforth-static -e bye} takes about 17.1ms user and
1437: 8.2ms system time.
1438:
1439: The next step to improve startup speed is to use a non-relocatable image
1440: @ref{Non-Relocatable Image Files}. You can create this image with
1441: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1442: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1443: and a part of the copy-on-write overhead. The disadvantage is that the
1444: nonrelocatable image does not work if the OS gives Gforth a different
1445: address for the dictionary, for whatever reason; so you better provide a
1446: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1447: bye} takes about 15.3ms user and 7.5ms system time.
1448:
1449: The final step is to disable dictionary hashing in Gforth. Gforth
1450: builds the hash table on startup, which takes much of the startup
1451: overhead. You can do this by commenting out the @code{include hash.fs}
1452: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1453: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1454: The disadvantages are that functionality like @code{table} and
1455: @code{ekey} is missing and that text interpretation (e.g., compiling)
1456: now takes much longer. So, you should only use this method if there is
1457: no significant text interpretation to perform (the script should be
1458: compiled into the image, among other things). @code{gforth-static -i
1459: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1460:
1461: @c ******************************************************************
1462: @node Tutorial, Introduction, Gforth Environment, Top
1463: @chapter Forth Tutorial
1464: @cindex Tutorial
1465: @cindex Forth Tutorial
1466:
1467: This tutorial can be used with any ANS-compliant Forth; any places that
1468: mention features specific to Gforth are marked as such and you can skip
1469: them, if you work with another Forth. This tutorial does not explain
1470: all features of Forth, just enough to get you started and give you some
1471: ideas about the facilities available in Forth. Read the rest of the
1472: manual and the standard when you are through this.
1473:
1474: The intended way to use this tutorial is that you work through it while
1475: sitting in front of the console, take a look at the examples and predict
1476: what they will do, then try them out; if the outcome is not as expected,
1477: find out why (e.g., by trying out variations of the example), so you
1478: understand what's going on. There are also some assignments that you
1479: should solve.
1480:
1481: This tutorial assumes that you have programmed before and know what,
1482: e.g., a loop is.
1483:
1484: @c !! explain compat library
1485:
1486: @menu
1487: * Starting Gforth Tutorial::
1488: * Syntax Tutorial::
1489: * Crash Course Tutorial::
1490: * Stack Tutorial::
1491: * Arithmetics Tutorial::
1492: * Stack Manipulation Tutorial::
1493: * Using files for Forth code Tutorial::
1494: * Comments Tutorial::
1495: * Colon Definitions Tutorial::
1496: * Decompilation Tutorial::
1497: * Stack-Effect Comments Tutorial::
1498: * Types Tutorial::
1499: * Factoring Tutorial::
1500: * Designing the stack effect Tutorial::
1501: * Local Variables Tutorial::
1502: * Conditional execution Tutorial::
1503: * Flags and Comparisons Tutorial::
1504: * General Loops Tutorial::
1505: * Counted loops Tutorial::
1506: * Recursion Tutorial::
1507: * Leaving definitions or loops Tutorial::
1508: * Return Stack Tutorial::
1509: * Memory Tutorial::
1510: * Characters and Strings Tutorial::
1511: * Alignment Tutorial::
1512: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1513: * Execution Tokens Tutorial::
1514: * Exceptions Tutorial::
1515: * Defining Words Tutorial::
1516: * Arrays and Records Tutorial::
1517: * POSTPONE Tutorial::
1518: * Literal Tutorial::
1519: * Advanced macros Tutorial::
1520: * Compilation Tokens Tutorial::
1521: * Wordlists and Search Order Tutorial::
1522: @end menu
1523:
1524: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1525: @section Starting Gforth
1526:
1527: You can start Gforth by typing its name:
1528:
1529: @example
1530: gforth
1531: @end example
1532:
1533: That puts you into interactive mode; you can leave Gforth by typing
1534: @code{bye}. While in Gforth, you can edit the command line and access
1535: the command line history with cursor keys, similar to bash.
1536:
1537:
1538: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1539: @section Syntax
1540:
1541: A @dfn{word} is a sequence of arbitrary characters (expcept white
1542: space). Words are separated by white space. E.g., each of the
1543: following lines contains exactly one word:
1544:
1545: @example
1546: word
1547: !@@#$%^&*()
1548: 1234567890
1549: 5!a
1550: @end example
1551:
1552: A frequent beginner's error is to leave away necessary white space,
1553: resulting in an error like @samp{Undefined word}; so if you see such an
1554: error, check if you have put spaces wherever necessary.
1555:
1556: @example
1557: ." hello, world" \ correct
1558: ."hello, world" \ gives an "Undefined word" error
1559: @end example
1560:
1561: Gforth and most other Forth systems ignores differences in case (it is
1562: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1563: your system is case-sensitive, you may have to type all the examples
1564: given here in upper case.
1565:
1566:
1567: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1568: @section Crash Course
1569:
1570: Type
1571:
1572: @example
1573: 0 0 !
1574: here execute
1575: ' catch >body 20 erase abort
1576: ' (quit) >body 20 erase
1577: @end example
1578:
1579: The last two examples are guaranteed to destroy parts of Gforth (and
1580: most other systems), so you better leave Gforth afterwards (if it has
1581: not finished by itself). On some systems you may have to kill gforth
1582: from outside (e.g., in Unix with @code{kill}).
1583:
1584: Now that you know how to produce crashes (and that there's not much to
1585: them), let's learn how to produce meaningful programs.
1586:
1587:
1588: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1589: @section Stack
1590:
1591: The most obvious feature of Forth is the stack. When you type in a
1592: number, it is pushed on the stack. You can display the content of the
1593: stack with @code{.s}.
1594:
1595: @example
1596: 1 2 .s
1597: 3 .s
1598: @end example
1599:
1600: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1601: appear in @code{.s} output as they appeared in the input.
1602:
1603: You can print the top of stack element with @code{.}.
1604:
1605: @example
1606: 1 2 3 . . .
1607: @end example
1608:
1609: In general, words consume their stack arguments (@code{.s} is an
1610: exception).
1611:
1612: @assignment
1613: What does the stack contain after @code{5 6 7 .}?
1614: @endassignment
1615:
1616:
1617: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1618: @section Arithmetics
1619:
1620: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1621: operate on the top two stack items:
1622:
1623: @example
1624: 2 2 + .
1625: 2 1 - .
1626: 7 3 mod .
1627: @end example
1628:
1629: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1630: as in the corresponding infix expression (this is generally the case in
1631: Forth).
1632:
1633: Parentheses are superfluous (and not available), because the order of
1634: the words unambiguously determines the order of evaluation and the
1635: operands:
1636:
1637: @example
1638: 3 4 + 5 * .
1639: 3 4 5 * + .
1640: @end example
1641:
1642: @assignment
1643: What are the infix expressions corresponding to the Forth code above?
1644: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1645: known as Postfix or RPN (Reverse Polish Notation).}.
1646: @endassignment
1647:
1648: To change the sign, use @code{negate}:
1649:
1650: @example
1651: 2 negate .
1652: @end example
1653:
1654: @assignment
1655: Convert -(-3)*4-5 to Forth.
1656: @endassignment
1657:
1658: @code{/mod} performs both @code{/} and @code{mod}.
1659:
1660: @example
1661: 7 3 /mod . .
1662: @end example
1663:
1664: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1665: @section Stack Manipulation
1666:
1667: Stack manipulation words rearrange the data on the stack.
1668:
1669: @example
1670: 1 .s drop .s
1671: 1 .s dup .s drop drop .s
1672: 1 2 .s over .s drop drop drop
1673: 1 2 .s swap .s drop drop
1674: 1 2 3 .s rot .s drop drop drop
1675: @end example
1676:
1677: These are the most important stack manipulation words. There are also
1678: variants that manipulate twice as many stack items:
1679:
1680: @example
1681: 1 2 3 4 .s 2swap .s 2drop 2drop
1682: @end example
1683:
1684: Two more stack manipulation words are:
1685:
1686: @example
1687: 1 2 .s nip .s drop
1688: 1 2 .s tuck .s 2drop drop
1689: @end example
1690:
1691: @assignment
1692: Replace @code{nip} and @code{tuck} with combinations of other stack
1693: manipulation words.
1694:
1695: @example
1696: Given: How do you get:
1697: 1 2 3 3 2 1
1698: 1 2 3 1 2 3 2
1699: 1 2 3 1 2 3 3
1700: 1 2 3 1 3 3
1701: 1 2 3 2 1 3
1702: 1 2 3 4 4 3 2 1
1703: 1 2 3 1 2 3 1 2 3
1704: 1 2 3 4 1 2 3 4 1 2
1705: 1 2 3
1706: 1 2 3 1 2 3 4
1707: 1 2 3 1 3
1708: @end example
1709: @endassignment
1710:
1711: @example
1712: 5 dup * .
1713: @end example
1714:
1715: @assignment
1716: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1717: Write a piece of Forth code that expects two numbers on the stack
1718: (@var{a} and @var{b}, with @var{b} on top) and computes
1719: @code{(a-b)(a+1)}.
1720: @endassignment
1721:
1722: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1723: @section Using files for Forth code
1724:
1725: While working at the Forth command line is convenient for one-line
1726: examples and short one-off code, you probably want to store your source
1727: code in files for convenient editing and persistence. You can use your
1728: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1729: Gforth}) to create @var{file} and use
1730:
1731: @example
1732: s" @var{file}" included
1733: @end example
1734:
1735: to load it into your Forth system. The file name extension I use for
1736: Forth files is @samp{.fs}.
1737:
1738: You can easily start Gforth with some files loaded like this:
1739:
1740: @example
1741: gforth @var{file1} @var{file2}
1742: @end example
1743:
1744: If an error occurs during loading these files, Gforth terminates,
1745: whereas an error during @code{INCLUDED} within Gforth usually gives you
1746: a Gforth command line. Starting the Forth system every time gives you a
1747: clean start every time, without interference from the results of earlier
1748: tries.
1749:
1750: I often put all the tests in a file, then load the code and run the
1751: tests with
1752:
1753: @example
1754: gforth @var{code} @var{tests} -e bye
1755: @end example
1756:
1757: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1758: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1759: restart this command without ado.
1760:
1761: The advantage of this approach is that the tests can be repeated easily
1762: every time the program ist changed, making it easy to catch bugs
1763: introduced by the change.
1764:
1765:
1766: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1767: @section Comments
1768:
1769: @example
1770: \ That's a comment; it ends at the end of the line
1771: ( Another comment; it ends here: ) .s
1772: @end example
1773:
1774: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1775: separated with white space from the following text.
1776:
1777: @example
1778: \This gives an "Undefined word" error
1779: @end example
1780:
1781: The first @code{)} ends a comment started with @code{(}, so you cannot
1782: nest @code{(}-comments; and you cannot comment out text containing a
1783: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1784: avoid @code{)} in word names.}.
1785:
1786: I use @code{\}-comments for descriptive text and for commenting out code
1787: of one or more line; I use @code{(}-comments for describing the stack
1788: effect, the stack contents, or for commenting out sub-line pieces of
1789: code.
1790:
1791: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1792: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1793: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1794: with @kbd{M-q}.
1795:
1796:
1797: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1798: @section Colon Definitions
1799:
1800: are similar to procedures and functions in other programming languages.
1801:
1802: @example
1803: : squared ( n -- n^2 )
1804: dup * ;
1805: 5 squared .
1806: 7 squared .
1807: @end example
1808:
1809: @code{:} starts the colon definition; its name is @code{squared}. The
1810: following comment describes its stack effect. The words @code{dup *}
1811: are not executed, but compiled into the definition. @code{;} ends the
1812: colon definition.
1813:
1814: The newly-defined word can be used like any other word, including using
1815: it in other definitions:
1816:
1817: @example
1818: : cubed ( n -- n^3 )
1819: dup squared * ;
1820: -5 cubed .
1821: : fourth-power ( n -- n^4 )
1822: squared squared ;
1823: 3 fourth-power .
1824: @end example
1825:
1826: @assignment
1827: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1828: @code{/mod} in terms of other Forth words, and check if they work (hint:
1829: test your tests on the originals first). Don't let the
1830: @samp{redefined}-Messages spook you, they are just warnings.
1831: @endassignment
1832:
1833:
1834: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1835: @section Decompilation
1836:
1837: You can decompile colon definitions with @code{see}:
1838:
1839: @example
1840: see squared
1841: see cubed
1842: @end example
1843:
1844: In Gforth @code{see} shows you a reconstruction of the source code from
1845: the executable code. Informations that were present in the source, but
1846: not in the executable code, are lost (e.g., comments).
1847:
1848: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1849: @section Stack-Effect Comments
1850:
1851: By convention the comment after the name of a definition describes the
1852: stack effect: The part in from of the @samp{--} describes the state of
1853: the stack before the execution of the definition, i.e., the parameters
1854: that are passed into the colon definition; the part behind the @samp{--}
1855: is the state of the stack after the execution of the definition, i.e.,
1856: the results of the definition. The stack comment only shows the top
1857: stack items that the definition accesses and/or changes.
1858:
1859: You should put a correct stack effect on every definition, even if it is
1860: just @code{( -- )}. You should also add some descriptive comment to
1861: more complicated words (I usually do this in the lines following
1862: @code{:}). If you don't do this, your code becomes unreadable (because
1863: you have to work through every definition before you can undertsand
1864: any).
1865:
1866: @assignment
1867: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1868: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1869: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1870: are done, you can compare your stack effects to this in this manual
1871: (@pxref{Word Index}).
1872: @endassignment
1873:
1874: Sometimes programmers put comments at various places in colon
1875: definitions that describe the contents of the stack at that place (stack
1876: comments); i.e., they are like the first part of a stack-effect
1877: comment. E.g.,
1878:
1879: @example
1880: : cubed ( n -- n^3 )
1881: dup squared ( n n^2 ) * ;
1882: @end example
1883:
1884: In this case the stack comment is pretty superfluous, because the word
1885: is simple enough. If you think it would be a good idea to add such a
1886: comment to increase readability, you should also consider factoring the
1887: word into several simpler words (@pxref{Factoring Tutorial,,
1888: Factoring}), which typically eliminates the need for the stack effect;
1889: however, if you decide not to refactor it, then having such a comment is
1890: better than not having it.
1891:
1892: The names of the stack items in stack-effect and stack comments in the
1893: standard, in this manual, and in many programs specify the type through
1894: a type prefix, similar to Fortran and Hungarian notation. The most
1895: frequent prefixes are:
1896:
1897: @table @code
1898: @item n
1899: signed integer
1900: @item u
1901: unsigned integer
1902: @item c
1903: character
1904: @item f
1905: Boolean flags, i.e. @code{false} or @code{true}.
1906: @item a-addr,a-
1907: Cell-aligned address
1908: @item c-addr,c-
1909: Char-aligned address (note that a Char may have two bytes in Windows NT)
1910: @item xt
1911: Execution token, same size as Cell
1912: @item w,x
1913: Cell, can contain an integer or an address. It usually takes 32, 64 or
1914: 16 bits (depending on your platform and Forth system). A cell is more
1915: commonly known as machine word, but the term @emph{word} already means
1916: something different in Forth.
1917: @item d
1918: signed double-cell integer
1919: @item ud
1920: unsigned double-cell integer
1921: @item r
1922: Float (on the FP stack)
1923: @end table
1924:
1925: You can find a more complete list in @ref{Notation}.
1926:
1927: @assignment
1928: Write stack-effect comments for all definitions you have written up to
1929: now.
1930: @endassignment
1931:
1932:
1933: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1934: @section Types
1935:
1936: In Forth the names of the operations are not overloaded; so similar
1937: operations on different types need different names; e.g., @code{+} adds
1938: integers, and you have to use @code{f+} to add floating-point numbers.
1939: The following prefixes are often used for related operations on
1940: different types:
1941:
1942: @table @code
1943: @item (none)
1944: signed integer
1945: @item u
1946: unsigned integer
1947: @item c
1948: character
1949: @item d
1950: signed double-cell integer
1951: @item ud, du
1952: unsigned double-cell integer
1953: @item 2
1954: two cells (not-necessarily double-cell numbers)
1955: @item m, um
1956: mixed single-cell and double-cell operations
1957: @item f
1958: floating-point (note that in stack comments @samp{f} represents flags,
1959: and @samp{r} represents FP number).
1960: @end table
1961:
1962: If there are no differences between the signed and the unsigned variant
1963: (e.g., for @code{+}), there is only the prefix-less variant.
1964:
1965: Forth does not perform type checking, neither at compile time, nor at
1966: run time. If you use the wrong oeration, the data are interpreted
1967: incorrectly:
1968:
1969: @example
1970: -1 u.
1971: @end example
1972:
1973: If you have only experience with type-checked languages until now, and
1974: have heard how important type-checking is, don't panic! In my
1975: experience (and that of other Forthers), type errors in Forth code are
1976: usually easy to find (once you get used to it), the increased vigilance
1977: of the programmer tends to catch some harder errors in addition to most
1978: type errors, and you never have to work around the type system, so in
1979: most situations the lack of type-checking seems to be a win (projects to
1980: add type checking to Forth have not caught on).
1981:
1982:
1983: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1984: @section Factoring
1985:
1986: If you try to write longer definitions, you will soon find it hard to
1987: keep track of the stack contents. Therefore, good Forth programmers
1988: tend to write only short definitions (e.g., three lines). The art of
1989: finding meaningful short definitions is known as factoring (as in
1990: factoring polynomials).
1991:
1992: Well-factored programs offer additional advantages: smaller, more
1993: general words, are easier to test and debug and can be reused more and
1994: better than larger, specialized words.
1995:
1996: So, if you run into difficulties with stack management, when writing
1997: code, try to define meaningful factors for the word, and define the word
1998: in terms of those. Even if a factor contains only two words, it is
1999: often helpful.
2000:
2001: Good factoring is not easy, and even experienced Forth programmers often
2002: don't find the right solution right away, but only when rewriting the
2003: program. So, if you don't come up with a good solution immediately,
2004: keep trying, don't despair.
2005:
2006: @c example !!
2007:
2008:
2009: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
2010: @section Designing the stack effect
2011:
2012: In other languages you can use an arbitrary order of parameters for a
2013: function; and since ther is only one result, you don't have to deal with
2014: the order of results, either.
2015:
2016: In Forth (and other stack-based languages, e.g., Postscript) the
2017: parameter and result order of a definition is important and should be
2018: designed well. The general guideline is to design the stack effect such
2019: that the word is simple to use in most cases, even if that complicates
2020: the implementation of the word. Some concrete rules are:
2021:
2022: @itemize @bullet
2023:
2024: @item
2025: Words consume all of their parameters (e.g., @code{.}).
2026:
2027: @item
2028: If there is a convention on the order of parameters (e.g., from
2029: mathematics or another programming language), stick with it (e.g.,
2030: @code{-}).
2031:
2032: @item
2033: If one parameter usually requires only a short computation (e.g., it is
2034: a constant), pass it on the top of the stack. Conversely, parameters
2035: that usually require a long sequence of code to compute should be passed
2036: as the bottom (i.e., first) parameter. This makes the code easier to
2037: read, because reader does not need to keep track of the bottom item
2038: through a long sequence of code (or, alternatively, through stack
2039: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
2040: address on top of the stack because it is usually simpler to compute
2041: than the stored value (often the address is just a variable).
2042:
2043: @item
2044: Similarly, results that are usually consumed quickly should be returned
2045: on the top of stack, whereas a result that is often used in long
2046: computations should be passed as bottom result. E.g., the file words
2047: like @code{open-file} return the error code on the top of stack, because
2048: it is usually consumed quickly by @code{throw}; moreover, the error code
2049: has to be checked before doing anything with the other results.
2050:
2051: @end itemize
2052:
2053: These rules are just general guidelines, don't lose sight of the overall
2054: goal to make the words easy to use. E.g., if the convention rule
2055: conflicts with the computation-length rule, you might decide in favour
2056: of the convention if the word will be used rarely, and in favour of the
2057: computation-length rule if the word will be used frequently (because
2058: with frequent use the cost of breaking the computation-length rule would
2059: be quite high, and frequent use makes it easier to remember an
2060: unconventional order).
2061:
2062: @c example !! structure package
2063:
2064: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2065: @section Local Variables
2066:
2067: You can define local variables (@emph{locals}) in a colon definition:
2068:
2069: @example
2070: : swap @{ a b -- b a @}
2071: b a ;
2072: 1 2 swap .s 2drop
2073: @end example
2074:
2075: (If your Forth system does not support this syntax, include
2076: @file{compat/anslocals.fs} first).
2077:
2078: In this example @code{@{ a b -- b a @}} is the locals definition; it
2079: takes two cells from the stack, puts the top of stack in @code{b} and
2080: the next stack element in @code{a}. @code{--} starts a comment ending
2081: with @code{@}}. After the locals definition, using the name of the
2082: local will push its value on the stack. You can leave the comment
2083: part (@code{-- b a}) away:
2084:
2085: @example
2086: : swap ( x1 x2 -- x2 x1 )
2087: @{ a b @} b a ;
2088: @end example
2089:
2090: In Gforth you can have several locals definitions, anywhere in a colon
2091: definition; in contrast, in a standard program you can have only one
2092: locals definition per colon definition, and that locals definition must
2093: be outside any controll structure.
2094:
2095: With locals you can write slightly longer definitions without running
2096: into stack trouble. However, I recommend trying to write colon
2097: definitions without locals for exercise purposes to help you gain the
2098: essential factoring skills.
2099:
2100: @assignment
2101: Rewrite your definitions until now with locals
2102: @endassignment
2103:
2104:
2105: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2106: @section Conditional execution
2107:
2108: In Forth you can use control structures only inside colon definitions.
2109: An @code{if}-structure looks like this:
2110:
2111: @example
2112: : abs ( n1 -- +n2 )
2113: dup 0 < if
2114: negate
2115: endif ;
2116: 5 abs .
2117: -5 abs .
2118: @end example
2119:
2120: @code{if} takes a flag from the stack. If the flag is non-zero (true),
2121: the following code is performed, otherwise execution continues after the
2122: @code{endif} (or @code{else}). @code{<} compares the top two stack
2123: elements and prioduces a flag:
2124:
2125: @example
2126: 1 2 < .
2127: 2 1 < .
2128: 1 1 < .
2129: @end example
2130:
2131: Actually the standard name for @code{endif} is @code{then}. This
2132: tutorial presents the examples using @code{endif}, because this is often
2133: less confusing for people familiar with other programming languages
2134: where @code{then} has a different meaning. If your system does not have
2135: @code{endif}, define it with
2136:
2137: @example
2138: : endif postpone then ; immediate
2139: @end example
2140:
2141: You can optionally use an @code{else}-part:
2142:
2143: @example
2144: : min ( n1 n2 -- n )
2145: 2dup < if
2146: drop
2147: else
2148: nip
2149: endif ;
2150: 2 3 min .
2151: 3 2 min .
2152: @end example
2153:
2154: @assignment
2155: Write @code{min} without @code{else}-part (hint: what's the definition
2156: of @code{nip}?).
2157: @endassignment
2158:
2159:
2160: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2161: @section Flags and Comparisons
2162:
2163: In a false-flag all bits are clear (0 when interpreted as integer). In
2164: a canonical true-flag all bits are set (-1 as a twos-complement signed
2165: integer); in many contexts (e.g., @code{if}) any non-zero value is
2166: treated as true flag.
2167:
2168: @example
2169: false .
2170: true .
2171: true hex u. decimal
2172: @end example
2173:
2174: Comparison words produce canonical flags:
2175:
2176: @example
2177: 1 1 = .
2178: 1 0= .
2179: 0 1 < .
2180: 0 0 < .
2181: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2182: -1 1 < .
2183: @end example
2184:
2185: Gforth supports all combinations of the prefixes @code{0 u d d0 du} (or
2186: none) and the comparisons @code{= <> < > <= >=}. Only a part of these
2187: combinations are standard (see the standard or !! the glossary for
2188: details).
2189:
2190: You can use @code{and or xor invert} can be used as operations on
2191: canonical flags. Actually they are bitwise operations:
2192:
2193: @example
2194: 1 2 and .
2195: 1 2 or .
2196: 1 3 xor .
2197: 1 invert .
2198: @end example
2199:
2200: You can convert a zero/non-zero flag into a canonical flag with
2201: @code{0<>} (and complement it on the way with @code{0=}).
2202:
2203: @example
2204: 1 0= .
2205: 1 0<> .
2206: @end example
2207:
2208: You can use the all-bits-set feature of canonicasl flags and the bitwise
2209: operation of the Boolean operations to avoid @code{if}s:
2210:
2211: @example
2212: : foo ( n1 -- n2 )
2213: 0= if
2214: 14
2215: else
2216: 0
2217: endif ;
2218: 0 foo .
2219: 1 foo .
2220:
2221: : foo ( n1 -- n2 )
2222: 0= 14 and ;
2223: 0 foo .
2224: 1 foo .
2225: @end example
2226:
2227: @assignment
2228: Write @code{min} without @code{if}.
2229: @endassignment
2230:
2231:
2232: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2233: @section General Loops
2234:
2235: The endless loop is the most simple one:
2236:
2237: @example
2238: : endless ( -- )
2239: 0 begin
2240: dup . 1+
2241: again ;
2242: endless
2243: @end example
2244:
2245: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2246: does nothing at run-time, @code{again} jumps back to @code{begin}.
2247:
2248: A loop with one exit at any place looks like this:
2249:
2250: @example
2251: : log2 ( +n1 -- n2 )
2252: \ logarithmus dualis of n1>0, rounded down to the next integer
2253: assert( dup 0> )
2254: 2/ 0 begin
2255: over 0> while
2256: 1+ swap 2/ swap
2257: repeat
2258: nip ;
2259: 7 log2 .
2260: 8 log2 .
2261: @end example
2262:
2263: At run-time @code{while} consumes a flag; if it is 0, execution
2264: continues behind the @code{repeat}; if the flag is non-zero, execution
2265: continues behind the @code{while}. @code{Repeat} jumps back to
2266: @code{begin}, just like @code{again}.
2267:
2268: In Forth there are many combinations/abbreviations, like @code{1+}.
2269: However, @code{2/} is not one of them; it shifts it's argument right by
2270: one bit (arithmetic shift right):
2271:
2272: @example
2273: -5 2 / .
2274: -5 2/ .
2275: @end example
2276:
2277: @code{assert(} is no standard word, but you can get it on systems other
2278: then Gforth by including @file{compat/assert.fs}. You can see what it
2279: does by trying
2280:
2281: @example
2282: 0 log2 .
2283: @end example
2284:
2285: Here's a loop with an exit at the end:
2286:
2287: @example
2288: : log2 ( +n1 -- n2 )
2289: \ logarithmus dualis of n1>0, rounded down to the next integer
2290: assert( dup 0 > )
2291: -1 begin
2292: 1+ swap 2/ swap
2293: over 0 <=
2294: until
2295: nip ;
2296: @end example
2297:
2298: @code{Until} consumes a flag; if it is non-zero, execution continues at
2299: the @code{begin}, otherwise after the @code{until}.
2300:
2301: @assignment
2302: Write a definition for computing the greatest common divisor.
2303: @endassignment
2304:
2305:
2306: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2307: @section Counted loops
2308:
2309: @example
2310: : ^ ( n1 u -- n )
2311: \ n = the uth power of u1
2312: 1 swap 0 u+do
2313: over *
2314: loop
2315: nip ;
2316: 3 2 ^ .
2317: 4 3 ^ .
2318: @end example
2319:
2320: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2321: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2322: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2323: times (or not at all, if @code{u3-u4<0}).
2324:
2325: You can see the stack effect design rules at work in the stack effect of
2326: the loop start words: Since the start value of the loop is more
2327: frequently constant than the end value, the start value is passed on
2328: the top-of-stack.
2329:
2330: You can access the counter of a counted loop with @code{i}:
2331:
2332: @example
2333: : fac ( u -- u! )
2334: 1 swap 1+ 1 u+do
2335: i *
2336: loop ;
2337: 5 fac .
2338: 7 fac .
2339: @end example
2340:
2341: There is also @code{+do}, which expects signed numbers (important for
2342: deciding whether to enter the loop).
2343:
2344: @assignment
2345: Write a definition for computing the nth Fibonacci number.
2346: @endassignment
2347:
2348: !! +DO...+LOOP
2349: !! -DO...-LOOP
2350:
2351:
2352: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2353: @section Recursion
2354:
2355: Usually the name of a definition is not visible in the definition; but
2356: earlier definitions are usually visible:
2357:
2358: @example
2359: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2360: : / ( n1 n2 -- n )
2361: dup 0= if
2362: -10 throw \ report division by zero
2363: endif
2364: / \ old version
2365: ;
2366: 1 0 /
2367: @end example
2368:
2369: For recursive definitions you can use @code{recursive} (non-standard) or
2370: @code{recurse}:
2371:
2372: @example
2373: : fac1 ( n -- n! ) recursive
2374: dup 0> if
2375: dup 1- fac1 *
2376: else
2377: drop 1
2378: endif ;
2379: 7 fac1 .
2380:
2381: : fac2 ( n -- n! )
2382: dup 0> if
2383: dup 1- recurse *
2384: else
2385: drop 1
2386: endif ;
2387: 8 fac2 .
2388: @end example
2389:
2390: @assignment
2391: Write a recursive definition for computing the nth Fibonacci number.
2392: @endassignment
2393:
2394:
2395: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2396: @section Leaving definitions or loops
2397:
2398: @code{EXIT} exits the current definition right away. For every counted
2399: loop that is left in this way, an @code{UNLOOP} has to be performed
2400: before the @code{EXIT}:
2401:
2402: @c !! real examples
2403: @example
2404: : ...
2405: ... u+do
2406: ... if
2407: ... unloop exit
2408: endif
2409: ...
2410: loop
2411: ... ;
2412: @end example
2413:
2414: @code{LEAVE} leaves the innermost counted loop right away:
2415:
2416: @example
2417: : ...
2418: ... u+do
2419: ... if
2420: ... leave
2421: endif
2422: ...
2423: loop
2424: ... ;
2425: @end example
2426:
2427:
2428: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2429: @section Return Stack
2430:
2431: In addition to the data stack Forth also has a second stack, the return
2432: stack; most Forth systems store the return addresses of procedure calls
2433: there (thus its name). Programmers can also use this stack:
2434:
2435: @example
2436: : foo ( n1 n2 -- )
2437: .s
2438: >r .s
2439: r@@ .
2440: >r .s
2441: r@@ .
2442: r> .
2443: r@@ .
2444: r> . ;
2445: 1 2 foo
2446: @end example
2447:
2448: @code{>r} takes an element from the data stack and pushes it onto the
2449: return stack; conversely, @code{r>} moves an elementm from the return to
2450: the data stack; @code{r@@} pushes a copy of the top of the return stack
2451: on the return stack.
2452:
2453: Forth programmers usually use the return stack for storing data
2454: temporarily, if using the data stack alone would be too complex, and
2455: factoring and locals are not an option:
2456:
2457: @example
2458: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2459: rot >r rot r> ;
2460: @end example
2461:
2462: The return address of the definition and the loop control parameters of
2463: counted loops usually reside on the return stack, so you have to take
2464: all items, that you have pushed on the return stack in a colon
2465: definition or counted loop, from the return stack before the definition
2466: or loop ends. You cannot access items that you pushed on the return
2467: stack outside some definition or loop within the definition of loop.
2468:
2469: If you miscount the return stack items, this usually ends in a crash:
2470:
2471: @example
2472: : crash ( n -- )
2473: >r ;
2474: 5 crash
2475: @end example
2476:
2477: You cannot mix using locals and using the return stack (according to the
2478: standard; Gforth has no problem). However, they solve the same
2479: problems, so this shouldn't be an issue.
2480:
2481: @assignment
2482: Can you rewrite any of the definitions you wrote until now in a better
2483: way using the return stack?
2484: @endassignment
2485:
2486:
2487: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2488: @section Memory
2489:
2490: You can create a global variable @code{v} with
2491:
2492: @example
2493: variable v ( -- addr )
2494: @end example
2495:
2496: @code{v} pushes the address of a cell in memory on the stack. This cell
2497: was reserved by @code{variable}. You can use @code{!} (store) to store
2498: values into this cell and @code{@@} (fetch) to load the value from the
2499: stack into memory:
2500:
2501: @example
2502: v .
2503: 5 v ! .s
2504: v @@ .
2505: @end example
2506:
2507: You can also reserve more memory:
2508:
2509: @example
2510: create v2 20 cells allot
2511: @end example
2512:
2513: creates a word @code{v2} and reserves 20 cells; the address pushed by
2514: @code{v2} points to the start of these 20 cells. You can use address
2515: arithmetic to access these cells:
2516:
2517: @example
2518: 3 v2 5 cells + !
2519: @end example
2520:
2521: You can reserve and initialize memory with @code{,}:
2522:
2523: @example
2524: create v3
2525: 5 , 4 , 3 , 2 , 1 ,
2526: v3 @@ .
2527: v3 cell+ @@ .
2528: v3 2 cells + @@ .
2529: @end example
2530:
2531: @assignment
2532: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2533: @code{u} cells, with the first of these cells at @code{addr}, the next
2534: one at @code{addr cell+} etc.
2535: @endassignment
2536:
2537: You can also reserve memory without creating a new word:
2538:
2539: @example
2540: here 10 cells allot
2541: .s
2542: @end example
2543:
2544: @code{Here} pushes the start address of the memory area. You should
2545: store it somewhere, or you will have a hard time finding the memory area
2546: again.
2547:
2548: @code{Allot} manages dictionary memory. The dictionary memory contains
2549: the system's data structures for words etc. on Gforth and most other
2550: Forth systems. It is managed like a stack: You can free the memory that
2551: you have just @code{allot}ed with
2552:
2553: @example
2554: -10 cells allot
2555: @end example
2556:
2557: Note that you cannot do this if you have created a new word in the
2558: meantime (because then your @code{allot}ed memory is no longer on the
2559: top of the dictionary ``stack'').
2560:
2561: Alternatively, you can use @code{allocate} and @code{free} which allow
2562: freeing memory in any order:
2563:
2564: @example
2565: 10 cells allocate throw .s
2566: 20 cells allocate throw .s
2567: swap
2568: free throw
2569: free throw
2570: @end example
2571:
2572: The @code{throw}s deal with errors (e.g., out of memory).
2573:
2574: And there is also a garbage collector @url{!!}, which eliminates the
2575: need to @code{free} memory explicitly.
2576:
2577:
2578: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2579: @section Characters and Strings
2580:
2581: On the stack characters take up a cell, like numbers. In memory they
2582: have their own size (one 8-bit byte on most systems), and therefore
2583: require their own words for memory access:
2584:
2585: @example
2586: create v4
2587: 104 c, 97 c, 108 c, 108 c, 111 c,
2588: v4 4 chars + c@@ .
2589: @end example
2590:
2591: The preferred representation of strings on the stack is @code{addr
2592: u-count}, where @code{addr} is the address of the first character and
2593: @code{u-count} is the number of characters in the string.
2594:
2595: @example
2596: v4 5 type
2597: @end example
2598:
2599: You get a string constant with
2600:
2601: @example
2602: s" hello, world" .s
2603: type
2604: @end example
2605:
2606: Make sure you have a space between @code{s"} and the string; @code{s"}
2607: is a normal Forth word and must be delimited with white space (try what
2608: happens when you remove the space).
2609:
2610: However, this interpretive use of @code{s"} is quite restricted: the
2611: string exists only until the next call of @code{s"} (some Forth systems
2612: keep more than one of these strings, but usually they still have a
2613: limited lifetime.
2614:
2615: @example
2616: s" hello," s" world" .s
2617: type
2618: type
2619: @end example
2620:
2621: However, you can also use @code{s"} in a definition, and the resulting
2622: strings then live forever (well, as long as the definition):
2623:
2624: @example
2625: : foo s" hello," s" world" ;
2626: foo .s
2627: type
2628: type
2629: @end example
2630:
2631: @assignment
2632: @code{Emit ( c -- )} types @code{c} as character (not a number).
2633: Implement @code{type ( addr u -- )}.
2634: @endassignment
2635:
2636: @node Alignment Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Characters and Strings Tutorial, Tutorial
2637: @section Alignment
2638:
2639: On many processors cells have to be aligned in memory, if you want to
2640: access them with @code{@@} and @code{!} (and even if the processor does
2641: not require alignment, access to aligned cells are faster).
2642:
2643: @code{Create} aligns @code{here} (i.e., the place where the next
2644: allocation will occur, and that the @code{create}d word points to).
2645: Likewise, the memory produced by @code{allocate} starts at an aligned
2646: address. Adding a number of @code{cells} to an aligned address produces
2647: another aligned address.
2648:
2649: However, address arithmetic involving @code{char+} and @code{chars} can
2650: create an address that is not cell-aligned. @code{Aligned ( addr --
2651: a-addr )} produces the next aligned address:
2652:
2653: @example
2654: v3 char+ aligned .s @@ .
2655: v3 char+ .s @@ .
2656: @end example
2657:
2658: Similarly, @code{align} advances @code{here} to the next aligned
2659: address:
2660:
2661: @example
2662: create v5 97 c,
2663: here .
2664: align here .
2665: 1000 ,
2666: @end example
2667:
2668: Note that you should use aligned addresses even if your processor does
2669: not require them, if you want your program to be portable.
2670:
2671:
2672: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Alignment Tutorial, Tutorial
2673: @section Interpretation and Compilation Semantics and Immediacy
2674:
2675: When a word is compiled, it behaves differently from being interpreted.
2676: E.g., consider @code{+}:
2677:
2678: @example
2679: 1 2 + .
2680: : foo + ;
2681: @end example
2682:
2683: These two behaviours are known as compilation and interpretation
2684: semantics. For normal words (e.g., @code{+}), the compilation semantics
2685: is to append the interpretation semantics to the currently defined word
2686: (@code{foo} in the example above). I.e., when @code{foo} is executed
2687: later, the interpretation semantics of @code{+} (i.e., adding two
2688: numbers) will be performed.
2689:
2690: However, there are words with non-default compilation semantics, e.g.,
2691: the control-flow words like @code{if}. You can use @code{immediate} to
2692: change the compilation semantics of the last defined word to be equal to
2693: the interpretation semantics:
2694:
2695: @example
2696: : [FOO] ( -- )
2697: 5 . ; immediate
2698:
2699: [FOO]
2700: : bar ( -- )
2701: [FOO] ;
2702: bar
2703: see bar
2704: @end example
2705:
2706: Two conventions to mark words with non-default compilation semnatics are
2707: names with brackets (more frequently used) and to write them all in
2708: upper case (less frequently used).
2709:
2710: In Gforth (and many other systems) you can also remove the
2711: interpretation semantics with @code{compile-only} (the compilation
2712: semantics is derived from the original interpretation semantics):
2713:
2714: @example
2715: : flip ( -- )
2716: 6 . ; compile-only \ but not immediate
2717: flip
2718:
2719: : flop ( -- )
2720: flip ;
2721: flop
2722: @end example
2723:
2724: In this example the interpretation semantics of @code{flop} is equal to
2725: the original interpretation semantics of @code{flip}.
2726:
2727: The text interpreter has two states: in interpret state, it performs the
2728: interpretation semantics of words it encounters; in compile state, it
2729: performs the compilation semantics of these words.
2730:
2731: Among other things, @code{:} switches into compile state, and @code{;}
2732: switches back to interpret state. They contain the factors @code{]}
2733: (switch to compile state) and @code{[} (switch to interpret state), that
2734: do nothing but switch the state.
2735:
2736: @example
2737: : xxx ( -- )
2738: [ 5 . ]
2739: ;
2740:
2741: xxx
2742: see xxx
2743: @end example
2744:
2745: These brackets are also the source of the naming convention mentioned
2746: above.
2747:
2748:
2749: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2750: @section Execution Tokens
2751:
2752: @code{' word} gives you the execution token (XT) of a word. The XT is a
2753: cell representing the interpretation semantics of a word. You can
2754: execute this semantics with @code{execute}:
2755:
2756: @example
2757: ' + .s
2758: 1 2 rot execute .
2759: @end example
2760:
2761: The XT is similar to a function pointer in C. However, parameter
2762: passing through the stack makes it a little more flexible:
2763:
2764: @example
2765: : map-array ( ... addr u xt -- ... )
2766: \ executes xt ( ... x -- ... ) for every element of the array starting
2767: \ at addr and containing u elements
2768: @{ xt @}
2769: cells over + swap ?do
2770: i @@ xt execute
2771: 1 cells +loop ;
2772:
2773: create a 3 , 4 , 2 , -1 , 4 ,
2774: a 5 ' . map-array .s
2775: 0 a 5 ' + map-array .
2776: s" max-n" environment? drop .s
2777: a 5 ' min map-array .
2778: @end example
2779:
2780: You can use map-array with the XTs of words that consume one element
2781: more than they produce. In theory you can also use it with other XTs,
2782: but the stack effect then depends on the size of the array, which is
2783: hard to understand.
2784:
2785: Since XTs are cell-sized, you can store them in memory and manipulate
2786: them on the stack like other cells. You can also compile the XT into a
2787: word with @code{compile,}:
2788:
2789: @example
2790: : foo1 ( n1 n2 -- n )
2791: [ ' + compile, ] ;
2792: see foo
2793: @end example
2794:
2795: This is non-standard, because @code{compile,} has no compilation
2796: semantics in the standard, but it works in good Forth systems. For the
2797: broken ones, use
2798:
2799: @example
2800: : [compile,] compile, ; immediate
2801:
2802: : foo1 ( n1 n2 -- n )
2803: [ ' + ] [compile,] ;
2804: see foo
2805: @end example
2806:
2807: @code{'} is a word with default compilation semantics; it parses the
2808: next word when its interpretation semantics are executed, not during
2809: compilation:
2810:
2811: @example
2812: : foo ( -- xt )
2813: ' ;
2814: see foo
2815: : bar ( ... "word" -- ... )
2816: ' execute ;
2817: see bar
2818: 1 2 bar +
2819: @end example
2820:
2821: You often want to parse a word during compilation and compile its XT so
2822: it will be pushed on the stack at run-time. @code{[']} does this:
2823:
2824: @example
2825: : xt-+ ( -- xt )
2826: ['] + ;
2827: see xt-+
2828: 1 2 xt-+ execute .
2829: @end example
2830:
2831: Many programmers tend to see @code{'} and the word it parses as one
2832: unit, and expect it to behave like @code{[']} when compiled, and are
2833: confused by the actual behaviour. If you are, just remember that the
2834: Forth system just takes @code{'} as one unit and has no idea that it is
2835: a parsing word (attempts to convenience programmers in this issue have
2836: usually resulted in even worse pitfalls, see
2837: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz}).
2838:
2839: Note that the state of the interpreter does not come into play when
2840: creating and executing XTs. I.e., even when you execute @code{'} in
2841: compile state, it still gives you the interpretation semantics. And
2842: whatever that state is, @code{execute} performs the semantics
2843: represented by the XT (i.e., the interpretation semantics).
2844:
2845:
2846: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
2847: @section Exceptions
2848:
2849: @code{throw ( n -- )} causes an exception unless n is zero.
2850:
2851: @example
2852: 100 throw .s
2853: 0 throw .s
2854: @end example
2855:
2856: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
2857: it catches exceptions and pushes the number of the exception on the
2858: stack (or 0, if the xt executed without exception). If there was an
2859: exception, the stacks have the same depth as when entering @code{catch}:
2860:
2861: @example
2862: .s
2863: 3 0 ' / catch .s
2864: 3 2 ' / catch .s
2865: @end example
2866:
2867: @assignment
2868: Try the same with @code{execute} instead of @code{catch}.
2869: @endassignment
2870:
2871: @code{Throw} always jumps to the dynamically next enclosing
2872: @code{catch}, even if it has to leave several call levels to achieve
2873: this:
2874:
2875: @example
2876: : foo 100 throw ;
2877: : foo1 foo ." after foo" ;
2878: : bar ['] foo1 catch ;
2879: bar
2880: @end example
2881:
2882: It is often important to restore a value upon leaving a definition, even
2883: if the definition is left through an exception. You can ensure this
2884: like this:
2885:
2886: @example
2887: : ...
2888: save-x
2889: ['] word-changing-x catch ( ... n )
2890: restore-x
2891: ( ... n ) throw ;
2892: @end example
2893:
2894: Gforth provides an alternative syntax in addition to @code{catch}:
2895: @code{try ... recover ... endtry}. If the code between @code{try} and
2896: @code{recover} has an exception, the stack depths are restored, the
2897: exception number is pushed on the stack, and the code between
2898: @code{recover} and @code{endtry} is performed. E.g., the definition for
2899: @code{catch} is
2900:
2901: @example
2902: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
2903: try
2904: execute 0
2905: recover
2906: nip
2907: endtry ;
2908: @end example
2909:
2910: The equivalent to the restoration code above is
2911:
2912: @example
2913: : ...
2914: save-x
2915: try
2916: word-changing-x
2917: end-try
2918: restore-x
2919: throw ;
2920: @end example
2921:
2922: As you can see, the @code{recover} part is optional.
2923:
2924:
2925: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
2926: @section Defining Words
2927:
2928: @code{:}, @code{create}, and @code{variable} are definition words: They
2929: define other words. @code{Constant} is another definition word:
2930:
2931: @example
2932: 5 constant foo
2933: foo .
2934: @end example
2935:
2936: You can also use the prefixes @code{2} (double-cell) and @code{f}
2937: (floating point) with @code{variable} and @code{constant}.
2938:
2939: You can also define your own defining words. E.g.:
2940:
2941: @example
2942: : variable ( "name" -- )
2943: create 0 , ;
2944: @end example
2945:
2946: You can also define defining words that create words that do something
2947: other than just producing their address:
2948:
2949: @example
2950: : constant ( n "name" -- )
2951: create ,
2952: does> ( -- n )
2953: ( addr ) @@ ;
2954:
2955: 5 constant foo
2956: foo .
2957: @end example
2958:
2959: The definition of @code{constant} above ends at the @code{does>}; i.e.,
2960: @code{does>} replaces @code{;}, but it also does something else: It
2961: changes the last defined word such that it pushes the address of the
2962: body of the word and then performs the code after the @code{does>}
2963: whenever it is called.
2964:
2965: In the example above, @code{constant} uses @code{,} to store 5 into the
2966: body of @code{foo}. When @code{foo} executes, it pushes the address of
2967: the body onto the stack, then (in the code after the @code{does>})
2968: fetches the 5 from there.
2969:
2970: The stack comment near the @code{does>} reflects the stack effect of the
2971: defined word, not the stack effect of the code after the @code{does>}
2972: (the difference is that the code expects the address of the body that
2973: the stack comment does not show).
2974:
2975: You can use these definition words to do factoring in cases that involve
2976: (other) definition words. E.g., a field offset is always added to an
2977: address. Instead of defining
2978:
2979: @example
2980: 2 cells constant offset-field1
2981: @end example
2982:
2983: and using this like
2984:
2985: @example
2986: ( addr ) offset-field1 +
2987: @end example
2988:
2989: you can define a definition word
2990:
2991: @example
2992: : simple-field ( n "name" -- )
2993: create ,
2994: does> ( n1 -- n1+n )
2995: ( addr ) @@ + ;
2996: @end example
2997:
2998: Definition and use of field offsets now look like this:
2999:
3000: @example
3001: 2 cells simple-field field1
3002: ( addr ) field1
3003: @end example
3004:
3005: If you want to do something with the word without performing the code
3006: after the @code{does>}, you can access the body of a @code{create}d word
3007: with @code{>body ( xt -- addr )}:
3008:
3009: @example
3010: : value ( n "name" -- )
3011: create ,
3012: does> ( -- n1 )
3013: @@ ;
3014: : to ( n "name" -- )
3015: ' >body ! ;
3016:
3017: 5 value foo
3018: foo .
3019: 7 to foo
3020: foo .
3021: @end example
3022:
3023: @assignment
3024: Define @code{defer ( "name" -- )}, which creates a word that stores an
3025: XT (at the start the XT of @code{abort}), and upon execution
3026: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3027: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3028: recursion is one application of @code{defer}.
3029: @endassignment
3030:
3031: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3032: @section Arrays and Records
3033:
3034: Forth has no standard words for defining data structures such as arrays
3035: and records (structs in C terminology), but you can build them yourself
3036: based on address arithmetic. You can also define words for defining
3037: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
3038:
3039: One of the first projects a Forth newcomer sets out upon when learning
3040: about defining words is an array defining word (possibly for
3041: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3042: learn something from it. However, don't be disappointed when you later
3043: learn that you have little use for these words (inappropriate use would
3044: be even worse). I have not yet found a set of useful array words yet;
3045: the needs are just too diverse, and named, global arrays (the result of
3046: naive use of defining words) are often not flexible enough (e.g.,
3047: consider how to pass them as parameters).
3048:
3049: On the other hand, there is a useful set of record words, and it has
3050: been defined in @file{compat/struct.fs}; these words are predefined in
3051: Gforth. They are explained in depth elsewhere in this manual (see
3052: @pxref{Structures}). The @code{simple-field} example above is
3053: simplified variant of fields in this package.
3054:
3055:
3056: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3057: @section @code{POSTPONE}
3058:
3059: You can compile the compilation semantics (instead of compiling the
3060: interpretation semantics) of a word with @code{POSTPONE}:
3061:
3062: @example
3063: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3064: POSTPONE + ; immediate
3065: : foo ( n1 n2 -- n )
3066: MY-+ ;
3067: 1 2 foo .
3068: see foo
3069: @end example
3070:
3071: During the definition of @code{foo} the text interpreter performs the
3072: compilation semantics of @code{MY-+}, which performs the compilation
3073: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3074:
3075: This example also displays separate stack comments for the compilation
3076: semantics and for the stack effect of the compiled code. For words with
3077: default compilation semantics these stack effects are usually not
3078: displayed; the stack effect of the compilation semantics is always
3079: @code{( -- )} for these words, the stack effect for the compiled code is
3080: the stack effect of the interpretation semantics.
3081:
3082: Note that the state of the interpreter does not come into play when
3083: performing the compilation semantics in this way. You can also perform
3084: it interpretively, e.g.:
3085:
3086: @example
3087: : foo2 ( n1 n2 -- n )
3088: [ MY-+ ] ;
3089: 1 2 foo .
3090: see foo
3091: @end example
3092:
3093: However, there are some broken Forth systems where this does not always
3094: work, and therefore this practice has been declared non-standard in
3095: 1999.
3096: @c !! repair.fs
3097:
3098: Here is another example for using @code{POSTPONE}:
3099:
3100: @example
3101: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3102: POSTPONE negate POSTPONE + ; immediate compile-only
3103: : bar ( n1 n2 -- n )
3104: MY-- ;
3105: 2 1 bar .
3106: see bar
3107: @end example
3108:
3109: You can define @code{ENDIF} in this way:
3110:
3111: @example
3112: : ENDIF ( Compilation: orig -- )
3113: POSTPONE then ; immediate
3114: @end example
3115:
3116: @assignment
3117: Write @code{MY-2DUP} that has compilation semantics equivalent to
3118: @code{2dup}, but compiles @code{over over}.
3119: @endassignment
3120:
3121: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3122: @section @code{Literal}
3123:
3124: You cannot @code{POSTPONE} numbers:
3125:
3126: @example
3127: : [FOO] POSTPONE 500 ; immediate
3128: @end example
3129:
3130: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
3131:
3132: @example
3133: : [FOO] ( compilation: --; run-time: -- n )
3134: 500 POSTPONE literal ; immediate
3135:
3136: : flip foo ;
3137: flip .
3138: see flip
3139: @end example
3140:
3141: @code{LITERAL} consumes a number at compile-time (when it's compilation
3142: semantics are executed) and pushes it at run-time (when the code it
3143: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3144: number computed at compile time into the current word:
3145:
3146: @example
3147: : bar ( -- n )
3148: [ 2 2 + ] literal ;
3149: see bar
3150: @end example
3151:
3152: @assignment
3153: Write @code{]L} which allows writing the example above as @code{: bar (
3154: -- n ) [ 2 2 + ]L ;}
3155: @endassignment
3156:
3157:
3158: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3159: @section Advanced macros
3160:
3161: Reconsider @code{map-array} from @ref{Execution Tokens
3162: Tutorial,, Execution Tokens}. It frequently performs @code{execute}, a
3163: relatively expensive operation in some implementations. You can use
3164: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3165: and produce a word that contains the word to be performed directly:
3166:
3167: @c use ]] ... [[
3168: @example
3169: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3170: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3171: \ array beginning at addr and containing u elements
3172: @{ xt @}
3173: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
3174: POSTPONE i POSTPONE @@ xt compile,
3175: 1 cells POSTPONE literal POSTPONE +loop ;
3176:
3177: : sum-array ( addr u -- n )
3178: 0 rot rot [ ' + compile-map-array ] ;
3179: see sum-array
3180: a 5 sum-array .
3181: @end example
3182:
3183: You can use the full power of Forth for generating the code; here's an
3184: example where the code is generated in a loop:
3185:
3186: @example
3187: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3188: \ n2=n1+(addr1)*n, addr2=addr1+cell
3189: POSTPONE tuck POSTPONE @@
3190: POSTPONE literal POSTPONE * POSTPONE +
3191: POSTPONE swap POSTPONE cell+ ;
3192:
3193: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
3194: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
3195: 0 postpone literal postpone swap
3196: [ ' compile-vmul-step compile-map-array ]
3197: postpone drop ;
3198: see compile-vmul
3199:
3200: : a-vmul ( addr -- n )
3201: \ n=a*v, where v is a vector that's as long as a and starts at addr
3202: [ a 5 compile-vmul ] ;
3203: see a-vmul
3204: a a-vmul .
3205: @end example
3206:
3207: This example uses @code{compile-map-array} to show off, but you could
3208: also use @code{map-array} instead (try it now).
3209:
3210: You can use this technique for efficient multiplication of large
3211: matrices. In matrix multiplication, you multiply every line of one
3212: matrix with every column of the other matrix. You can generate the code
3213: for one line once, and use it for every column. The only downside of
3214: this technique is that it is cumbersome to recover the memory consumed
3215: by the generated code when you are done (and in more complicated cases
3216: it is not possible portably).
3217:
3218: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3219: @section Compilation Tokens
3220:
3221: This section is Gforth-specific. You can skip it.
3222:
3223: @code{' word compile,} compiles the interpretation semantics. For words
3224: with default compilation semantics this is the same as performing the
3225: compilation semantics. To represent the compilation semantics of other
3226: words (e.g., words like @code{if} that have no interpretation
3227: semantics), Gforth has the concept of a compilation token (CT,
3228: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3229: You can perform the compilation semantics represented by a CT with
3230: @code{execute}:
3231:
3232: @example
3233: : foo2 ( n1 n2 -- n )
3234: [ comp' + execute ] ;
3235: see foo
3236: @end example
3237:
3238: You can compile the compilation semantics represented by a CT with
3239: @code{postpone,}:
3240:
3241: @example
3242: : foo3 ( -- )
3243: [ comp' + postpone, ] ;
3244: see foo3
3245: @end example
3246:
3247: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
3248: @code{comp'} is particularly useful for words that have no
3249: interpretation semantics:
3250:
3251: @example
3252: ' if
3253: comp' if .s
3254: @end example
3255:
3256:
3257: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3258: @section Wordlists and Search Order
3259:
3260: The dictionary is not just a memory area that allows you to allocate
3261: memory with @code{allot}, it also contains the Forth words, arranged in
3262: several wordlists. When searching for a word in a wordlist,
3263: conceptually you start searching at the youngest and proceed towards
3264: older words (in reality most systems nowadays use hash-tables); i.e., if
3265: you define a word with the same name as an older word, the new word
3266: shadows the older word.
3267:
3268: Which wordlists are searched in which order is determined by the search
3269: order. You can display the search order with @code{order}. It displays
3270: first the search order, starting with the wordlist searched first, then
3271: it displays the wordlist that will contain newly defined words.
3272:
3273: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
3274:
3275: @example
3276: wordlist constant mywords
3277: @end example
3278:
3279: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3280: defined words (the @emph{current} wordlist):
3281:
3282: @example
3283: mywords set-current
3284: order
3285: @end example
3286:
3287: Gforth does not display a name for the wordlist in @code{mywords}
3288: because this wordlist was created anonymously with @code{wordlist}.
3289:
3290: You can get the current wordlist with @code{get-current ( -- wid)}. If
3291: you want to put something into a specific wordlist without overall
3292: effect on the current wordlist, this typically looks like this:
3293:
3294: @example
3295: get-current mywords set-current ( wid )
3296: create someword
3297: ( wid ) set-current
3298: @end example
3299:
3300: You can write the search order with @code{set-order ( wid1 .. widn n --
3301: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3302: searched wordlist is topmost.
3303:
3304: @example
3305: get-order mywords swap 1+ set-order
3306: order
3307: @end example
3308:
3309: Yes, the order of wordlists in the output of @code{order} is reversed
3310: from stack comments and the output of @code{.s} and thus unintuitive.
3311:
3312: @assignment
3313: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3314: wordlist to the search order. Define @code{previous ( -- )}, which
3315: removes the first searched wordlist from the search order. Experiment
3316: with boundary conditions (you will see some crashes or situations that
3317: are hard or impossible to leave).
3318: @endassignment
3319:
3320: The search order is a powerful foundation for providing features similar
3321: to Modula-2 modules and C++ namespaces. However, trying to modularize
3322: programs in this way has disadvantages for debugging and reuse/factoring
3323: that overcome the advantages in my experience (I don't do huge projects,
3324: though). These disadvantages are not so clear in other
3325: languages/programming environments, because these langauges are not so
3326: strong in debugging and reuse.
3327:
3328:
3329: @c ******************************************************************
3330: @node Introduction, Words, Tutorial, Top
3331: @comment node-name, next, previous, up
3332: @chapter An Introduction to ANS Forth
3333: @cindex Forth - an introduction
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: :1: Undefined word
3414: qwer341
3415: ^^^^^^^
3416: $400D2BA8 Bounce
3417: $400DBDA8 no.extensions
3418: @end example
3419:
3420: The exact text, other than the ``Undefined word'' may differ slightly on
3421: 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 description of error messages see @ref{Error
3424: 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: :1: Undefined word
3475: 12 dup fred dup
3476: ^^^^
3477: $400D2BA8 Bounce
3478: $400DBDA8 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{clearstack 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{clearstack} and executes
3596: it; it tidies up the stack 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 will report 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''. I reworded big chunks of the ``how does that work''
3846: @c section (and, unusually for me, I think I even made it shorter!). See
3847: @c what you think -- I know I have not addressed your primary concern
3848: @c that it is too heavy-going for an introduction. From what I understood
3849: @c of your course notes it looks as though they might be a good framework.
3850: @c Things that I've tried to capture here are some things that came as a
3851: @c great revelation here when I first understood them. Also, I like the
3852: @c fact that a very simple code example shows up almost all of the issues
3853: @c that you need to understand to see how Forth works. That's unique and
3854: @c worthwhile to emphasise.
3855:
3856: Now we're going to take another look at the definition of @code{add-two}
3857: from the previous section. From our knowledge of the way that the text
3858: interpreter works, we would have expected this result when we tried to
3859: define @code{add-two}:
3860:
3861: @example
3862: @kbd{: add-two 2 + . ;@key{RET}}
3863: ^^^^^^^
3864: Error: Undefined word
3865: @end example
3866:
3867: The reason that this didn't happen is bound up in the way that @code{:}
3868: works. The word @code{:} does two special things. The first special
3869: thing that it does prevents the text interpreter from ever seeing the
3870: characters @code{add-two}. The text interpreter uses a variable called
3871: @cindex modifying >IN
3872: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
3873: input line. When it encounters the word @code{:} it behaves in exactly
3874: the same way as it does for any other word; it looks it up in the name
3875: dictionary, finds its xt and executes it. When @code{:} executes, it
3876: looks at the input buffer, finds the word @code{add-two} and advances the
3877: value of @code{>IN} to point past it. It then does some other stuff
3878: associated with creating the new definition (including creating an entry
3879: for @code{add-two} in the name dictionary). When the execution of @code{:}
3880: completes, control returns to the text interpreter, which is oblivious
3881: to the fact that it has been tricked into ignoring part of the input
3882: line.
3883:
3884: @cindex parsing words
3885: Words like @code{:} -- words that advance the value of @code{>IN} and so
3886: prevent the text interpreter from acting on the whole of the input line
3887: -- are called @dfn{parsing words}.
3888:
3889: @cindex @code{state} - effect on the text interpreter
3890: @cindex text interpreter - effect of state
3891: The second special thing that @code{:} does is change the value of a
3892: variable called @code{state}, which affects the way that the text
3893: interpreter behaves. When Gforth starts up, @code{state} has the value
3894: 0, and the text interpreter is said to be @dfn{interpreting}. During a
3895: colon definition (started with @code{:}), @code{state} is set to -1 and
3896: the text interpreter is said to be @dfn{compiling}.
3897:
3898: In this example, the text interpreter is compiling when it processes the
3899: string ``@code{2 + . ;}''. It still breaks the string down into
3900: character sequences in the same way. However, instead of pushing the
3901: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
3902: into the definition of @code{add-two} that will make the number @code{2} get
3903: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
3904: the behaviours of @code{+} and @code{.} are also compiled into the
3905: definition.
3906:
3907: One category of words don't get compiled. These so-called @dfn{immediate
3908: words} get executed (performed @i{now}) regardless of whether the text
3909: interpreter is interpreting or compiling. The word @code{;} is an
3910: immediate word. Rather than being compiled into the definition, it
3911: executes. Its effect is to terminate the current definition, which
3912: includes changing the value of @code{state} back to 0.
3913:
3914: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
3915: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
3916: definition.
3917:
3918: In Forth, every word or number can be described in terms of two
3919: properties:
3920:
3921: @itemize @bullet
3922: @item
3923: @cindex interpretation semantics
3924: Its @dfn{interpretation semantics} describe how it will behave when the
3925: text interpreter encounters it in @dfn{interpret} state. The
3926: interpretation semantics of a word are represented by an @dfn{execution
3927: token}.
3928: @item
3929: @cindex compilation semantics
3930: Its @dfn{compilation semantics} describe how it will behave when the
3931: text interpreter encounters it in @dfn{compile} state. The compilation
3932: semantics of a word are represented in an implementation-dependent way;
3933: Gforth uses a @dfn{compilation token}.
3934: @end itemize
3935:
3936: @noindent
3937: Numbers are always treated in a fixed way:
3938:
3939: @itemize @bullet
3940: @item
3941: When the number is @dfn{interpreted}, its behaviour is to push the
3942: number onto the stack.
3943: @item
3944: When the number is @dfn{compiled}, a piece of code is appended to the
3945: current definition that pushes the number when it runs. (In other words,
3946: the compilation semantics of a number are to postpone its interpretation
3947: semantics until the run-time of the definition that it is being compiled
3948: into.)
3949: @end itemize
3950:
3951: Words don't behave in such a regular way, but most have @i{default
3952: semantics} which means that they behave like this:
3953:
3954: @itemize @bullet
3955: @item
3956: The @dfn{interpretation semantics} of the word are to do something useful.
3957: @item
3958: The @dfn{compilation semantics} of the word are to append its
3959: @dfn{interpretation semantics} to the current definition (so that its
3960: run-time behaviour is to do something useful).
3961: @end itemize
3962:
3963: @cindex immediate words
3964: The actual behaviour of any particular word can be controlled by using
3965: the words @code{immediate} and @code{compile-only} when the word is
3966: defined. These words set flags in the name dictionary entry of the most
3967: recently defined word, and these flags are retrieved by the text
3968: interpreter when it finds the word in the name dictionary.
3969:
3970: A word that is marked as @dfn{immediate} has compilation semantics that
3971: are identical to its interpretation semantics. In other words, it
3972: behaves like this:
3973:
3974: @itemize @bullet
3975: @item
3976: The @dfn{interpretation semantics} of the word are to do something useful.
3977: @item
3978: The @dfn{compilation semantics} of the word are to do something useful
3979: (and actually the same thing); i.e., it is executed during compilation.
3980: @end itemize
3981:
3982: Marking a word as @dfn{compile-only} prohibits the text interpreter from
3983: performing the interpretation semantics of the word directly; an attempt
3984: to do so will generate an error. It is never necessary to use
3985: @code{compile-only} (and it is not even part of ANS Forth, though it is
3986: provided by many implementations) but it is good etiquette to apply it
3987: to a word that will not behave correctly (and might have unexpected
3988: side-effects) in interpret state. For example, it is only legal to use
3989: the conditional word @code{IF} within a definition. If you forget this
3990: and try to use it elsewhere, the fact that (in Gforth) it is marked as
3991: @code{compile-only} allows the text interpreter to generate a helpful
3992: error message rather than subjecting you to the consequences of your
3993: folly.
3994:
3995: This example shows the difference between an immediate and a
3996: non-immediate word:
3997:
3998: @example
3999: : show-state state @@ . ;
4000: : show-state-now show-state ; immediate
4001: : word1 show-state ;
4002: : word2 show-state-now ;
4003: @end example
4004:
4005: The word @code{immediate} after the definition of @code{show-state-now}
4006: makes that word an immediate word. These definitions introduce a new
4007: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4008: variable, and leaves it on the stack. Therefore, the behaviour of
4009: @code{show-state} is to print a number that represents the current value
4010: of @code{state}.
4011:
4012: When you execute @code{word1}, it prints the number 0, indicating that
4013: the system is interpreting. When the text interpreter compiled the
4014: definition of @code{word1}, it encountered @code{show-state} whose
4015: compilation semantics are to append its interpretation semantics to the
4016: current definition. When you execute @code{word1}, it performs the
4017: interpretation semantics of @code{show-state}. At the time that @code{word1}
4018: (and therefore @code{show-state}) are executed, the system is
4019: interpreting.
4020:
4021: When you pressed @key{RET} after entering the definition of @code{word2},
4022: you should have seen the number -1 printed, followed by ``@code{
4023: ok}''. When the text interpreter compiled the definition of
4024: @code{word2}, it encountered @code{show-state-now}, an immediate word,
4025: whose compilation semantics are therefore to perform its interpretation
4026: semantics. It is executed straight away (even before the text
4027: interpreter has moved on to process another group of characters; the
4028: @code{;} in this example). The effect of executing it are to display the
4029: value of @code{state} @i{at the time that the definition of}
4030: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4031: system is compiling at this time. If you execute @code{word2} it does
4032: nothing at all.
4033:
4034: @cindex @code{."}, how it works
4035: Before leaving the subject of immediate words, consider the behaviour of
4036: @code{."} in the definition of @code{greet}, in the previous
4037: section. This word is both a parsing word and an immediate word. Notice
4038: that there is a space between @code{."} and the start of the text
4039: @code{Hello and welcome}, but that there is no space between the last
4040: letter of @code{welcome} and the @code{"} character. The reason for this
4041: is that @code{."} is a Forth word; it must have a space after it so that
4042: the text interpreter can identify it. The @code{"} is not a Forth word;
4043: it is a @dfn{delimiter}. The examples earlier show that, when the string
4044: is displayed, there is neither a space before the @code{H} nor after the
4045: @code{e}. Since @code{."} is an immediate word, it executes at the time
4046: that @code{greet} is defined. When it executes, its behaviour is to
4047: search forward in the input line looking for the delimiter. When it
4048: finds the delimiter, it updates @code{>IN} to point past the
4049: delimiter. It also compiles some magic code into the definition of
4050: @code{greet}; the xt of a run-time routine that prints a text string. It
4051: compiles the string @code{Hello and welcome} into memory so that it is
4052: available to be printed later. When the text interpreter gains control,
4053: the next word it finds in the input stream is @code{;} and so it
4054: terminates the definition of @code{greet}.
4055:
4056:
4057: @comment ----------------------------------------------
4058: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4059: @section Forth is written in Forth
4060: @cindex structure of Forth programs
4061:
4062: When you start up a Forth compiler, a large number of definitions
4063: already exist. In Forth, you develop a new application using bottom-up
4064: programming techniques to create new definitions that are defined in
4065: terms of existing definitions. As you create each definition you can
4066: test and debug it interactively.
4067:
4068: If you have tried out the examples in this section, you will probably
4069: have typed them in by hand; when you leave Gforth, your definitions will
4070: be lost. You can avoid this by using a text editor to enter Forth source
4071: code into a file, and then loading code from the file using
4072: @code{include} (@pxref{Forth source files}). A Forth source file is
4073: processed by the text interpreter, just as though you had typed it in by
4074: hand@footnote{Actually, there are some subtle differences -- see
4075: @ref{The Text Interpreter}.}.
4076:
4077: Gforth also supports the traditional Forth alternative to using text
4078: files for program entry (@pxref{Blocks}).
4079:
4080: In common with many, if not most, Forth compilers, most of Gforth is
4081: actually written in Forth. All of the @file{.fs} files in the
4082: installation directory@footnote{For example,
4083: @file{/usr/local/share/gforth...}} are Forth source files, which you can
4084: study to see examples of Forth programming.
4085:
4086: Gforth maintains a history file that records every line that you type to
4087: the text interpreter. This file is preserved between sessions, and is
4088: used to provide a command-line recall facility. If you enter long
4089: definitions by hand, you can use a text editor to paste them out of the
4090: history file into a Forth source file for reuse at a later time
4091: (for more information @pxref{Command-line editing}).
4092:
4093:
4094: @comment ----------------------------------------------
4095: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4096: @section Review - elements of a Forth system
4097: @cindex elements of a Forth system
4098:
4099: To summarise this chapter:
4100:
4101: @itemize @bullet
4102: @item
4103: Forth programs use @dfn{factoring} to break a problem down into small
4104: fragments called @dfn{words} or @dfn{definitions}.
4105: @item
4106: Forth program development is an interactive process.
4107: @item
4108: The main command loop that accepts input, and controls both
4109: interpretation and compilation, is called the @dfn{text interpreter}
4110: (also known as the @dfn{outer interpreter}).
4111: @item
4112: Forth has a very simple syntax, consisting of words and numbers
4113: separated by spaces or carriage-return characters. Any additional syntax
4114: is imposed by @dfn{parsing words}.
4115: @item
4116: Forth uses a stack to pass parameters between words. As a result, it
4117: uses postfix notation.
4118: @item
4119: To use a word that has previously been defined, the text interpreter
4120: searches for the word in the @dfn{name dictionary}.
4121: @item
4122: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
4123: @item
4124: The text interpreter uses the value of @code{state} to select between
4125: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4126: semantics} of a word that it encounters.
4127: @item
4128: The relationship between the @dfn{interpretation semantics} and
4129: @dfn{compilation semantics} for a word
4130: depend upon the way in which the word was defined (for example, whether
4131: it is an @dfn{immediate} word).
4132: @item
4133: Forth definitions can be implemented in Forth (called @dfn{high-level
4134: definitions}) or in some other way (usually a lower-level language and
4135: as a result often called @dfn{low-level definitions}, @dfn{code
4136: definitions} or @dfn{primitives}).
4137: @item
4138: Many Forth systems are implemented mainly in Forth.
4139: @end itemize
4140:
4141:
4142: @comment ----------------------------------------------
4143: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
4144: @section Where To Go Next
4145: @cindex where to go next
4146:
4147: Amazing as it may seem, if you have read (and understood) this far, you
4148: know almost all the fundamentals about the inner workings of a Forth
4149: system. You certainly know enough to be able to read and understand the
4150: rest of this manual and the ANS Forth document, to learn more about the
4151: facilities that Forth in general and Gforth in particular provide. Even
4152: scarier, you know almost enough to implement your own Forth system.
4153: However, that's not a good idea just yet... better to try writing some
4154: programs in Gforth.
4155:
4156: Forth has such a rich vocabulary that it can be hard to know where to
4157: start in learning it. This section suggests a few sets of words that are
4158: enough to write small but useful programs. Use the word index in this
4159: document to learn more about each word, then try it out and try to write
4160: small definitions using it. Start by experimenting with these words:
4161:
4162: @itemize @bullet
4163: @item
4164: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4165: @item
4166: Comparison: @code{MIN MAX =}
4167: @item
4168: Logic: @code{AND OR XOR NOT}
4169: @item
4170: Stack manipulation: @code{DUP DROP SWAP OVER}
4171: @item
4172: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
4173: @item
4174: Input/Output: @code{. ." EMIT CR KEY}
4175: @item
4176: Defining words: @code{: ; CREATE}
4177: @item
4178: Memory allocation words: @code{ALLOT ,}
4179: @item
4180: Tools: @code{SEE WORDS .S MARKER}
4181: @end itemize
4182:
4183: When you have mastered those, go on to:
4184:
4185: @itemize @bullet
4186: @item
4187: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
4188: @item
4189: Memory access: @code{@@ !}
4190: @end itemize
4191:
4192: When you have mastered these, there's nothing for it but to read through
4193: the whole of this manual and find out what you've missed.
4194:
4195: @comment ----------------------------------------------
4196: @node Exercises, , Where to go next, Introduction
4197: @section Exercises
4198: @cindex exercises
4199:
4200: TODO: provide a set of programming excercises linked into the stuff done
4201: already and into other sections of the manual. Provide solutions to all
4202: the exercises in a .fs file in the distribution.
4203:
4204: @c Get some inspiration from Starting Forth and Kelly&Spies.
4205:
4206: @c excercises:
4207: @c 1. take inches and convert to feet and inches.
4208: @c 2. take temperature and convert from fahrenheight to celcius;
4209: @c may need to care about symmetric vs floored??
4210: @c 3. take input line and do character substitution
4211: @c to encipher or decipher
4212: @c 4. as above but work on a file for in and out
4213: @c 5. take input line and convert to pig-latin
4214: @c
4215: @c thing of sets of things to exercise then come up with
4216: @c problems that need those things.
4217:
4218:
4219: @c ******************************************************************
4220: @node Words, Error messages, Introduction, Top
4221: @chapter Forth Words
4222: @cindex words
4223:
4224: @menu
4225: * Notation::
4226: * Comments::
4227: * Boolean Flags::
4228: * Arithmetic::
4229: * Stack Manipulation::
4230: * Memory::
4231: * Control Structures::
4232: * Defining Words::
4233: * Interpretation and Compilation Semantics::
4234: * Tokens for Words::
4235: * The Text Interpreter::
4236: * Word Lists::
4237: * Environmental Queries::
4238: * Files::
4239: * Blocks::
4240: * Other I/O::
4241: * Programming Tools::
4242: * Assembler and Code Words::
4243: * Threading Words::
4244: * Locals::
4245: * Structures::
4246: * Object-oriented Forth::
4247: * Passing Commands to the OS::
4248: * Keeping track of Time::
4249: * Miscellaneous Words::
4250: @end menu
4251:
4252: @node Notation, Comments, Words, Words
4253: @section Notation
4254: @cindex notation of glossary entries
4255: @cindex format of glossary entries
4256: @cindex glossary notation format
4257: @cindex word glossary entry format
4258:
4259: The Forth words are described in this section in the glossary notation
4260: that has become a de-facto standard for Forth texts, i.e.,
4261:
4262: @format
4263: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
4264: @end format
4265: @i{Description}
4266:
4267: @table @var
4268: @item word
4269: The name of the word.
4270:
4271: @item Stack effect
4272: @cindex stack effect
4273: The stack effect is written in the notation @code{@i{before} --
4274: @i{after}}, where @i{before} and @i{after} describe the top of
4275: stack entries before and after the execution of the word. The rest of
4276: the stack is not touched by the word. The top of stack is rightmost,
4277: i.e., a stack sequence is written as it is typed in. Note that Gforth
4278: uses a separate floating point stack, but a unified stack
4279: notation. Also, return stack effects are not shown in @i{stack
4280: effect}, but in @i{Description}. The name of a stack item describes
4281: the type and/or the function of the item. See below for a discussion of
4282: the types.
4283:
4284: All words have two stack effects: A compile-time stack effect and a
4285: run-time stack effect. The compile-time stack-effect of most words is
4286: @i{ -- }. If the compile-time stack-effect of a word deviates from
4287: this standard behaviour, or the word does other unusual things at
4288: compile time, both stack effects are shown; otherwise only the run-time
4289: stack effect is shown.
4290:
4291: @cindex pronounciation of words
4292: @item pronunciation
4293: How the word is pronounced.
4294:
4295: @cindex wordset
4296: @item wordset
4297: The ANS Forth standard is divided into several word sets. A standard
4298: system need not support all of them. Therefore, in theory, the fewer
4299: word sets your program uses the more portable it will be. However, we
4300: suspect that most ANS Forth systems on personal machines will feature
4301: all word sets. Words that are not defined in ANS Forth have
4302: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
4303: describes words that will work in future releases of Gforth;
4304: @code{gforth-internal} words are more volatile. Environmental query
4305: strings are also displayed like words; you can recognize them by the
4306: @code{environment} in the word set field.
4307:
4308: @item Description
4309: A description of the behaviour of the word.
4310: @end table
4311:
4312: @cindex types of stack items
4313: @cindex stack item types
4314: The type of a stack item is specified by the character(s) the name
4315: starts with:
4316:
4317: @table @code
4318: @item f
4319: @cindex @code{f}, stack item type
4320: Boolean flags, i.e. @code{false} or @code{true}.
4321: @item c
4322: @cindex @code{c}, stack item type
4323: Char
4324: @item w
4325: @cindex @code{w}, stack item type
4326: Cell, can contain an integer or an address
4327: @item n
4328: @cindex @code{n}, stack item type
4329: signed integer
4330: @item u
4331: @cindex @code{u}, stack item type
4332: unsigned integer
4333: @item d
4334: @cindex @code{d}, stack item type
4335: double sized signed integer
4336: @item ud
4337: @cindex @code{ud}, stack item type
4338: double sized unsigned integer
4339: @item r
4340: @cindex @code{r}, stack item type
4341: Float (on the FP stack)
4342: @item a-
4343: @cindex @code{a_}, stack item type
4344: Cell-aligned address
4345: @item c-
4346: @cindex @code{c_}, stack item type
4347: Char-aligned address (note that a Char may have two bytes in Windows NT)
4348: @item f-
4349: @cindex @code{f_}, stack item type
4350: Float-aligned address
4351: @item df-
4352: @cindex @code{df_}, stack item type
4353: Address aligned for IEEE double precision float
4354: @item sf-
4355: @cindex @code{sf_}, stack item type
4356: Address aligned for IEEE single precision float
4357: @item xt
4358: @cindex @code{xt}, stack item type
4359: Execution token, same size as Cell
4360: @item wid
4361: @cindex @code{wid}, stack item type
4362: Word list ID, same size as Cell
4363: @item f83name
4364: @cindex @code{f83name}, stack item type
4365: Pointer to a name structure
4366: @item "
4367: @cindex @code{"}, stack item type
4368: string in the input stream (not on the stack). The terminating character
4369: is a blank by default. If it is not a blank, it is shown in @code{<>}
4370: quotes.
4371: @end table
4372:
4373: @node Comments, Boolean Flags, Notation, Words
4374: @section Comments
4375: @cindex comments
4376:
4377: Forth supports two styles of comment; the traditional @i{in-line} comment,
4378: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
4379:
4380:
4381: doc-(
4382: doc-\
4383: doc-\G
4384:
4385:
4386: @node Boolean Flags, Arithmetic, Comments, Words
4387: @section Boolean Flags
4388: @cindex Boolean flags
4389:
4390: A Boolean flag is cell-sized. A cell with all bits clear represents the
4391: flag @code{false} and a flag with all bits set represents the flag
4392: @code{true}. Words that check a flag (for example, @code{IF}) will treat
4393: a cell that has @i{any} bit set as @code{true}.
4394:
4395:
4396: doc-true
4397: doc-false
4398: doc-on
4399: doc-off
4400:
4401:
4402: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
4403: @section Arithmetic
4404: @cindex arithmetic words
4405:
4406: @cindex division with potentially negative operands
4407: Forth arithmetic is not checked, i.e., you will not hear about integer
4408: overflow on addition or multiplication, you may hear about division by
4409: zero if you are lucky. The operator is written after the operands, but
4410: the operands are still in the original order. I.e., the infix @code{2-1}
4411: corresponds to @code{2 1 -}. Forth offers a variety of division
4412: operators. If you perform division with potentially negative operands,
4413: you do not want to use @code{/} or @code{/mod} with its undefined
4414: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4415: former, @pxref{Mixed precision}).
4416: @comment TODO discuss the different division forms and the std approach
4417:
4418: @menu
4419: * Single precision::
4420: * Bitwise operations::
4421: * Double precision:: Double-cell integer arithmetic
4422: * Numeric comparison::
4423: * Mixed precision:: Operations with single and double-cell integers
4424: * Floating Point::
4425: @end menu
4426:
4427: @node Single precision, Bitwise operations, Arithmetic, Arithmetic
4428: @subsection Single precision
4429: @cindex single precision arithmetic words
4430:
4431: By default, numbers in Forth are single-precision integers that are 1
4432: cell in size. They can be signed or unsigned, depending upon how you
4433: treat them. For the rules used by the text interpreter for recognising
4434: single-precision integers see @ref{Number Conversion}.
4435:
4436:
4437: doc-+
4438: doc-1+
4439: doc--
4440: doc-1-
4441: doc-*
4442: doc-/
4443: doc-mod
4444: doc-/mod
4445: doc-negate
4446: doc-abs
4447: doc-min
4448: doc-max
4449: doc-d>s
4450: doc-floored
4451:
4452:
4453: @node Bitwise operations, Double precision, Single precision, Arithmetic
4454: @subsection Bitwise operations
4455: @cindex bitwise operation words
4456:
4457:
4458: doc-and
4459: doc-or
4460: doc-xor
4461: doc-invert
4462: doc-lshift
4463: doc-rshift
4464: doc-2*
4465: doc-d2*
4466: doc-2/
4467: doc-d2/
4468:
4469:
4470: @node Double precision, Numeric comparison, Bitwise operations, Arithmetic
4471: @subsection Double precision
4472: @cindex double precision arithmetic words
4473:
4474: For the rules used by the text interpreter for
4475: recognising double-precision integers, see @ref{Number Conversion}.
4476:
4477: A double precision number is represented by a cell pair, with the most
4478: significant cell at the TOS. It is trivial to convert an unsigned
4479: single to an (unsigned) double; simply push a @code{0} onto the
4480: TOS. Since numbers are represented by Gforth using 2's complement
4481: arithmetic, converting a signed single to a (signed) double requires
4482: sign-extension across the most significant cell. This can be achieved
4483: using @code{s>d}. The moral of the story is that you cannot convert a
4484: number without knowing whether it represents an unsigned or a
4485: signed number.
4486:
4487:
4488: doc-s>d
4489: doc-d+
4490: doc-d-
4491: doc-dnegate
4492: doc-dabs
4493: doc-dmin
4494: doc-dmax
4495:
4496:
4497: @node Numeric comparison, Mixed precision, Double precision, Arithmetic
4498: @subsection Numeric comparison
4499: @cindex numeric comparison words
4500:
4501:
4502: doc-<
4503: doc-<=
4504: doc-<>
4505: doc-=
4506: doc->
4507: doc->=
4508:
4509: doc-0<
4510: doc-0<=
4511: doc-0<>
4512: doc-0=
4513: doc-0>
4514: doc-0>=
4515:
4516: doc-u<
4517: doc-u<=
4518: @c u<> and u= exist but are the same as <> and =
4519: @c doc-u<>
4520: @c doc-u=
4521: doc-u>
4522: doc-u>=
4523:
4524: doc-within
4525:
4526: doc-d<
4527: doc-d<=
4528: doc-d<>
4529: doc-d=
4530: doc-d>
4531: doc-d>=
4532:
4533: doc-d0<
4534: doc-d0<=
4535: doc-d0<>
4536: doc-d0=
4537: doc-d0>
4538: doc-d0>=
4539:
4540: doc-du<
4541: doc-du<=
4542: @c du<> and du= exist but are the same as d<> and d=
4543: @c doc-du<>
4544: @c doc-du=
4545: doc-du>
4546: doc-du>=
4547:
4548:
4549: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
4550: @subsection Mixed precision
4551: @cindex mixed precision arithmetic words
4552:
4553:
4554: doc-m+
4555: doc-*/
4556: doc-*/mod
4557: doc-m*
4558: doc-um*
4559: doc-m*/
4560: doc-um/mod
4561: doc-fm/mod
4562: doc-sm/rem
4563:
4564:
4565: @node Floating Point, , Mixed precision, Arithmetic
4566: @subsection Floating Point
4567: @cindex floating point arithmetic words
4568:
4569: For the rules used by the text interpreter for
4570: recognising floating-point numbers see @ref{Number Conversion}.
4571:
4572: Gforth has a separate floating point
4573: stack, but the documentation uses the unified notation.
4574:
4575: @cindex floating-point arithmetic, pitfalls
4576: Floating point numbers have a number of unpleasant surprises for the
4577: unwary (e.g., floating point addition is not associative) and even a few
4578: for the wary. You should not use them unless you know what you are doing
4579: or you don't care that the results you get are totally bogus. If you
4580: want to learn about the problems of floating point numbers (and how to
4581: avoid them), you might start with @cite{David Goldberg, What Every
4582: Computer Scientist Should Know About Floating-Point Arithmetic, ACM
4583: Computing Surveys 23(1):5@minus{}48, March 1991}
4584: (@uref{http://www.validgh.com/goldberg/paper.ps}).
4585:
4586:
4587: doc-d>f
4588: doc-f>d
4589: doc-f+
4590: doc-f-
4591: doc-f*
4592: doc-f/
4593: doc-fnegate
4594: doc-fabs
4595: doc-fmax
4596: doc-fmin
4597: doc-floor
4598: doc-fround
4599: doc-f**
4600: doc-fsqrt
4601: doc-fexp
4602: doc-fexpm1
4603: doc-fln
4604: doc-flnp1
4605: doc-flog
4606: doc-falog
4607: doc-f2*
4608: doc-f2/
4609: doc-1/f
4610: doc-precision
4611: doc-set-precision
4612:
4613: @cindex angles in trigonometric operations
4614: @cindex trigonometric operations
4615: Angles in floating point operations are given in radians (a full circle
4616: has 2 pi radians).
4617:
4618: doc-fsin
4619: doc-fcos
4620: doc-fsincos
4621: doc-ftan
4622: doc-fasin
4623: doc-facos
4624: doc-fatan
4625: doc-fatan2
4626: doc-fsinh
4627: doc-fcosh
4628: doc-ftanh
4629: doc-fasinh
4630: doc-facosh
4631: doc-fatanh
4632: doc-pi
4633:
4634: @cindex equality of floats
4635: @cindex floating-point comparisons
4636: One particular problem with floating-point arithmetic is that comparison
4637: for equality often fails when you would expect it to succeed. For this
4638: reason approximate equality is often preferred (but you still have to
4639: know what you are doing). The comparison words are:
4640:
4641: doc-f~rel
4642: doc-f~abs
4643: doc-f=
4644: doc-f~
4645: doc-f<>
4646:
4647: doc-f<
4648: doc-f<=
4649: doc-f>
4650: doc-f>=
4651:
4652: doc-f0<
4653: doc-f0<=
4654: doc-f0<>
4655: doc-f0=
4656: doc-f0>
4657: doc-f0>=
4658:
4659:
4660: @node Stack Manipulation, Memory, Arithmetic, Words
4661: @section Stack Manipulation
4662: @cindex stack manipulation words
4663:
4664: @cindex floating-point stack in the standard
4665: Gforth maintains a number of separate stacks:
4666:
4667: @cindex data stack
4668: @cindex parameter stack
4669: @itemize @bullet
4670: @item
4671: A data stack (also known as the @dfn{parameter stack}) -- for
4672: characters, cells, addresses, and double cells.
4673:
4674: @cindex floating-point stack
4675: @item
4676: A floating point stack -- for holding floating point (FP) numbers.
4677:
4678: @cindex return stack
4679: @item
4680: A return stack -- for holding the return addresses of colon
4681: definitions and other (non-FP) data.
4682:
4683: @cindex locals stack
4684: @item
4685: A locals stack -- for holding local variables.
4686: @end itemize
4687:
4688: @menu
4689: * Data stack::
4690: * Floating point stack::
4691: * Return stack::
4692: * Locals stack::
4693: * Stack pointer manipulation::
4694: @end menu
4695:
4696: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
4697: @subsection Data stack
4698: @cindex data stack manipulation words
4699: @cindex stack manipulations words, data stack
4700:
4701:
4702: doc-drop
4703: doc-nip
4704: doc-dup
4705: doc-over
4706: doc-tuck
4707: doc-swap
4708: doc-pick
4709: doc-rot
4710: doc--rot
4711: doc-?dup
4712: doc-roll
4713: doc-2drop
4714: doc-2nip
4715: doc-2dup
4716: doc-2over
4717: doc-2tuck
4718: doc-2swap
4719: doc-2rot
4720:
4721:
4722: @node Floating point stack, Return stack, Data stack, Stack Manipulation
4723: @subsection Floating point stack
4724: @cindex floating-point stack manipulation words
4725: @cindex stack manipulation words, floating-point stack
4726:
4727: Whilst every sane Forth has a separate floating-point stack, it is not
4728: strictly required; an ANS Forth system could theoretically keep
4729: floating-point numbers on the data stack. As an additional difficulty,
4730: you don't know how many cells a floating-point number takes. It is
4731: reportedly possible to write words in a way that they work also for a
4732: unified stack model, but we do not recommend trying it. Instead, just
4733: say that your program has an environmental dependency on a separate
4734: floating-point stack.
4735:
4736: doc-floating-stack
4737:
4738: doc-fdrop
4739: doc-fnip
4740: doc-fdup
4741: doc-fover
4742: doc-ftuck
4743: doc-fswap
4744: doc-fpick
4745: doc-frot
4746:
4747:
4748: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
4749: @subsection Return stack
4750: @cindex return stack manipulation words
4751: @cindex stack manipulation words, return stack
4752:
4753: @cindex return stack and locals
4754: @cindex locals and return stack
4755: A Forth system is allowed to keep local variables on the
4756: return stack. This is reasonable, as local variables usually eliminate
4757: the need to use the return stack explicitly. So, if you want to produce
4758: a standard compliant program and you are using local variables in a
4759: word, forget about return stack manipulations in that word (refer to the
4760: standard document for the exact rules).
4761:
4762: doc->r
4763: doc-r>
4764: doc-r@
4765: doc-rdrop
4766: doc-2>r
4767: doc-2r>
4768: doc-2r@
4769: doc-2rdrop
4770:
4771:
4772: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
4773: @subsection Locals stack
4774:
4775: Gforth uses an extra locals stack. It is described, along with the
4776: reasons for its existence, in @ref{Implementation,Implementation of locals}.
4777:
4778: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
4779: @subsection Stack pointer manipulation
4780: @cindex stack pointer manipulation words
4781:
4782: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
4783: doc-sp0
4784: doc-sp@
4785: doc-sp!
4786: doc-fp0
4787: doc-fp@
4788: doc-fp!
4789: doc-rp0
4790: doc-rp@
4791: doc-rp!
4792: doc-lp0
4793: doc-lp@
4794: doc-lp!
4795:
4796:
4797: @node Memory, Control Structures, Stack Manipulation, Words
4798: @section Memory
4799: @cindex memory words
4800:
4801: @menu
4802: * Memory model::
4803: * Dictionary allocation::
4804: * Heap Allocation::
4805: * Memory Access::
4806: * Address arithmetic::
4807: * Memory Blocks::
4808: @end menu
4809:
4810: @node Memory model, Dictionary allocation, Memory, Memory
4811: @subsection ANS Forth and Gforth memory models
4812:
4813: @c The ANS Forth description is a mess (e.g., is the heap part of
4814: @c the dictionary?), so let's not stick to closely with it.
4815:
4816: ANS Forth considers a Forth system as consisting of several memories, of
4817: which only @dfn{data space} is managed and accessible with the memory
4818: words. Memory not necessarily in data space includes the stacks, the
4819: code (called code space) and the headers (called name space). In Gforth
4820: everything is in data space, but the code for the primitives is usually
4821: read-only.
4822:
4823: Data space is divided into a number of areas: The (data space portion of
4824: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
4825: refer to the search data structure embodied in word lists and headers,
4826: because it is used for looking up names, just as you would in a
4827: conventional dictionary.}, the heap, and a number of system-allocated
4828: buffers.
4829:
4830: In ANS Forth data space is also divided into contiguous regions. You
4831: can only use address arithmetic within a contiguous region, not between
4832: them. Usually each allocation gives you one contiguous region, but the
4833: dictionary allocation words have additional rules (@pxref{Dictionary
4834: allocation}).
4835:
4836: Gforth provides one big address space, and address arithmetic can be
4837: performed between any addresses. However, in the dictionary headers or
4838: code are interleaved with data, so almost the only contiguous data space
4839: regions there are those described by ANS Forth as contiguous; but you
4840: can be sure that the dictionary is allocated towards increasing
4841: addresses even between contiguous regions. The memory order of
4842: allocations in the heap is platform-dependent (and possibly different
4843: from one run to the next).
4844:
4845: @subsubsection ANS Forth dictionary details
4846:
4847: This section is just informative, you can skip it if you are in a hurry.
4848:
4849: When you create a colon definition, the text interpreter compiles the
4850: code for the definition into the code space and compiles the name
4851: of the definition into the header space, together with other
4852: information about the definition (such as its execution token).
4853:
4854: When you create a variable, the execution of @code{Variable} will
4855: compile some code, assign one cell in data space, and compile the name
4856: of the variable into the header space.
4857:
4858: @cindex memory regions - relationship between them
4859: ANS Forth does not specify the relationship between the three memory
4860: regions, and specifies that a Standard program must not access code or
4861: data space directly -- it may only access data space directly. In
4862: addition, the Standard defines what relationships you may and may not
4863: rely on when allocating regions in data space. These constraints are
4864: simply a reflection of the many diverse techniques that are used to
4865: implement Forth systems; understanding and following the requirements of
4866: the Standard allows you to write portable programs -- programs that run
4867: in the same way on any of these diverse systems. Another way of looking
4868: at this is to say that ANS Forth was designed to permit compliant Forth
4869: systems to be implemented in many diverse ways.
4870:
4871: @cindex memory regions - how they are assigned
4872: Here are some examples of ways in which name, code and data spaces
4873: might be assigned in different Forth implementations:
4874:
4875: @itemize @bullet
4876: @item
4877: For a Forth system that runs from RAM under a general-purpose operating
4878: system, it can be convenient to interleave name, code and data spaces in
4879: a single contiguous memory region. This organisation can be
4880: memory-efficient (for example, because the relationship between the name
4881: dictionary entry and the associated code space entry can be
4882: implicit, rather than requiring an explicit memory pointer to reference
4883: from the header space and the code space). This is the
4884: organisation used by Gforth, as this example@footnote{The addresses
4885: in the example have been truncated to fit it onto the page, and the
4886: addresses and data shown will not match the output from your system} shows:
4887: @example
4888: hex
4889: variable fred 123456 fred !
4890: variable jim abcd jim !
4891: : foo + / - ;
4892: ' fred 10 - 50 dump
4893: ..80: 5C 46 0E 40 84 66 72 65 - 64 20 20 20 20 20 20 20 \F.@.fred
4894: ..90: D0 9B 04 08 00 00 00 00 - 56 34 12 00 80 46 0E 40 ........V4...F.@@
4895: ..A0: 83 6A 69 6D 20 20 20 20 - D0 9B 04 08 00 00 00 00 .jim ........
4896: ..B0: CD AB 00 00 9C 46 0E 40 - 83 66 6F 6F 20 20 20 20 .....F.@.foo
4897: ..C0: 80 9B 04 08 00 00 00 00 - E4 2E 05 08 0C 2F 05 08 ............./..
4898: @end example
4899:
4900: @item
4901: For a high-performance system running on a modern RISC processor with a
4902: modified Harvard architecture (one that has a unified main memory but
4903: separate instruction and data caches), it is desirable to separate
4904: processor instructions from processor data. This encourages a high cache
4905: density and therefore a high cache hit rate. The Forth code space
4906: is not necessarily made up entirely of processor instructions; its
4907: nature is dependent upon the Forth implementation.
4908:
4909: @item
4910: A Forth compiler that runs on a segmented 8086 processor could be
4911: designed to interleave the name, code and data spaces within a single
4912: 64Kbyte segment. A more common implementation choice is to use a
4913: separate 64Kbyte segment for each region, which provides more memory
4914: overall but provides an address map in which only the data space is
4915: accessible.
4916:
4917: @item
4918: Microprocessors exist that run Forth (or many of the primitives required
4919: to implement the Forth virtual machine efficiently) directly. On these
4920: processors, the relationship between name, code and data spaces may be
4921: imposed as a side-effect of the architecture of the processor.
4922:
4923: @item
4924: A Forth compiler that executes from ROM on an embedded system needs its
4925: data space separated from the name and code spaces so that the data
4926: space can be mapped to a RAM area.
4927:
4928: @item
4929: A Forth compiler that runs on an embedded system may have a requirement
4930: for a small memory footprint. On such a system it can be useful to
4931: separate the header space from the data and code spaces; once the
4932: application has been compiled, the header space is no longer
4933: required@footnote{more strictly speaking, most applications can be
4934: designed so that this is the case}. The header space can be deleted
4935: entirely, or could be stored in memory on a remote @i{host} system for
4936: debug and development purposes. In the latter case, the compiler running
4937: on the @i{target} system could implement a protocol across a
4938: communication link that would allow it to interrogate the header space.
4939: @end itemize
4940:
4941: @node Dictionary allocation, Heap Allocation, Memory model, Memory
4942: @subsection Dictionary allocation
4943: @cindex reserving data space
4944: @cindex data space - reserving some
4945:
4946: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
4947: you want to deallocate X, you also deallocate everything
4948: allocated after X.
4949:
4950: The allocations using the words below are contiguous and grow the region
4951: towards increasing addresses. Other words that allocate dictionary
4952: memory of any kind (i.e., defining words including @code{:noname}) end
4953: the contiguous region and start a new one.
4954:
4955: In ANS Forth only @code{create}d words are guaranteed to produce an
4956: address that is the start of the following contiguous region. In
4957: particular, the cell allocated by @code{variable} is not guaranteed to
4958: be contiguous with following @code{allot}ed memory.
4959:
4960: You can deallocate memory by using @code{allot} with a negative argument
4961: (with some restrictions, see @code{allot}). For larger deallocations use
4962: @code{marker}.
4963:
4964:
4965: doc-here
4966: doc-unused
4967: doc-allot
4968: doc-c,
4969: doc-f,
4970: doc-,
4971: doc-2,
4972: @cindex user space
4973: doc-udp
4974: doc-uallot
4975:
4976: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
4977: course you should allocate memory in an aligned way, too. I.e., before
4978: allocating allocating a cell, @code{here} must be cell-aligned, etc.
4979: The words below align @code{here} if it is not already. Basically it is
4980: only already aligned for a type, if the last allocation was a multiple
4981: of the size of this type and if @code{here} was aligned for this type
4982: before.
4983:
4984: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
4985: ANS Forth (@code{maxalign}ed in Gforth).
4986:
4987: doc-align
4988: doc-falign
4989: doc-sfalign
4990: doc-dfalign
4991: doc-maxalign
4992: doc-cfalign
4993:
4994:
4995: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
4996: @subsection Heap allocation
4997: @cindex heap allocation
4998: @cindex dynamic allocation of memory
4999: @cindex memory-allocation word set
5000:
5001: Heap allocation supports deallocation of allocated memory in any
5002: order. Dictionary allocation is not affected by it (i.e., it does not
5003: end a contiguous region). In Gforth, these words are implemented using
5004: the standard C library calls malloc(), free() and resize().
5005:
5006: doc-allocate
5007: doc-free
5008: doc-resize
5009:
5010:
5011: @node Memory Access, Address arithmetic, Heap Allocation, Memory
5012: @subsection Memory Access
5013: @cindex memory access words
5014:
5015:
5016: doc-@
5017: doc-!
5018: doc-+!
5019: doc-c@
5020: doc-c!
5021: doc-2@
5022: doc-2!
5023: doc-f@
5024: doc-f!
5025: doc-sf@
5026: doc-sf!
5027: doc-df@
5028: doc-df!
5029:
5030: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5031: @subsection Address arithmetic
5032: @cindex address arithmetic words
5033:
5034: Address arithmetic is the foundation on which data structures like
5035: arrays, records (@pxref{Structures}) and objects (@pxref{Object-oriented
5036: Forth}) are built.
5037:
5038: ANS Forth does not specify the sizes of the data types. Instead, it
5039: offers a number of words for computing sizes and doing address
5040: arithmetic. Address arithmetic is performed in terms of address units
5041: (aus); on most systems the address unit is one byte. Note that a
5042: character may have more than one au, so @code{chars} is no noop (on
5043: systems where it is a noop, it compiles to nothing).
5044:
5045: @cindex alignment of addresses for types
5046: ANS Forth also defines words for aligning addresses for specific
5047: types. Many computers require that accesses to specific data types
5048: must only occur at specific addresses; e.g., that cells may only be
5049: accessed at addresses divisible by 4. Even if a machine allows unaligned
5050: accesses, it can usually perform aligned accesses faster.
5051:
5052: For the performance-conscious: alignment operations are usually only
5053: necessary during the definition of a data structure, not during the
5054: (more frequent) accesses to it.
5055:
5056: ANS Forth defines no words for character-aligning addresses. This is not
5057: an oversight, but reflects the fact that addresses that are not
5058: char-aligned have no use in the standard and therefore will not be
5059: created.
5060:
5061: @cindex @code{CREATE} and alignment
5062: ANS Forth guarantees that addresses returned by @code{CREATE}d words
5063: are cell-aligned; in addition, Gforth guarantees that these addresses
5064: are aligned for all purposes.
5065:
5066: Note that the ANS Forth word @code{char} has nothing to do with address
5067: arithmetic.
5068:
5069:
5070: doc-chars
5071: doc-char+
5072: doc-cells
5073: doc-cell+
5074: doc-cell
5075: doc-aligned
5076: doc-floats
5077: doc-float+
5078: doc-float
5079: doc-faligned
5080: doc-sfloats
5081: doc-sfloat+
5082: doc-sfaligned
5083: doc-dfloats
5084: doc-dfloat+
5085: doc-dfaligned
5086: doc-maxaligned
5087: doc-cfaligned
5088: doc-address-unit-bits
5089:
5090:
5091: @node Memory Blocks, , Address arithmetic, Memory
5092: @subsection Memory Blocks
5093: @cindex memory block words
5094: @cindex character strings - moving and copying
5095:
5096: Memory blocks often represent character strings; For ways of storing
5097: character strings in memory see @ref{String Formats}. For other
5098: string-processing words see @ref{Displaying characters and strings}.
5099:
5100: Some of these words work on address units. Others work on character
5101: units (increments of @code{CHAR}), and expect a @code{CHAR}-aligned
5102: address. Choose the correct operation depending upon your data type.
5103:
5104: When copying characters between overlapping memory regions, choose
5105: carefully between @code{cmove} and @code{cmove>}.
5106:
5107: You can only use any of these words @i{portably} to access data space.
5108:
5109: @comment TODO - think the naming of the arguments is wrong for move
5110: @comment well, really it seems to be the Standard that's wrong; it
5111: @comment describes MOVE as a word that requires a CELL-aligned source
5112: @comment and destination address but a xtranfer count that need not
5113: @comment be a multiple of CELL.
5114:
5115: doc-move
5116: doc-erase
5117: doc-cmove
5118: doc-cmove>
5119: doc-fill
5120: doc-blank
5121: doc-compare
5122: doc-search
5123: doc--trailing
5124: doc-/string
5125:
5126:
5127: @comment TODO examples
5128:
5129:
5130: @node Control Structures, Defining Words, Memory, Words
5131: @section Control Structures
5132: @cindex control structures
5133:
5134: Control structures in Forth cannot be used interpretively, only in a
5135: colon definition@footnote{To be precise, they have no interpretation
5136: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5137: not like this limitation, but have not seen a satisfying way around it
5138: yet, although many schemes have been proposed.
5139:
5140: @menu
5141: * Selection:: IF ... ELSE ... ENDIF
5142: * Simple Loops:: BEGIN ...
5143: * Counted Loops:: DO
5144: * Arbitrary control structures::
5145: * Calls and returns::
5146: * Exception Handling::
5147: @end menu
5148:
5149: @node Selection, Simple Loops, Control Structures, Control Structures
5150: @subsection Selection
5151: @cindex selection control structures
5152: @cindex control structures for selection
5153:
5154: @c what's the purpose of all these @i? Maybe we should define a macro
5155: @c so we can produce logical markup. - anton
5156:
5157: @c nac-> When I started working on the manual, a mixture of @i and @var
5158: @c were used inconsistently in code examples and \Glossary entries. These
5159: @c two behave differently in info format so I decided to standardize on @i.
5160: @c Logical markup would be better but texi isn't really upto it, and
5161: @c texi2html just ignores macros.
5162: @c nac02dec1999-> update: the latest texinfo release can spit out html
5163: @c and it handles macros, so we could do some logical markup. Unfortunately
5164: @c texinfo will not split html output, which would be a big pain if you
5165: @c wanted to put the document on the web, which would be nice.
5166:
5167: @cindex @code{IF} control structure
5168: @example
5169: @i{flag}
5170: IF
5171: @i{code}
5172: ENDIF
5173: @end example
5174: @noindent
5175:
5176: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5177: with any bit set represents truth) @i{code} is executed.
5178:
5179: @example
5180: @i{flag}
5181: IF
5182: @i{code1}
5183: ELSE
5184: @i{code2}
5185: ENDIF
5186: @end example
5187:
5188: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5189: executed.
5190:
5191: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5192: standard, and @code{ENDIF} is not, although it is quite popular. We
5193: recommend using @code{ENDIF}, because it is less confusing for people
5194: who also know other languages (and is not prone to reinforcing negative
5195: prejudices against Forth in these people). Adding @code{ENDIF} to a
5196: system that only supplies @code{THEN} is simple:
5197: @example
5198: : ENDIF POSTPONE THEN ; immediate
5199: @end example
5200:
5201: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5202: (adv.)} has the following meanings:
5203: @quotation
5204: ... 2b: following next after in order ... 3d: as a necessary consequence
5205: (if you were there, then you saw them).
5206: @end quotation
5207: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5208: and many other programming languages has the meaning 3d.]
5209:
5210: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
5211: you can avoid using @code{?dup}. Using these alternatives is also more
5212: efficient than using @code{?dup}. Definitions in ANS Forth
5213: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5214: @file{compat/control.fs}.
5215:
5216: @cindex @code{CASE} control structure
5217: @example
5218: @i{n}
5219: CASE
5220: @i{n1} OF @i{code1} ENDOF
5221: @i{n2} OF @i{code2} ENDOF
5222: @dots{}
5223: ENDCASE
5224: @end example
5225:
5226: Executes the first @i{codei}, where the @i{ni} is equal to
5227: @i{n}. A default case can be added by simply writing the code after
5228: the last @code{ENDOF}. It may use @i{n}, which is on top of the stack,
5229: but must not consume it.
5230:
5231: @node Simple Loops, Counted Loops, Selection, Control Structures
5232: @subsection Simple Loops
5233: @cindex simple loops
5234: @cindex loops without count
5235:
5236: @cindex @code{WHILE} loop
5237: @example
5238: BEGIN
5239: @i{code1}
5240: @i{flag}
5241: WHILE
5242: @i{code2}
5243: REPEAT
5244: @end example
5245:
5246: @i{code1} is executed and @i{flag} is computed. If it is true,
5247: @i{code2} is executed and the loop is restarted; If @i{flag} is
5248: false, execution continues after the @code{REPEAT}.
5249:
5250: @cindex @code{UNTIL} loop
5251: @example
5252: BEGIN
5253: @i{code}
5254: @i{flag}
5255: UNTIL
5256: @end example
5257:
5258: @i{code} is executed. The loop is restarted if @code{flag} is false.
5259:
5260: @cindex endless loop
5261: @cindex loops, endless
5262: @example
5263: BEGIN
5264: @i{code}
5265: AGAIN
5266: @end example
5267:
5268: This is an endless loop.
5269:
5270: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5271: @subsection Counted Loops
5272: @cindex counted loops
5273: @cindex loops, counted
5274: @cindex @code{DO} loops
5275:
5276: The basic counted loop is:
5277: @example
5278: @i{limit} @i{start}
5279: ?DO
5280: @i{body}
5281: LOOP
5282: @end example
5283:
5284: This performs one iteration for every integer, starting from @i{start}
5285: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
5286: accessed with @code{i}. For example, the loop:
5287: @example
5288: 10 0 ?DO
5289: i .
5290: LOOP
5291: @end example
5292: @noindent
5293: prints @code{0 1 2 3 4 5 6 7 8 9}
5294:
5295: The index of the innermost loop can be accessed with @code{i}, the index
5296: of the next loop with @code{j}, and the index of the third loop with
5297: @code{k}.
5298:
5299:
5300: doc-i
5301: doc-j
5302: doc-k
5303:
5304:
5305: The loop control data are kept on the return stack, so there are some
5306: restrictions on mixing return stack accesses and counted loop words. In
5307: particuler, if you put values on the return stack outside the loop, you
5308: cannot read them inside the loop@footnote{well, not in a way that is
5309: portable.}. If you put values on the return stack within a loop, you
5310: have to remove them before the end of the loop and before accessing the
5311: index of the loop.
5312:
5313: There are several variations on the counted loop:
5314:
5315: @itemize @bullet
5316: @item
5317: @code{LEAVE} leaves the innermost counted loop immediately; execution
5318: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5319:
5320: @example
5321: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5322: @end example
5323: prints @code{0 1 2 3}
5324:
5325:
5326: @item
5327: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5328: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5329: return stack so @code{EXIT} can get to its return address. For example:
5330:
5331: @example
5332: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5333: @end example
5334: prints @code{0 1 2 3}
5335:
5336:
5337: @item
5338: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
5339: (and @code{LOOP} iterates until they become equal by wrap-around
5340: arithmetic). This behaviour is usually not what you want. Therefore,
5341: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
5342: @code{?DO}), which do not enter the loop if @i{start} is greater than
5343: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
5344: unsigned loop parameters.
5345:
5346: @item
5347: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5348: the loop, independent of the loop parameters. Do not use @code{DO}, even
5349: if you know that the loop is entered in any case. Such knowledge tends
5350: to become invalid during maintenance of a program, and then the
5351: @code{DO} will make trouble.
5352:
5353: @item
5354: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5355: index by @i{n} instead of by 1. The loop is terminated when the border
5356: between @i{limit-1} and @i{limit} is crossed. E.g.:
5357:
5358: @example
5359: 4 0 +DO i . 2 +LOOP
5360: @end example
5361: @noindent
5362: prints @code{0 2}
5363:
5364: @example
5365: 4 1 +DO i . 2 +LOOP
5366: @end example
5367: @noindent
5368: prints @code{1 3}
5369:
5370:
5371: @cindex negative increment for counted loops
5372: @cindex counted loops with negative increment
5373: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
5374:
5375: @example
5376: -1 0 ?DO i . -1 +LOOP
5377: @end example
5378: @noindent
5379: prints @code{0 -1}
5380:
5381: @example
5382: 0 0 ?DO i . -1 +LOOP
5383: @end example
5384: prints nothing.
5385:
5386: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5387: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5388: index by @i{u} each iteration. The loop is terminated when the border
5389: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
5390: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5391:
5392: @example
5393: -2 0 -DO i . 1 -LOOP
5394: @end example
5395: @noindent
5396: prints @code{0 -1}
5397:
5398: @example
5399: -1 0 -DO i . 1 -LOOP
5400: @end example
5401: @noindent
5402: prints @code{0}
5403:
5404: @example
5405: 0 0 -DO i . 1 -LOOP
5406: @end example
5407: @noindent
5408: prints nothing.
5409:
5410: @end itemize
5411:
5412: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
5413: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5414: for these words that uses only standard words is provided in
5415: @file{compat/loops.fs}.
5416:
5417:
5418: @cindex @code{FOR} loops
5419: Another counted loop is:
5420: @example
5421: @i{n}
5422: FOR
5423: @i{body}
5424: NEXT
5425: @end example
5426: This is the preferred loop of native code compiler writers who are too
5427: lazy to optimize @code{?DO} loops properly. This loop structure is not
5428: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5429: @code{i} produces values starting with @i{n} and ending with 0. Other
5430: Forth systems may behave differently, even if they support @code{FOR}
5431: loops. To avoid problems, don't use @code{FOR} loops.
5432:
5433: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5434: @subsection Arbitrary control structures
5435: @cindex control structures, user-defined
5436:
5437: @cindex control-flow stack
5438: ANS Forth permits and supports using control structures in a non-nested
5439: way. Information about incomplete control structures is stored on the
5440: control-flow stack. This stack may be implemented on the Forth data
5441: stack, and this is what we have done in Gforth.
5442:
5443: @cindex @code{orig}, control-flow stack item
5444: @cindex @code{dest}, control-flow stack item
5445: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5446: entry represents a backward branch target. A few words are the basis for
5447: building any control structure possible (except control structures that
5448: need storage, like calls, coroutines, and backtracking).
5449:
5450:
5451: doc-if
5452: doc-ahead
5453: doc-then
5454: doc-begin
5455: doc-until
5456: doc-again
5457: doc-cs-pick
5458: doc-cs-roll
5459:
5460:
5461: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5462: manipulate the control-flow stack in a portable way. Without them, you
5463: would need to know how many stack items are occupied by a control-flow
5464: entry (many systems use one cell. In Gforth they currently take three,
5465: but this may change in the future).
5466:
5467: Some standard control structure words are built from these words:
5468:
5469:
5470: doc-else
5471: doc-while
5472: doc-repeat
5473:
5474:
5475: @noindent
5476: Gforth adds some more control-structure words:
5477:
5478:
5479: doc-endif
5480: doc-?dup-if
5481: doc-?dup-0=-if
5482:
5483:
5484: @noindent
5485: Counted loop words constitute a separate group of words:
5486:
5487:
5488: doc-?do
5489: doc-+do
5490: doc-u+do
5491: doc--do
5492: doc-u-do
5493: doc-do
5494: doc-for
5495: doc-loop
5496: doc-+loop
5497: doc--loop
5498: doc-next
5499: doc-leave
5500: doc-?leave
5501: doc-unloop
5502: doc-done
5503:
5504:
5505: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5506: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
5507: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5508: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5509: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5510: resolved (by using one of the loop-ending words or @code{DONE}).
5511:
5512: @noindent
5513: Another group of control structure words are:
5514:
5515:
5516: doc-case
5517: doc-endcase
5518: doc-of
5519: doc-endof
5520:
5521:
5522: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5523: @code{CS-ROLL}.
5524:
5525: @subsubsection Programming Style
5526: @cindex control structures programming style
5527: @cindex programming style, arbitrary control structures
5528:
5529: In order to ensure readability we recommend that you do not create
5530: arbitrary control structures directly, but define new control structure
5531: words for the control structure you want and use these words in your
5532: program. For example, instead of writing:
5533:
5534: @example
5535: BEGIN
5536: ...
5537: IF [ 1 CS-ROLL ]
5538: ...
5539: AGAIN THEN
5540: @end example
5541:
5542: @noindent
5543: we recommend defining control structure words, e.g.,
5544:
5545: @example
5546: : WHILE ( DEST -- ORIG DEST )
5547: POSTPONE IF
5548: 1 CS-ROLL ; immediate
5549:
5550: : REPEAT ( orig dest -- )
5551: POSTPONE AGAIN
5552: POSTPONE THEN ; immediate
5553: @end example
5554:
5555: @noindent
5556: and then using these to create the control structure:
5557:
5558: @example
5559: BEGIN
5560: ...
5561: WHILE
5562: ...
5563: REPEAT
5564: @end example
5565:
5566: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5567: @code{WHILE} are predefined, so in this example it would not be
5568: necessary to define them.
5569:
5570: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5571: @subsection Calls and returns
5572: @cindex calling a definition
5573: @cindex returning from a definition
5574:
5575: @cindex recursive definitions
5576: A definition can be called simply be writing the name of the definition
5577: to be called. Normally a definition is invisible during its own
5578: definition. If you want to write a directly recursive definition, you
5579: can use @code{recursive} to make the current definition visible, or
5580: @code{recurse} to call the current definition directly.
5581:
5582:
5583: doc-recursive
5584: doc-recurse
5585:
5586:
5587: @comment TODO add example of the two recursion methods
5588: @quotation
5589: @progstyle
5590: I prefer using @code{recursive} to @code{recurse}, because calling the
5591: definition by name is more descriptive (if the name is well-chosen) than
5592: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5593: implementation, it is much better to read (and think) ``now sort the
5594: partitions'' than to read ``now do a recursive call''.
5595: @end quotation
5596:
5597: For mutual recursion, use @code{Defer}red words, like this:
5598:
5599: @example
5600: Defer foo
5601:
5602: : bar ( ... -- ... )
5603: ... foo ... ;
5604:
5605: :noname ( ... -- ... )
5606: ... bar ... ;
5607: IS foo
5608: @end example
5609:
5610: Deferred words are discussed in more detail in @ref{Deferred words}.
5611:
5612: The current definition returns control to the calling definition when
5613: the end of the definition is reached or @code{EXIT} is encountered.
5614:
5615: doc-exit
5616: doc-;s
5617:
5618:
5619: @node Exception Handling, , Calls and returns, Control Structures
5620: @subsection Exception Handling
5621: @cindex exceptions
5622:
5623: If your program detects a fatal error condition, the simplest action
5624: that it can take is to @code{quit}. This resets the return stack and
5625: restarts the text interpreter, but does not print any error message.
5626:
5627: The next stage in severity is to execute @code{abort}, which has the
5628: same effect as @code{quit}, with the addition that it resets the data
5629: stack.
5630:
5631: A slightly more sophisticated approach is use use @code{abort"}, which
5632: compiles a string to be used as an error message and does a conditional
5633: @code{abort} at run-time. For example:
5634:
5635: @example
5636: @kbd{: checker abort" That flag was true" ." A false flag" ;@key{RET}} ok
5637: @kbd{0 checker@key{RET}} A false flag ok
5638: @kbd{1 checker@key{RET}}
5639: :1: That flag was true
5640: 1 checker
5641: ^^^^^^^
5642: $400D1648 throw
5643: $400E4660
5644: @end example
5645:
5646: These simple techniques allow a program to react to a fatal error
5647: condition, but they are not exactly user-friendly. The ANS Forth
5648: Exception word set provides the pair of words @code{throw} and
5649: @code{catch}, which can be used to provide sophisticated error-handling.
5650:
5651: @code{catch} has a similar behaviour to @code{execute}, in that it takes
5652: an @i{xt} as a parameter and starts execution of the xt. However,
5653: before passing control to the xt, @code{catch} pushes an
5654: @dfn{exception frame} onto the @dfn{exception stack}. This exception
5655: frame is used to restore the system to a known state if a detected error
5656: occurs during the execution of the xt. A typical way to use @code{catch}
5657: would be:
5658:
5659: @example
5660: ... ['] foo catch IF ...
5661: @end example
5662:
5663: @c TOS is undefined. - anton
5664:
5665: @c nac-> TODO -- I need to look at this example again.
5666:
5667: Whilst @code{foo} executes, it can call other words to any level of
5668: nesting, as usual. If @code{foo} (and all the words that it calls)
5669: execute successfully, control will ultimately pass to the word following
5670: the @code{catch}, and there will be a 0 at TOS. However, if any word
5671: detects an error, it can terminate the execution of @code{foo} by
5672: pushing a non-zero error code onto the stack and then performing a
5673: @code{throw}. The execution of @code{throw} will pass control to the
5674: word following the @code{catch}, but this time the TOS will hold the
5675: error code. Therefore, the @code{IF} in the example can be used to
5676: determine whether @code{foo} executed successfully.
5677:
5678: This simple example shows how you can use @code{throw} and @code{catch}
5679: to ``take over'' exception handling from the system:
5680: @example
5681: : my-div ['] / catch if ." DIVIDE ERROR" else ." OK.. " . then ;
5682: @end example
5683:
5684: The next example is more sophisticated and shows a multi-level
5685: @code{throw} and @code{catch}. To understand this example, start at the
5686: definition of @code{top-level} and work backwards:
5687:
5688: @example
5689: : lowest-level ( -- c )
5690: key dup 27 = if
5691: 1 throw \ ESCAPE key pressed
5692: else
5693: ." lowest-level successful" CR
5694: then
5695: ;
5696:
5697: : lower-level ( -- c )
5698: lowest-level
5699: \ at this level consider a CTRL-U to be a fatal error
5700: dup 21 = if \ CTRL-U
5701: 2 throw
5702: else
5703: ." lower-level successful" CR
5704: then
5705: ;
5706:
5707: : low-level ( -- c )
5708: ['] lower-level catch
5709: ?dup if
5710: \ error occurred - do we recognise it?
5711: dup 1 = if
5712: \ ESCAPE key pressed.. pretend it was an E
5713: [char] E
5714: else throw \ propogate the error upwards
5715: then
5716: then
5717: ." low-level successfull" CR
5718: ;
5719:
5720: : top-level ( -- )
5721: CR ['] low-level catch \ CATCH is used like EXECUTE
5722: ?dup if \ error occurred..
5723: ." Error " . ." occurred - contact your supplier"
5724: else
5725: ." The '" emit ." ' key was pressed" CR
5726: then
5727: ;
5728: @end example
5729:
5730: The ANS Forth document assigns @code{throw} codes thus:
5731:
5732: @itemize @bullet
5733: @item
5734: codes in the range -1 -- -255 are reserved to be assigned by the
5735: Standard. Assignments for codes in the range -1 -- -58 are currently
5736: documented in the Standard. In particular, @code{-1 throw} is equivalent
5737: to @code{abort} and @code{-2 throw} is equivalent to @code{abort"}.
5738: @item
5739: codes in the range -256 -- -4095 are reserved to be assigned by the system.
5740: @item
5741: all other codes may be assigned by programs.
5742: @end itemize
5743:
5744: Gforth provides the word @code{exception} as a mechanism for assigning
5745: system throw codes to applications. This allows multiple applications to
5746: co-exist in memory without any clash of @code{throw} codes. A definition
5747: of @code{exception} in ANS Forth is provided in
5748: @file{compat/exception.fs}.
5749:
5750:
5751: doc-quit
5752: doc-abort
5753: doc-abort"
5754:
5755: doc-catch
5756: doc-throw
5757: doc---exception-exception
5758:
5759:
5760:
5761: @c -------------------------------------------------------------
5762: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
5763: @section Defining Words
5764: @cindex defining words
5765:
5766: Defining words are used to extend Forth by creating new entries in the dictionary.
5767:
5768: @menu
5769: * CREATE::
5770: * Variables:: Variables and user variables
5771: * Constants::
5772: * Values:: Initialised variables
5773: * Colon Definitions::
5774: * Anonymous Definitions:: Definitions without names
5775: * User-defined Defining Words::
5776: * Deferred words:: Allow forward references
5777: * Aliases::
5778: * Supplying names::
5779: @end menu
5780:
5781: @node CREATE, Variables, Defining Words, Defining Words
5782: @subsection @code{CREATE}
5783: @cindex simple defining words
5784: @cindex defining words, simple
5785:
5786: Defining words are used to create new entries in the dictionary. The
5787: simplest defining word is @code{CREATE}. @code{CREATE} is used like
5788: this:
5789:
5790: @example
5791: CREATE new-word1
5792: @end example
5793:
5794: @code{CREATE} is a parsing word that generates a dictionary entry for
5795: @code{new-word1}. When @code{new-word1} is executed, all that it does is
5796: leave an address on the stack. The address represents the value of
5797: the data space pointer (@code{HERE}) at the time that @code{new-word1}
5798: was defined. Therefore, @code{CREATE} is a way of associating a name
5799: with the address of a region of memory.
5800:
5801: doc-create
5802:
5803: By extending this example to reserve some memory in data space, we end
5804: up with a @i{variable}. Here are two different ways to do it:
5805:
5806: @example
5807: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
5808: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
5809: @end example
5810:
5811: The variable can be examined and modified using @code{@@} (``fetch'') and
5812: @code{!} (``store'') like this:
5813:
5814: @example
5815: new-word2 @@ . \ get address, fetch from it and display
5816: 1234 new-word2 ! \ new value, get address, store to it
5817: @end example
5818:
5819: @cindex arrays
5820: A similar mechanism can be used to create arrays. For example, an
5821: 80-character text input buffer:
5822:
5823: @example
5824: CREATE text-buf 80 chars allot
5825:
5826: text-buf 0 chars c@@ \ the 1st character (offset 0)
5827: text-buf 3 chars c@@ \ the 4th character (offset 3)
5828: @end example
5829:
5830: You can build arbitrarily complex data structures by allocating
5831: appropriate areas of memory. For further discussions of this, and to
5832: learn about some Gforth tools that make it easier, see
5833: @xref{Structures}.
5834:
5835:
5836: @node Variables, Constants, CREATE, Defining Words
5837: @subsection Variables
5838: @cindex variables
5839:
5840: The previous section showed how a sequence of commands could be used to
5841: generate a variable. As a final refinement, the whole code sequence can
5842: be wrapped up in a defining word (pre-empting the subject of the next
5843: section), making it easier to create new variables:
5844:
5845: @example
5846: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
5847: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
5848:
5849: myvariableX foo \ variable foo starts off with an unknown value
5850: myvariable0 joe \ whilst joe is initialised to 0
5851:
5852: 45 3 * foo ! \ set foo to 135
5853: 1234 joe ! \ set joe to 1234
5854: 3 joe +! \ increment joe by 3.. to 1237
5855: @end example
5856:
5857: Not surprisingly, there is no need to define @code{myvariable}, since
5858: Forth already has a definition @code{Variable}. ANS Forth does not
5859: require a @code{Variable} to be initialised when it is created (i.e., it
5860: behaves like @code{myvariableX}). In contrast, Gforth's @code{Variable}
5861: initialises the variable to 0 (i.e., it behaves exactly like
5862: @code{myvariable0}). Forth also provides @code{2Variable} and
5863: @code{fvariable} for double and floating-point variables, respectively
5864: -- both are initialised to 0 in Gforth. If you use a @code{Variable} to
5865: store a boolean, you can use @code{on} and @code{off} to toggle its
5866: state.
5867:
5868: doc-variable
5869: doc-2variable
5870: doc-fvariable
5871:
5872: @cindex user variables
5873: @cindex user space
5874: The defining word @code{User} behaves in the same way as @code{Variable}.
5875: The difference is that it reserves space in @i{user (data) space} rather
5876: than normal data space. In a Forth system that has a multi-tasker, each
5877: task has its own set of user variables.
5878:
5879: doc-user
5880:
5881: @comment TODO is that stuff about user variables strictly correct? Is it
5882: @comment just terminal tasks that have user variables?
5883: @comment should document tasker.fs (with some examples) elsewhere
5884: @comment in this manual, then expand on user space and user variables.
5885:
5886:
5887: @node Constants, Values, Variables, Defining Words
5888: @subsection Constants
5889: @cindex constants
5890:
5891: @code{Constant} allows you to declare a fixed value and refer to it by
5892: name. For example:
5893:
5894: @example
5895: 12 Constant INCHES-PER-FOOT
5896: 3E+08 fconstant SPEED-O-LIGHT
5897: @end example
5898:
5899: A @code{Variable} can be both read and written, so its run-time
5900: behaviour is to supply an address through which its current value can be
5901: manipulated. In contrast, the value of a @code{Constant} cannot be
5902: changed once it has been declared@footnote{Well, often it can be -- but
5903: not in a Standard, portable way. It's safer to use a @code{Value} (read
5904: on).} so it's not necessary to supply the address -- it is more
5905: efficient to return the value of the constant directly. That's exactly
5906: what happens; the run-time effect of a constant is to put its value on
5907: the top of the stack (You can find one
5908: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
5909:
5910: Gforth also provides @code{2Constant} and @code{fconstant} for defining
5911: double and floating-point constants, respectively.
5912:
5913: doc-constant
5914: doc-2constant
5915: doc-fconstant
5916:
5917: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
5918: @c nac-> How could that not be true in an ANS Forth? You can't define a
5919: @c constant, use it and then delete the definition of the constant..
5920: @c I agree that it's rather deep, but IMO it is an important difference
5921: @c relative to other programming languages.. often it's annoying: it
5922: @c certainly changes my programming style relative to C.
5923:
5924: Constants in Forth behave differently from their equivalents in other
5925: programming languages. In other languages, a constant (such as an EQU in
5926: assembler or a #define in C) only exists at compile-time; in the
5927: executable program the constant has been translated into an absolute
5928: number and, unless you are using a symbolic debugger, it's impossible to
5929: know what abstract thing that number represents. In Forth a constant has
5930: an entry in the header space and remains there after the code that uses
5931: it has been defined. In fact, it must remain in the dictionary since it
5932: has run-time duties to perform. For example:
5933:
5934: @example
5935: 12 Constant INCHES-PER-FOOT
5936: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
5937: @end example
5938:
5939: @cindex in-lining of constants
5940: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
5941: associated with the constant @code{INCHES-PER-FOOT}. If you use
5942: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
5943: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
5944: attempt to optimise constants by in-lining them where they are used. You
5945: can force Gforth to in-line a constant like this:
5946:
5947: @example
5948: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
5949: @end example
5950:
5951: If you use @code{see} to decompile @i{this} version of
5952: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
5953: longer present. To understand how this works, read
5954: @ref{Interpret/Compile states}, and @ref{Literals}.
5955:
5956: In-lining constants in this way might improve execution time
5957: fractionally, and can ensure that a constant is now only referenced at
5958: compile-time. However, the definition of the constant still remains in
5959: the dictionary. Some Forth compilers provide a mechanism for controlling
5960: a second dictionary for holding transient words such that this second
5961: dictionary can be deleted later in order to recover memory
5962: space. However, there is no standard way of doing this.
5963:
5964:
5965: @node Values, Colon Definitions, Constants, Defining Words
5966: @subsection Values
5967: @cindex values
5968:
5969: A @code{Value} is like a @code{Variable} but with two important
5970: differences:
5971:
5972: @itemize @bullet
5973: @item
5974: A @code{Value} is initialised when it is declared; like a
5975: @code{Constant} but unlike a @code{Variable}.
5976: @item
5977: A @code{Value} returns its value rather than its address when it is
5978: executed; i.e., it has the same run-time behaviour as @code{Constant}.
5979: @end itemize
5980:
5981: A @code{Value} needs an additional word, @code{TO} to allow its value to
5982: be changed. Here are some examples:
5983:
5984: @example
5985: 12 Value APPLES \ Define APPLES with an initial value of 12
5986: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
5987: APPLES \ puts 34 on the top of the stack.
5988: @end example
5989:
5990: doc-value
5991: doc-to
5992:
5993:
5994: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
5995: @subsection Colon Definitions
5996: @cindex colon definitions
5997:
5998: @example
5999: : name ( ... -- ... )
6000: word1 word2 word3 ;
6001: @end example
6002:
6003: @noindent
6004: Creates a word called @code{name} that, upon execution, executes
6005: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
6006:
6007: The explanation above is somewhat superficial. For simple examples of
6008: colon definitions see @ref{Your first definition}. For an in-depth
6009: discussion of some of the issues involved, see @xref{Interpretation and
6010: Compilation Semantics}.
6011:
6012: doc-:
6013: doc-;
6014:
6015:
6016: @node Anonymous Definitions, User-defined Defining Words, Colon Definitions, Defining Words
6017: @subsection Anonymous Definitions
6018: @cindex colon definitions
6019: @cindex defining words without name
6020:
6021: Sometimes you want to define an @dfn{anonymous word}; a word without a
6022: name. You can do this with:
6023:
6024: doc-:noname
6025:
6026: This leaves the execution token for the word on the stack after the
6027: closing @code{;}. Here's an example in which a deferred word is
6028: initialised with an @code{xt} from an anonymous colon definition:
6029:
6030: @example
6031: Defer deferred
6032: :noname ( ... -- ... )
6033: ... ;
6034: IS deferred
6035: @end example
6036:
6037: @noindent
6038: Gforth provides an alternative way of doing this, using two separate
6039: words:
6040:
6041: doc-noname
6042: @cindex execution token of last defined word
6043: doc-lastxt
6044:
6045: @noindent
6046: The previous example can be rewritten using @code{noname} and
6047: @code{lastxt}:
6048:
6049: @example
6050: Defer deferred
6051: noname : ( ... -- ... )
6052: ... ;
6053: lastxt IS deferred
6054: @end example
6055:
6056: @noindent
6057: @code{noname} works with any defining word, not just @code{:}.
6058:
6059: @code{lastxt} also works when the last word was not defined as
6060: @code{noname}. It also has the useful property that is is valid as soon
6061: as the header for a definition has been built. Thus:
6062:
6063: @example
6064: lastxt . : foo [ lastxt . ] ; ' foo .
6065: @end example
6066:
6067: @noindent
6068: prints 3 numbers; the last two are the same.
6069:
6070:
6071: @node User-defined Defining Words, Deferred words, Anonymous Definitions, Defining Words
6072: @subsection User-defined Defining Words
6073: @cindex user-defined defining words
6074: @cindex defining words, user-defined
6075:
6076: You can create a new defining word by wrapping defining-time code around
6077: an existing defining word and putting the sequence in a colon
6078: definition. For example, suppose that you have a word @code{stats} that
6079: gathers statistics about colon definitions given the @i{xt} of the
6080: definition, and you want every colon definition in your application to
6081: make a call to @code{stats}. You can define and use a new version of
6082: @code{:} like this:
6083:
6084: @example
6085: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6086: ... ; \ other code
6087:
6088: : my: : lastxt postpone literal ['] stats compile, ;
6089:
6090: my: foo + - ;
6091: @end example
6092:
6093: When @code{foo} is defined using @code{my:} these steps occur:
6094:
6095: @itemize @bullet
6096: @item
6097: @code{my:} is executed.
6098: @item
6099: The @code{:} within the definition (the one between @code{my:} and
6100: @code{lastxt}) is executed, and does just what it always does; it parses
6101: the input stream for a name, builds a dictionary header for the name
6102: @code{foo} and switches @code{state} from interpret to compile.
6103: @item
6104: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
6105: being defined -- @code{foo} -- onto the stack.
6106: @item
6107: The code that was produced by @code{postpone literal} is executed; this
6108: causes the value on the stack to be compiled as a literal in the code
6109: area of @code{foo}.
6110: @item
6111: The code @code{['] stats} compiles a literal into the definition of
6112: @code{my:}. When @code{compile,} is executed, that literal -- the
6113: execution token for @code{stats} -- is layed down in the code area of
6114: @code{foo} , following the literal@footnote{Strictly speaking, the
6115: mechanism that @code{compile,} uses to convert an @i{xt} into something
6116: in the code area is implementation-dependent. A threaded implementation
6117: might spit out the execution token directly whilst another
6118: implementation might spit out a native code sequence.}.
6119: @item
6120: At this point, the execution of @code{my:} is complete, and control
6121: returns to the text interpreter. The text interpreter is in compile
6122: state, so subsequent text @code{+ -} is compiled into the definition of
6123: @code{foo} and the @code{;} terminates the definition as always.
6124: @end itemize
6125:
6126: You can use @code{see} to decompile a word that was defined using
6127: @code{my:} and see how it is different from a normal @code{:}
6128: definition. For example:
6129:
6130: @example
6131: : bar + - ; \ like foo but using : rather than my:
6132: see bar
6133: : bar
6134: + - ;
6135: see foo
6136: : foo
6137: 107645672 stats + - ;
6138:
6139: \ use ' stats . to show that 107645672 is the xt for stats
6140: @end example
6141:
6142: You can use techniques like this to make new defining words in terms of
6143: @i{any} existing defining word.
6144:
6145:
6146: @cindex defining defining words
6147: @cindex @code{CREATE} ... @code{DOES>}
6148: If you want the words defined with your defining words to behave
6149: differently from words defined with standard defining words, you can
6150: write your defining word like this:
6151:
6152: @example
6153: : def-word ( "name" -- )
6154: CREATE @i{code1}
6155: DOES> ( ... -- ... )
6156: @i{code2} ;
6157:
6158: def-word name
6159: @end example
6160:
6161: @cindex child words
6162: This fragment defines a @dfn{defining word} @code{def-word} and then
6163: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6164: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6165: is not executed at this time. The word @code{name} is sometimes called a
6166: @dfn{child} of @code{def-word}.
6167:
6168: When you execute @code{name}, the address of the body of @code{name} is
6169: put on the data stack and @i{code2} is executed (the address of the body
6170: of @code{name} is the address @code{HERE} returns immediately after the
6171: @code{CREATE}).
6172:
6173: @cindex atavism in child words
6174: You can use @code{def-word} to define a set of child words that behave
6175: differently, though atavistically; they all have a common run-time
6176: behaviour determined by @i{code2}. Typically, the @i{code1} sequence
6177: builds a data area in the body of the child word. The structure of the
6178: data is common to all children of @code{def-word}, but the data values
6179: are specific -- and private -- to each child word. When a child word is
6180: executed, the address of its private data area is passed as a parameter
6181: on TOS to be used and manipulated@footnote{It is legitimate both to read
6182: and write to this data area.} by @i{code2}.
6183:
6184: The two fragments of code that make up the defining words act (are
6185: executed) at two completely separate times:
6186:
6187: @itemize @bullet
6188: @item
6189: At @i{define time}, the defining word executes @i{code1} to generate a
6190: child word
6191: @item
6192: At @i{child execution time}, when a child word is invoked, @i{code2}
6193: is executed, using parameters (data) that are private and specific to
6194: the child word.
6195: @end itemize
6196:
6197: Another way of understanding the behaviour of @code{def-word} and
6198: @code{name} is to say that, if you make the following definitions:
6199: @example
6200: : def-word1 ( "name" -- )
6201: CREATE @i{code1} ;
6202:
6203: : action1 ( ... -- ... )
6204: @i{code2} ;
6205:
6206: def-word1 name1
6207: @end example
6208:
6209: @noindent
6210: Then using @code{name1 action1} is equivalent to using @code{name}.
6211:
6212: The classic example is that you can define @code{CONSTANT} in this way:
6213:
6214: @example
6215: : CONSTANT ( w "name" -- )
6216: CREATE ,
6217: DOES> ( -- w )
6218: @@ ;
6219: @end example
6220:
6221: @comment There is a beautiful description of how this works and what
6222: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6223: @comment commentary on the Counting Fruits problem.
6224:
6225: When you create a constant with @code{5 CONSTANT five}, a set of
6226: define-time actions take place; first a new word @code{five} is created,
6227: then the value 5 is laid down in the body of @code{five} with
6228: @code{,}. When @code{five} is executed, the address of the body is put on
6229: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6230: no code of its own; it simply contains a data field and a pointer to the
6231: code that follows @code{DOES>} in its defining word. That makes words
6232: created in this way very compact.
6233:
6234: The final example in this section is intended to remind you that space
6235: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6236: both read and written by a Standard program@footnote{Exercise: use this
6237: example as a starting point for your own implementation of @code{Value}
6238: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6239: @code{[']}.}:
6240:
6241: @example
6242: : foo ( "name" -- )
6243: CREATE -1 ,
6244: DOES> ( -- )
6245: @@ . ;
6246:
6247: foo first-word
6248: foo second-word
6249:
6250: 123 ' first-word >BODY !
6251: @end example
6252:
6253: If @code{first-word} had been a @code{CREATE}d word, we could simply
6254: have executed it to get the address of its data field. However, since it
6255: was defined to have @code{DOES>} actions, its execution semantics are to
6256: perform those @code{DOES>} actions. To get the address of its data field
6257: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6258: translate the xt into the address of the data field. When you execute
6259: @code{first-word}, it will display @code{123}. When you execute
6260: @code{second-word} it will display @code{-1}.
6261:
6262: @cindex stack effect of @code{DOES>}-parts
6263: @cindex @code{DOES>}-parts, stack effect
6264: In the examples above the stack comment after the @code{DOES>} specifies
6265: the stack effect of the defined words, not the stack effect of the
6266: following code (the following code expects the address of the body on
6267: the top of stack, which is not reflected in the stack comment). This is
6268: the convention that I use and recommend (it clashes a bit with using
6269: locals declarations for stack effect specification, though).
6270:
6271: @menu
6272: * CREATE..DOES> applications::
6273: * CREATE..DOES> details::
6274: * Advanced does> usage::
6275: @end menu
6276:
6277: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
6278: @subsubsection Applications of @code{CREATE..DOES>}
6279: @cindex @code{CREATE} ... @code{DOES>}, applications
6280:
6281: You may wonder how to use this feature. Here are some usage patterns:
6282:
6283: @cindex factoring similar colon definitions
6284: When you see a sequence of code occurring several times, and you can
6285: identify a meaning, you will factor it out as a colon definition. When
6286: you see similar colon definitions, you can factor them using
6287: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6288: that look very similar:
6289: @example
6290: : ori, ( reg-target reg-source n -- )
6291: 0 asm-reg-reg-imm ;
6292: : andi, ( reg-target reg-source n -- )
6293: 1 asm-reg-reg-imm ;
6294: @end example
6295:
6296: @noindent
6297: This could be factored with:
6298: @example
6299: : reg-reg-imm ( op-code -- )
6300: CREATE ,
6301: DOES> ( reg-target reg-source n -- )
6302: @@ asm-reg-reg-imm ;
6303:
6304: 0 reg-reg-imm ori,
6305: 1 reg-reg-imm andi,
6306: @end example
6307:
6308: @cindex currying
6309: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6310: supply a part of the parameters for a word (known as @dfn{currying} in
6311: the functional language community). E.g., @code{+} needs two
6312: parameters. Creating versions of @code{+} with one parameter fixed can
6313: be done like this:
6314: @example
6315: : curry+ ( n1 -- )
6316: CREATE ,
6317: DOES> ( n2 -- n1+n2 )
6318: @@ + ;
6319:
6320: 3 curry+ 3+
6321: -2 curry+ 2-
6322: @end example
6323:
6324: @node CREATE..DOES> details, Advanced does> usage, CREATE..DOES> applications, User-defined Defining Words
6325: @subsubsection The gory details of @code{CREATE..DOES>}
6326: @cindex @code{CREATE} ... @code{DOES>}, details
6327:
6328: doc-does>
6329:
6330: @cindex @code{DOES>} in a separate definition
6331: This means that you need not use @code{CREATE} and @code{DOES>} in the
6332: same definition; you can put the @code{DOES>}-part in a separate
6333: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
6334: @example
6335: : does1
6336: DOES> ( ... -- ... )
6337: ... ;
6338:
6339: : does2
6340: DOES> ( ... -- ... )
6341: ... ;
6342:
6343: : def-word ( ... -- ... )
6344: create ...
6345: IF
6346: does1
6347: ELSE
6348: does2
6349: ENDIF ;
6350: @end example
6351:
6352: In this example, the selection of whether to use @code{does1} or
6353: @code{does2} is made at compile-time; at the time that the child word is
6354: @code{CREATE}d.
6355:
6356: @cindex @code{DOES>} in interpretation state
6357: In a standard program you can apply a @code{DOES>}-part only if the last
6358: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6359: will override the behaviour of the last word defined in any case. In a
6360: standard program, you can use @code{DOES>} only in a colon
6361: definition. In Gforth, you can also use it in interpretation state, in a
6362: kind of one-shot mode; for example:
6363: @example
6364: CREATE name ( ... -- ... )
6365: @i{initialization}
6366: DOES>
6367: @i{code} ;
6368: @end example
6369:
6370: @noindent
6371: is equivalent to the standard:
6372: @example
6373: :noname
6374: DOES>
6375: @i{code} ;
6376: CREATE name EXECUTE ( ... -- ... )
6377: @i{initialization}
6378: @end example
6379:
6380: doc->body
6381:
6382: @node Advanced does> usage, , CREATE..DOES> details, User-defined Defining Words
6383: @subsubsection Advanced does> usage
6384:
6385:
6386:
6387: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6388: @subsection Deferred words
6389: @cindex deferred words
6390:
6391: The defining word @code{Defer} allows you to define a word by name
6392: without defining its behaviour; the definition of its behaviour is
6393: deferred. Here are two situation where this can be useful:
6394:
6395: @itemize @bullet
6396: @item
6397: Where you want to allow the behaviour of a word to be altered later, and
6398: for all precompiled references to the word to change when its behaviour
6399: is changed.
6400: @item
6401: For mutual recursion; @xref{Calls and returns}.
6402: @end itemize
6403:
6404: In the following example, @code{foo} always invokes the version of
6405: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6406: always invokes the version that prints ``@code{Hello}''. There is no way
6407: of getting @code{foo} to use the later version without re-ordering the
6408: source code and recompiling it.
6409:
6410: @example
6411: : greet ." Good morning" ;
6412: : foo ... greet ... ;
6413: : greet ." Hello" ;
6414: : bar ... greet ... ;
6415: @end example
6416:
6417: This problem can be solved by defining @code{greet} as a @code{Defer}red
6418: word. The behaviour of a @code{Defer}red word can be defined and
6419: redefined at any time by using @code{IS} to associate the xt of a
6420: previously-defined word with it. The previous example becomes:
6421:
6422: @example
6423: Defer greet
6424: : foo ... greet ... ;
6425: : bar ... greet ... ;
6426: : greet1 ." Good morning" ;
6427: : greet2 ." Hello" ;
6428: ' greet2 <IS> greet \ make greet behave like greet2
6429: @end example
6430:
6431: A deferred word can be used to improve the statistics-gathering example
6432: from @ref{User-defined Defining Words}; rather than edit the
6433: application's source code to change every @code{:} to a @code{my:}, do
6434: this:
6435:
6436: @example
6437: : real: : ; \ retain access to the original
6438: defer : \ redefine as a deferred word
6439: ' my: IS : \ use special version of :
6440: \
6441: \ load application here
6442: \
6443: ' real: IS : \ go back to the original
6444: @end example
6445:
6446:
6447: One thing to note is that @code{<IS>} consumes its name when it is
6448: executed. If you want to specify the name at compile time, use
6449: @code{[IS]}:
6450:
6451: @example
6452: : set-greet ( xt -- )
6453: [IS] greet ;
6454:
6455: ' greet1 set-greet
6456: @end example
6457:
6458: A deferred word can only inherit default semantics from the xt (because
6459: that is all that an xt can represent -- for more discussion of this
6460: @pxref{Tokens for Words}). However, the semantics of the deferred word
6461: itself can be modified at the time that it is defined. For example:
6462:
6463: @example
6464: : bar .... ; compile-only
6465: Defer fred immediate
6466: Defer jim
6467:
6468: ' bar <IS> jim \ jim has default semantics
6469: ' bar <IS> fred \ fred is immediate
6470: @end example
6471:
6472: doc-defer
6473: doc-<is>
6474: doc-[is]
6475: doc-is
6476: @comment TODO document these: what's defers [is]
6477: doc-what's
6478: doc-defers
6479:
6480: @c Use @code{words-deferred} to see a list of deferred words.
6481:
6482: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
6483: are provided in @file{compat/defer.fs}.
6484:
6485:
6486: @node Aliases, Supplying names, Deferred words, Defining Words
6487: @subsection Aliases
6488: @cindex aliases
6489:
6490: The defining word @code{Alias} allows you to define a word by name that
6491: has the same behaviour as some other word. Here are two situation where
6492: this can be useful:
6493:
6494: @itemize @bullet
6495: @item
6496: When you want access to a word's definition from a different word list
6497: (for an example of this, see the definition of the @code{Root} word list
6498: in the Gforth source).
6499: @item
6500: When you want to create a synonym; a definition that can be known by
6501: either of two names (for example, @code{THEN} and @code{ENDIF} are
6502: aliases).
6503: @end itemize
6504:
6505: The word whose behaviour the alias is to inherit is represented by an
6506: xt. Therefore, the alias only inherits default semantics from its
6507: ancestor. The semantics of the alias itself can be modified at the time
6508: that it is defined. For example:
6509:
6510: @example
6511: : foo ... ; immediate
6512:
6513: ' foo Alias bar \ bar is not an immediate word
6514: ' foo Alias fooby immediate \ fooby is an immediate word
6515: @end example
6516:
6517: Words that are aliases have the same xt, different headers in the
6518: dictionary, and consequently different name tokens (@pxref{Tokens for
6519: Words}) and possibly different immediate flags. An alias can only have
6520: default or immediate compilation semantics; you can define aliases for
6521: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
6522:
6523: doc-alias
6524:
6525:
6526: @node Supplying names, , Aliases, Defining Words
6527: @subsection Supplying the name of a defined word
6528: @cindex names for defined words
6529: @cindex defining words, name given in a string
6530:
6531: By default, a defining word takes the name for the defined word from the
6532: input stream. Sometimes you want to supply the name from a string. You
6533: can do this with:
6534:
6535: doc-nextname
6536:
6537: For example:
6538:
6539: @example
6540: s" foo" nextname create
6541: @end example
6542:
6543: @noindent
6544: is equivalent to:
6545:
6546: @example
6547: create foo
6548: @end example
6549:
6550: @noindent
6551: @code{nextname} works with any defining word, not just @code{:}.
6552:
6553:
6554: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6555: @section Interpretation and Compilation Semantics
6556: @cindex semantics, interpretation and compilation
6557:
6558: @cindex interpretation semantics
6559: The @dfn{interpretation semantics} of a word are what the text
6560: interpreter does when it encounters the word in interpret state. It also
6561: appears in some other contexts, e.g., the execution token returned by
6562: @code{' @i{word}} identifies the interpretation semantics of
6563: @i{word} (in other words, @code{' @i{word} execute} is equivalent to
6564: interpret-state text interpretation of @code{@i{word}}).
6565:
6566: @cindex compilation semantics
6567: The @dfn{compilation semantics} of a word are what the text interpreter
6568: does when it encounters the word in compile state. It also appears in
6569: other contexts, e.g, @code{POSTPONE @i{word}} compiles@footnote{In
6570: standard terminology, ``appends to the current definition''.} the
6571: compilation semantics of @i{word}.
6572:
6573: @cindex execution semantics
6574: The standard also talks about @dfn{execution semantics}. They are used
6575: only for defining the interpretation and compilation semantics of many
6576: words. By default, the interpretation semantics of a word are to
6577: @code{execute} its execution semantics, and the compilation semantics of
6578: a word are to @code{compile,} its execution semantics.@footnote{In
6579: standard terminology: The default interpretation semantics are its
6580: execution semantics; the default compilation semantics are to append its
6581: execution semantics to the execution semantics of the current
6582: definition.}
6583:
6584: @comment TODO expand, make it co-operate with new sections on text interpreter.
6585:
6586: @cindex immediate words
6587: @cindex compile-only words
6588: You can change the semantics of the most-recently defined word:
6589:
6590:
6591: doc-immediate
6592: doc-compile-only
6593: doc-restrict
6594:
6595:
6596: Note that ticking (@code{'}) a compile-only word gives an error
6597: (``Interpreting a compile-only word'').
6598:
6599: @menu
6600: * Combined words::
6601: @end menu
6602:
6603: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
6604: @subsection Combined Words
6605: @cindex combined words
6606:
6607: Gforth allows you to define @dfn{combined words} -- words that have an
6608: arbitrary combination of interpretation and compilation semantics.
6609:
6610:
6611: doc-interpret/compile:
6612:
6613:
6614: This feature was introduced for implementing @code{TO} and @code{S"}. I
6615: recommend that you do not define such words, as cute as they may be:
6616: they make it hard to get at both parts of the word in some contexts.
6617: E.g., assume you want to get an execution token for the compilation
6618: part. Instead, define two words, one that embodies the interpretation
6619: part, and one that embodies the compilation part. Once you have done
6620: that, you can define a combined word with @code{interpret/compile:} for
6621: the convenience of your users.
6622:
6623: You might try to use this feature to provide an optimizing
6624: implementation of the default compilation semantics of a word. For
6625: example, by defining:
6626: @example
6627: :noname
6628: foo bar ;
6629: :noname
6630: POSTPONE foo POSTPONE bar ;
6631: interpret/compile: opti-foobar
6632: @end example
6633:
6634: @noindent
6635: as an optimizing version of:
6636:
6637: @example
6638: : foobar
6639: foo bar ;
6640: @end example
6641:
6642: Unfortunately, this does not work correctly with @code{[compile]},
6643: because @code{[compile]} assumes that the compilation semantics of all
6644: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
6645: opti-foobar} would compile compilation semantics, whereas
6646: @code{[compile] foobar} would compile interpretation semantics.
6647:
6648: @cindex state-smart words (are a bad idea)
6649: Some people try to use @dfn{state-smart} words to emulate the feature provided
6650: by @code{interpret/compile:} (words are state-smart if they check
6651: @code{STATE} during execution). E.g., they would try to code
6652: @code{foobar} like this:
6653:
6654: @example
6655: : foobar
6656: STATE @@
6657: IF ( compilation state )
6658: POSTPONE foo POSTPONE bar
6659: ELSE
6660: foo bar
6661: ENDIF ; immediate
6662: @end example
6663:
6664: Although this works if @code{foobar} is only processed by the text
6665: interpreter, it does not work in other contexts (like @code{'} or
6666: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
6667: for a state-smart word, not for the interpretation semantics of the
6668: original @code{foobar}; when you execute this execution token (directly
6669: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
6670: state, the result will not be what you expected (i.e., it will not
6671: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
6672: write them@footnote{For a more detailed discussion of this topic, see
6673: @cite{@code{State}-smartness -- Why it is Evil and How to Exorcise it} by Anton
6674: Ertl; presented at EuroForth '98 and available from
6675: @*@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz}}!
6676:
6677: @cindex defining words with arbitrary semantics combinations
6678: It is also possible to write defining words that define words with
6679: arbitrary combinations of interpretation and compilation semantics. In
6680: general, they look like this:
6681:
6682: @example
6683: : def-word
6684: create-interpret/compile
6685: @i{code1}
6686: interpretation>
6687: @i{code2}
6688: <interpretation
6689: compilation>
6690: @i{code3}
6691: <compilation ;
6692: @end example
6693:
6694: For a @i{word} defined with @code{def-word}, the interpretation
6695: semantics are to push the address of the body of @i{word} and perform
6696: @i{code2}, and the compilation semantics are to push the address of
6697: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
6698: can also be defined like this (except that the defined constants don't
6699: behave correctly when @code{[compile]}d):
6700:
6701: @example
6702: : constant ( n "name" -- )
6703: create-interpret/compile
6704: ,
6705: interpretation> ( -- n )
6706: @@
6707: <interpretation
6708: compilation> ( compilation. -- ; run-time. -- n )
6709: @@ postpone literal
6710: <compilation ;
6711: @end example
6712:
6713:
6714: doc-create-interpret/compile
6715: doc-interpretation>
6716: doc-<interpretation
6717: doc-compilation>
6718: doc-<compilation
6719:
6720:
6721: Words defined with @code{interpret/compile:} and
6722: @code{create-interpret/compile} have an extended header structure that
6723: differs from other words; however, unless you try to access them with
6724: plain address arithmetic, you should not notice this. Words for
6725: accessing the header structure usually know how to deal with this; e.g.,
6726: @code{'} @i{word} @code{>body} also gives you the body of a word created
6727: with @code{create-interpret/compile}.
6728:
6729:
6730: doc-postpone
6731:
6732: @comment TODO -- expand glossary text for POSTPONE
6733:
6734:
6735: @c -------------------------------------------------------------
6736: @node Tokens for Words, The Text Interpreter, Interpretation and Compilation Semantics, Words
6737: @section Tokens for Words
6738: @cindex tokens for words
6739:
6740: This section describes the creation and use of tokens that represent
6741: words.
6742:
6743: Named words have information stored in their header space entries to
6744: indicate any non-default semantics (@pxref{Interpretation and
6745: Compilation Semantics}). The semantics can be modified, using
6746: @code{immediate} and/or @code{compile-only}, at the time that the words
6747: are defined. Unnamed words have (by definition) no header space
6748: entry, and therefore must have default semantics.
6749:
6750: Named words have interpretation and compilation semantics. Unnamed words
6751: just have execution semantics.
6752:
6753: @cindex xt
6754: @cindex execution token
6755: The execution semantics of an unnamed word are represented by an
6756: @dfn{execution token} (@i{xt}). As explained in @ref{Supplying names},
6757: the execution token of the last word defined can be produced with
6758: @code{lastxt}.
6759:
6760: The interpretation semantics of a named word are also represented by an
6761: execution token. You can produce the execution token using @code{'} or
6762: @code{[']}. A simple example shows the difference between the two:
6763:
6764: @example
6765: : greet ( -- ) ." Hello" ;
6766: : foo ( -- xt ) ['] greet execute ; \ ['] parses greet at compile-time
6767: : bar ( -- ) ' execute ; \ ' parses at run-time
6768:
6769: \ the next four lines all do the same thing
6770: foo
6771: bar greet
6772: greet
6773: ' greet EXECUTE
6774: @end example
6775:
6776: An execution token occupies one cell.
6777: @cindex code field address
6778: @cindex CFA
6779: In Gforth, the abstract data type @i{execution token} is implemented
6780: as a code field address (CFA).
6781: @comment TODO note that the standard does not say what it represents..
6782: @comment and you cannot necessarily compile it in all Forths (eg native
6783: @comment compilers?).
6784:
6785: For literals, use @code{'} in interpreted code and @code{[']} in
6786: compiled code. Gforth's @code{'} and @code{[']} behave somewhat
6787: unusually by complaining about compile-only words. To get the execution
6788: token for a compile-only word @i{name}, use @code{COMP' @i{name} DROP}
6789: or @code{[COMP'] @i{name} DROP}.
6790:
6791: @cindex compilation token
6792: The compilation semantics of a named word are represented by a
6793: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
6794: @i{xt} is an execution token. The compilation semantics represented by
6795: the compilation token can be performed with @code{execute}, which
6796: consumes the whole compilation token, with an additional stack effect
6797: determined by the represented compilation semantics.
6798:
6799: At present, the @i{w} part of a compilation token is an execution token,
6800: and the @i{xt} part represents either @code{execute} or
6801: @code{compile,}@footnote{Depending upon the compilation semantics of the
6802: word. If the word has default compilation semantics, the @i{xt} will
6803: represent @code{compile,}. Otherwise (e.g., for immediate words), the
6804: @i{xt} will represent @code{execute}.}. However, don't rely on that
6805: knowledge, unless necessary; future versions of Gforth may introduce
6806: unusual compilation tokens (e.g., a compilation token that represents
6807: the compilation semantics of a literal).
6808:
6809: You can compile the compilation semantics with @code{postpone,}. I.e.,
6810: @code{COMP' @i{word} postpone,} is equivalent to @code{postpone
6811: @i{word}}.
6812:
6813: @cindex name token
6814: @cindex name field address
6815: @cindex NFA
6816: Named words are also represented by the @dfn{name token}, (@i{nt}). In
6817: Gforth, the abstract data type @emph{name token} is implemented as a
6818: name field address (NFA).
6819:
6820:
6821: doc-execute
6822: doc-perform
6823: doc-compile,
6824: doc-[']
6825: doc-'
6826: doc-[comp']
6827: doc-comp'
6828: doc-postpone,
6829:
6830: doc-find-name
6831: doc-name>int
6832: doc-name?int
6833: doc-name>comp
6834: doc-name>string
6835:
6836:
6837: @c ----------------------------------------------------------
6838: @node The Text Interpreter, Word Lists, Tokens for Words, Words
6839: @section The Text Interpreter
6840: @cindex interpreter - outer
6841: @cindex text interpreter
6842: @cindex outer interpreter
6843:
6844: @c Should we really describe all these ugly details? IMO the text
6845: @c interpreter should be much cleaner, but that may not be possible within
6846: @c ANS Forth. - anton
6847: @c nac-> I wanted to explain how it works to show how you can exploit
6848: @c it in your own programs. When I was writing a cross-compiler, figuring out
6849: @c some of these gory details was very helpful to me. None of the textbooks
6850: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
6851: @c seems to positively avoid going into too much detail for some of
6852: @c the internals.
6853:
6854: The text interpreter@footnote{This is an expanded version of the
6855: material in @ref{Introducing the Text Interpreter}.} is an endless loop
6856: that processes input from the current input device. It is also called
6857: the outer interpreter, in contrast to the inner interpreter
6858: (@pxref{Engine}) which executes the compiled Forth code on interpretive
6859: implementations.
6860:
6861: @cindex interpret state
6862: @cindex compile state
6863: The text interpreter operates in one of two states: @dfn{interpret
6864: state} and @dfn{compile state}. The current state is defined by the
6865: aptly-named variable, @code{state}.
6866:
6867: This section starts by describing how the text interpreter behaves when
6868: it is in interpret state, processing input from the user input device --
6869: the keyboard. This is the mode that a Forth system is in after it starts
6870: up.
6871:
6872: @cindex input buffer
6873: @cindex terminal input buffer
6874: The text interpreter works from an area of memory called the @dfn{input
6875: buffer}@footnote{When the text interpreter is processing input from the
6876: keyboard, this area of memory is called the @dfn{terminal input buffer}
6877: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
6878: @code{#TIB}.}, which stores your keyboard input when you press the
6879: @key{RET} key. Starting at the beginning of the input buffer, it skips
6880: leading spaces (called @dfn{delimiters}) then parses a string (a
6881: sequence of non-space characters) until it reaches either a space
6882: character or the end of the buffer. Having parsed a string, it makes two
6883: attempts to process it:
6884:
6885: @cindex dictionary
6886: @itemize @bullet
6887: @item
6888: It looks for the string in a @dfn{dictionary} of definitions. If the
6889: string is found, the string names a @dfn{definition} (also known as a
6890: @dfn{word}) and the dictionary search returns information that allows
6891: the text interpreter to perform the word's @dfn{interpretation
6892: semantics}. In most cases, this simply means that the word will be
6893: executed.
6894: @item
6895: If the string is not found in the dictionary, the text interpreter
6896: attempts to treat it as a number, using the rules described in
6897: @ref{Number Conversion}. If the string represents a legal number in the
6898: current radix, the number is pushed onto a parameter stack (the data
6899: stack for integers, the floating-point stack for floating-point
6900: numbers).
6901: @end itemize
6902:
6903: If both attempts fail, or if the word is found in the dictionary but has
6904: no interpretation semantics@footnote{This happens if the word was
6905: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
6906: remainder of the input buffer, issues an error message and waits for
6907: more input. If one of the attempts succeeds, the text interpreter
6908: repeats the parsing process until the whole of the input buffer has been
6909: processed, at which point it prints the status message ``@code{ ok}''
6910: and waits for more input.
6911:
6912: @cindex parse area
6913: The text interpreter keeps track of its position in the input buffer by
6914: updating a variable called @code{>IN} (pronounced ``to-in''). The value
6915: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
6916: of the input buffer. The region from offset @code{>IN @@} to the end of
6917: the input buffer is called the @dfn{parse area}@footnote{In other words,
6918: the text interpreter processes the contents of the input buffer by
6919: parsing strings from the parse area until the parse area is empty.}.
6920: This example shows how @code{>IN} changes as the text interpreter parses
6921: the input buffer:
6922:
6923: @example
6924: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
6925: CR ." ->" TYPE ." <-" ; IMMEDIATE
6926:
6927: 1 2 3 remaining + remaining .
6928:
6929: : foo 1 2 3 remaining SWAP remaining ;
6930: @end example
6931:
6932: @noindent
6933: The result is:
6934:
6935: @example
6936: ->+ remaining .<-
6937: ->.<-5 ok
6938:
6939: ->SWAP remaining ;-<
6940: ->;<- ok
6941: @end example
6942:
6943: @cindex parsing words
6944: The value of @code{>IN} can also be modified by a word in the input
6945: buffer that is executed by the text interpreter. This means that a word
6946: can ``trick'' the text interpreter into either skipping a section of the
6947: input buffer@footnote{This is how parsing words work.} or into parsing a
6948: section twice. For example:
6949:
6950: @example
6951: : lat ." <<lat>>" ;
6952: : flat ." <<flat>>" >IN DUP @@ 3 - SWAP ! ;
6953: @end example
6954:
6955: @noindent
6956: When @code{flat} is executed, this output is produced@footnote{Exercise
6957: for the reader: what would happen if the @code{3} were replaced with
6958: @code{4}?}:
6959:
6960: @example
6961: <<flat>><<lat>>
6962: @end example
6963:
6964: @noindent
6965: Two important notes about the behaviour of the text interpreter:
6966:
6967: @itemize @bullet
6968: @item
6969: It processes each input string to completion before parsing additional
6970: characters from the input buffer.
6971: @item
6972: It treats the input buffer as a read-only region (and so must your code).
6973: @end itemize
6974:
6975: @noindent
6976: When the text interpreter is in compile state, its behaviour changes in
6977: these ways:
6978:
6979: @itemize @bullet
6980: @item
6981: If a parsed string is found in the dictionary, the text interpreter will
6982: perform the word's @dfn{compilation semantics}. In most cases, this
6983: simply means that the execution semantics of the word will be appended
6984: to the current definition.
6985: @item
6986: When a number is encountered, it is compiled into the current definition
6987: (as a literal) rather than being pushed onto a parameter stack.
6988: @item
6989: If an error occurs, @code{state} is modified to put the text interpreter
6990: back into interpret state.
6991: @item
6992: Each time a line is entered from the keyboard, Gforth prints
6993: ``@code{ compiled}'' rather than `` @code{ok}''.
6994: @end itemize
6995:
6996: @cindex text interpreter - input sources
6997: When the text interpreter is using an input device other than the
6998: keyboard, its behaviour changes in these ways:
6999:
7000: @itemize @bullet
7001: @item
7002: When the parse area is empty, the text interpreter attempts to refill
7003: the input buffer from the input source. When the input source is
7004: exhausted, the input source is set back to the user input device.
7005: @item
7006: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7007: time the parse area is emptied.
7008: @item
7009: If an error occurs, the input source is set back to the user input
7010: device.
7011: @end itemize
7012:
7013: You can read about this in more detail in @ref{Input Sources}.
7014:
7015: doc->in
7016: doc-source
7017:
7018: doc-tib
7019: doc-#tib
7020:
7021:
7022: @menu
7023: * Input Sources::
7024: * Number Conversion::
7025: * Interpret/Compile states::
7026: * Literals::
7027: * Interpreter Directives::
7028: @end menu
7029:
7030: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7031: @subsection Input Sources
7032: @cindex input sources
7033: @cindex text interpreter - input sources
7034:
7035: By default, the text interpreter processes input from the user input
7036: device (the keyboard) when Forth starts up. The text interpreter can
7037: process input from any of these sources:
7038:
7039: @itemize @bullet
7040: @item
7041: The user input device -- the keyboard.
7042: @item
7043: A file, using the words described in @ref{Forth source files}.
7044: @item
7045: A block, using the words described in @ref{Blocks}.
7046: @item
7047: A text string, using @code{evaluate}.
7048: @end itemize
7049:
7050: A program can identify the current input device from the values of
7051: @code{source-id} and @code{blk}.
7052:
7053:
7054: doc-source-id
7055: doc-blk
7056:
7057: doc-save-input
7058: doc-restore-input
7059:
7060: doc-evaluate
7061:
7062:
7063:
7064: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
7065: @subsection Number Conversion
7066: @cindex number conversion
7067: @cindex double-cell numbers, input format
7068: @cindex input format for double-cell numbers
7069: @cindex single-cell numbers, input format
7070: @cindex input format for single-cell numbers
7071: @cindex floating-point numbers, input format
7072: @cindex input format for floating-point numbers
7073:
7074: This section describes the rules that the text interpreter uses when it
7075: tries to convert a string into a number.
7076:
7077: Let <digit> represent any character that is a legal digit in the current
7078: number base@footnote{For example, 0-9 when the number base is decimal or
7079: 0-9, A-F when the number base is hexadecimal.}.
7080:
7081: Let <decimal digit> represent any character in the range 0-9.
7082:
7083: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7084: in the braces (@i{a} or @i{b} or neither).
7085:
7086: Let * represent any number of instances of the previous character
7087: (including none).
7088:
7089: Let any other character represent itself.
7090:
7091: @noindent
7092: Now, the conversion rules are:
7093:
7094: @itemize @bullet
7095: @item
7096: A string of the form <digit><digit>* is treated as a single-precision
7097: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
7098: @item
7099: A string of the form -<digit><digit>* is treated as a single-precision
7100: (cell-sized) negative integer, and is represented using 2's-complement
7101: arithmetic. Examples are -45 -5681 -0
7102: @item
7103: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
7104: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7105: (all three of these represent the same number).
7106: @item
7107: A string of the form -<digit><digit>*.<digit>* is treated as a
7108: double-precision (double-cell-sized) negative integer, and is
7109: represented using 2's-complement arithmetic. Examples are -3465. -3.465
7110: -34.65 (all three of these represent the same number).
7111: @item
7112: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7113: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
7114: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
7115: number) +12.E-4
7116: @end itemize
7117:
7118: By default, the number base used for integer number conversion is given
7119: by the contents of the variable @code{base}. Note that a lot of
7120: confusion can result from unexpected values of @code{base}. If you
7121: change @code{base} anywhere, make sure to save the old value and restore
7122: it afterwards. In general I recommend keeping @code{base} decimal, and
7123: using the prefixes described below for the popular non-decimal bases.
7124:
7125: doc-dpl
7126: doc-base
7127: doc-hex
7128: doc-decimal
7129:
7130:
7131: @cindex '-prefix for character strings
7132: @cindex &-prefix for decimal numbers
7133: @cindex %-prefix for binary numbers
7134: @cindex $-prefix for hexadecimal numbers
7135: Gforth allows you to override the value of @code{base} by using a
7136: prefix@footnote{Some Forth implementations provide a similar scheme by
7137: implementing @code{$} etc. as parsing words that process the subsequent
7138: number in the input stream and push it onto the stack. For example, see
7139: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7140: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7141: is required between the prefix and the number.} before the first digit
7142: of an (integer) number. Four prefixes are supported:
7143:
7144: @itemize @bullet
7145: @item
7146: @code{&} -- decimal
7147: @item
7148: @code{%} -- binary
7149: @item
7150: @code{$} -- hexadecimal
7151: @item
7152: @code{'} -- base @code{max-char+1}
7153: @end itemize
7154:
7155: Here are some examples, with the equivalent decimal number shown after
7156: in braces:
7157:
7158: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7159: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
7160: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
7161: &905 (905), $abc (2478), $ABC (2478).
7162:
7163: @cindex number conversion - traps for the unwary
7164: @noindent
7165: Number conversion has a number of traps for the unwary:
7166:
7167: @itemize @bullet
7168: @item
7169: You cannot determine the current number base using the code sequence
7170: @code{base @@ .} -- the number base is always 10 in the current number
7171: base. Instead, use something like @code{base @@ dec.}
7172: @item
7173: If the number base is set to a value greater than 14 (for example,
7174: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7175: it to be intepreted as either a single-precision integer or a
7176: floating-point number (Gforth treats it as an integer). The ambiguity
7177: can be resolved by explicitly stating the sign of the mantissa and/or
7178: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7179: ambiguity arises; either representation will be treated as a
7180: floating-point number.
7181: @item
7182: There is a word @code{bin} but it does @i{not} set the number base!
7183: It is used to specify file types.
7184: @item
7185: ANS Forth requires the @code{.} of a double-precision number to
7186: be the final character in the string. Allowing the @code{.} to be
7187: anywhere after the first digit is a Gforth extension.
7188: @item
7189: The number conversion process does not check for overflow.
7190: @item
7191: In Gforth, number conversion to floating-point numbers always use base
7192: 10, irrespective of the value of @code{base}. In ANS Forth,
7193: conversion to floating-point numbers whilst the value of
7194: @code{base} is not 10 is an ambiguous condition.
7195: @end itemize
7196:
7197: You can read numbers into your programs with the words described in
7198: @ref{Input}.
7199:
7200: @node Interpret/Compile states, Literals, Number Conversion, The Text Interpreter
7201: @subsection Interpret/Compile states
7202: @cindex Interpret/Compile states
7203:
7204: A standard program is not permitted to change @code{state}
7205: explicitly. However, it can change @code{state} implicitly, using the
7206: words @code{[} and @code{]}. When @code{[} is executed it switches
7207: @code{state} to interpret state, and therefore the text interpreter
7208: starts interpreting. When @code{]} is executed it switches @code{state}
7209: to compile state and therefore the text interpreter starts
7210: compiling. The most common usage for these words is for switching into
7211: interpret state and back from within a colon definition; this technique
7212: can be used to compile a literal (for an example, @pxref{Literals}) or
7213: for conditional compilation (for an example, @pxref{Interpreter
7214: Directives}).
7215:
7216:
7217: @c This is a bad example: It's non-standard, and it's not necessary.
7218: @c However, I can't think of a good example for switching into compile
7219: @c state when there is no current word (@code{state}-smart words are not a
7220: @c good reason). So maybe we should use an example for switching into
7221: @c interpret @code{state} in a colon def. - anton
7222: @c nac-> I agree. I started out by putting in the example, then realised
7223: @c that it was non-ANS, so wrote more words around it. I hope this
7224: @c re-written version is acceptable to you. I do want to keep the example
7225: @c as it is helpful for showing what is and what is not portable, particularly
7226: @c where it outlaws a style in common use.
7227:
7228:
7229: @code{[} and @code{]} also give you the ability to switch into compile
7230: state and back, but we cannot think of any useful Standard application
7231: for this ability. Pre-ANS Forth textbooks have examples like this:
7232:
7233: @example
7234: : AA ." this is A" ;
7235: : BB ." this is B" ;
7236: : CC ." this is C" ;
7237:
7238: create table ] aa bb cc [
7239:
7240: : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7241: cells table + @ execute ;
7242: @end example
7243:
7244: This example builds a jump table; @code{0 go} will display ``@code{this
7245: is A}''. Using @code{[} and @code{]} in this example is equivalent to
7246: defining @code{table} like this:
7247:
7248: @example
7249: create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
7250: @end example
7251:
7252: The problem with this code is that the definition of @code{table} is not
7253: portable -- it @i{compile}s execution tokens into code space. Whilst it
7254: @i{may} work on systems where code space and data space co-incide, the
7255: Standard only allows data space to be assigned for a @code{CREATE}d
7256: word. In addition, the Standard only allows @code{@@} to access data
7257: space, whilst this example is using it to access code space. The only
7258: portable, Standard way to build this table is to build it in data space,
7259: like this:
7260:
7261: @example
7262: create table ' aa , ' bb , ' cc ,
7263: @end example
7264:
7265: doc-state
7266: doc-[
7267: doc-]
7268:
7269:
7270: @node Literals, Interpreter Directives, Interpret/Compile states, The Text Interpreter
7271: @subsection Literals
7272: @cindex Literals
7273:
7274: Often, you want to use a number within a colon definition. When you do
7275: this, the text interpreter automatically compiles the number as a
7276: @i{literal}. A literal is a number whose run-time effect is to be pushed
7277: onto the stack. If you had to do some maths to generate the number, you
7278: might write it like this:
7279:
7280: @example
7281: : HOUR-TO-SEC ( n1 -- n2 )
7282: 60 * \ to minutes
7283: 60 * ; \ to seconds
7284: @end example
7285:
7286: It is very clear what this definition is doing, but it's inefficient
7287: since it is performing 2 multiples at run-time. An alternative would be
7288: to write:
7289:
7290: @example
7291: : HOUR-TO-SEC ( n1 -- n2 )
7292: 3600 * ; \ to seconds
7293: @end example
7294:
7295: Which does the same thing, and has the advantage of using a single
7296: multiply. Ideally, we'd like the efficiency of the second with the
7297: readability of the first.
7298:
7299: @code{Literal} allows us to achieve that. It takes a number from the
7300: stack and lays it down in the current definition just as though the
7301: number had been typed directly into the definition. Our first attempt
7302: might look like this:
7303:
7304: @example
7305: 60 \ mins per hour
7306: 60 * \ seconds per minute
7307: : HOUR-TO-SEC ( n1 -- n2 )
7308: Literal * ; \ to seconds
7309: @end example
7310:
7311: But this produces the error message @code{unstructured}. What happened?
7312: The stack notation for @code{:} is (@i{ -- colon-sys}) and the size of
7313: @i{colon-sys} is implementation-defined. In other words, once we start a
7314: colon definition we can't portably access anything that was on the stack
7315: before the definition began@footnote{@cite{Two Problems in ANS Forth},
7316: by Thomas Worthington; Forth Dimensions 20(2) pages 32--34 describes
7317: some situations where you might want to access stack items above
7318: colon-sys, and provides a solution to the problem.}. The correct way of
7319: solving this problem in this instance is to use @code{[ ]} like this:
7320:
7321: @example
7322: : HOUR-TO-SEC ( n1 -- n2 )
7323: [ 60 \ minutes per hour
7324: 60 * ] \ seconds per minute
7325: LITERAL * ; \ to seconds
7326: @end example
7327:
7328:
7329: doc-literal
7330: doc-]L
7331: doc-2literal
7332: doc-fliteral
7333:
7334:
7335: @node Interpreter Directives, , Literals, The Text Interpreter
7336: @subsection Interpreter Directives
7337: @cindex interpreter directives
7338:
7339: These words are usually used in interpret state; typically to control
7340: which parts of a source file are processed by the text
7341: interpreter. There are only a few ANS Forth Standard words, but Gforth
7342: supplements these with a rich set of immediate control structure words
7343: to compensate for the fact that the non-immediate versions can only be
7344: used in compile state (@pxref{Control Structures}). Typical usages:
7345:
7346: @example
7347: FALSE Constant ASSEMBLER
7348: .
7349: .
7350: ASSEMBLER [IF]
7351: : ASSEMBLER-FEATURE
7352: ...
7353: ;
7354: [ENDIF]
7355: .
7356: .
7357: : SEE
7358: ... \ general-purpose SEE code
7359: [ ASSEMBLER [IF] ]
7360: ... \ assembler-specific SEE code
7361: [ [ENDIF] ]
7362: ;
7363: @end example
7364:
7365:
7366: doc-[IF]
7367: doc-[ELSE]
7368: doc-[THEN]
7369: doc-[ENDIF]
7370:
7371: doc-[IFDEF]
7372: doc-[IFUNDEF]
7373:
7374: doc-[?DO]
7375: doc-[DO]
7376: doc-[FOR]
7377: doc-[LOOP]
7378: doc-[+LOOP]
7379: doc-[NEXT]
7380:
7381: doc-[BEGIN]
7382: doc-[UNTIL]
7383: doc-[AGAIN]
7384: doc-[WHILE]
7385: doc-[REPEAT]
7386:
7387:
7388: @c -------------------------------------------------------------
7389: @node Word Lists, Environmental Queries, The Text Interpreter, Words
7390: @section Word Lists
7391: @cindex word lists
7392: @cindex header space
7393:
7394: A wordlist is a list of named words; you can add new words and look up
7395: words by name (and you can remove words in a restricted way with
7396: markers). Every named (and @code{reveal}ed) word is in one wordlist.
7397:
7398: @cindex search order stack
7399: The text interpreter searches the wordlists present in the search order
7400: (a stack of wordlists), from the top to the bottom. Within each
7401: wordlist, the search starts conceptually at the newest word; i.e., if
7402: two words in a wordlist have the same name, the newer word is found.
7403:
7404: @cindex compilation word list
7405: New words are added to the @dfn{compilation wordlist} (aka current
7406: wordlist).
7407:
7408: @cindex wid
7409: A word list is identified by a cell-sized word list identifier (@i{wid})
7410: in much the same way as a file is identified by a file handle. The
7411: numerical value of the wid has no (portable) meaning, and might change
7412: from session to session.
7413:
7414: The ANS Forth ``Search order'' word set is intended to provide a set of
7415: low-level tools that allow various different schemes to be
7416: implemented. Gforth provides @code{vocabulary}, a traditional Forth
7417: word. @file{compat/vocabulary.fs} provides an implementation in ANS
7418: Forth.
7419:
7420: @comment TODO: locals section refers to here, saying that every word list (aka
7421: @comment vocabulary) has its own methods for searching etc. Need to document that.
7422:
7423: @comment TODO: document markers, reveal, tables, mappedwordlist
7424:
7425: @comment the gforthman- prefix is used to pick out the true definition of a
7426: @comment word from the source files, rather than some alias.
7427:
7428: doc-forth-wordlist
7429: doc-definitions
7430: doc-get-current
7431: doc-set-current
7432: doc-get-order
7433: doc---gforthman-set-order
7434: doc-wordlist
7435: doc-table
7436: doc-push-order
7437: doc-previous
7438: doc-also
7439: doc---gforthman-forth
7440: doc-only
7441: doc---gforthman-order
7442:
7443: doc-find
7444: doc-search-wordlist
7445:
7446: doc-words
7447: doc-vlist
7448: @c doc-words-deferred
7449:
7450: doc-mappedwordlist
7451: doc-root
7452: doc-vocabulary
7453: doc-seal
7454: doc-vocs
7455: doc-current
7456: doc-context
7457:
7458:
7459: @menu
7460: * Why use word lists?::
7461: * Word list examples::
7462: @end menu
7463:
7464: @node Why use word lists?, Word list examples, Word Lists, Word Lists
7465: @subsection Why use word lists?
7466: @cindex word lists - why use them?
7467:
7468: Here are some reasons for using multiple word lists:
7469:
7470: @itemize @bullet
7471: @item
7472: To improve compilation speed by reducing the number of header space
7473: entries that must be searched. This is achieved by creating a new
7474: word list that contains all of the definitions that are used in the
7475: definition of a Forth system but which would not usually be used by
7476: programs running on that system. That word list would be on the search
7477: list when the Forth system was compiled but would be removed from the
7478: search list for normal operation. This can be a useful technique for
7479: low-performance systems (for example, 8-bit processors in embedded
7480: systems) but is unlikely to be necessary in high-performance desktop
7481: systems.
7482: @item
7483: To prevent a set of words from being used outside the context in which
7484: they are valid. Two classic examples of this are an integrated editor
7485: (all of the edit commands are defined in a separate word list; the
7486: search order is set to the editor word list when the editor is invoked;
7487: the old search order is restored when the editor is terminated) and an
7488: integrated assembler (the op-codes for the machine are defined in a
7489: separate word list which is used when a @code{CODE} word is defined).
7490: @item
7491: To prevent a name-space clash between multiple definitions with the same
7492: name. For example, when building a cross-compiler you might have a word
7493: @code{IF} that generates conditional code for your target system. By
7494: placing this definition in a different word list you can control whether
7495: the host system's @code{IF} or the target system's @code{IF} get used in
7496: any particular context by controlling the order of the word lists on the
7497: search order stack.
7498: @end itemize
7499:
7500: @node Word list examples, , Why use word lists?, Word Lists
7501: @subsection Word list examples
7502: @cindex word lists - examples
7503:
7504: Here is an example of creating and using a new wordlist using ANS
7505: Forth Standard words:
7506:
7507: @example
7508: wordlist constant my-new-words-wordlist
7509: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
7510:
7511: \ add it to the search order
7512: also my-new-words
7513:
7514: \ alternatively, add it to the search order and make it
7515: \ the compilation word list
7516: also my-new-words definitions
7517: \ type "order" to see the problem
7518: @end example
7519:
7520: The problem with this example is that @code{order} has no way to
7521: associate the name @code{my-new-words} with the wid of the word list (in
7522: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
7523: that has no associated name). There is no Standard way of associating a
7524: name with a wid.
7525:
7526: In Gforth, this example can be re-coded using @code{vocabulary}, which
7527: associates a name with a wid:
7528:
7529: @example
7530: vocabulary my-new-words
7531:
7532: \ add it to the search order
7533: also my-new-words
7534:
7535: \ alternatively, add it to the search order and make it
7536: \ the compilation word list
7537: my-new-words definitions
7538: \ type "order" to see that the problem is solved
7539: @end example
7540:
7541: @c -------------------------------------------------------------
7542: @node Environmental Queries, Files, Word Lists, Words
7543: @section Environmental Queries
7544: @cindex environmental queries
7545:
7546: ANS Forth introduced the idea of ``environmental queries'' as a way
7547: for a program running on a system to determine certain characteristics of the system.
7548: The Standard specifies a number of strings that might be recognised by a system.
7549:
7550: The Standard requires that the header space used for environmental queries
7551: be distinct from the header space used for definitions.
7552:
7553: Typically, environmental queries are supported by creating a set of
7554: definitions in a word list that is @i{only} used during environmental
7555: queries; that is what Gforth does. There is no Standard way of adding
7556: definitions to the set of recognised environmental queries, but any
7557: implementation that supports the loading of optional word sets must have
7558: some mechanism for doing this (after loading the word set, the
7559: associated environmental query string must return @code{true}). In
7560: Gforth, the word list used to honour environmental queries can be
7561: manipulated just like any other word list.
7562:
7563:
7564: doc-environment?
7565: doc-environment-wordlist
7566:
7567: doc-gforth
7568: doc-os-class
7569:
7570:
7571: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
7572: returning two items on the stack, querying it using @code{environment?}
7573: will return an additional item; the @code{true} flag that shows that the
7574: string was recognised.
7575:
7576: @comment TODO Document the standard strings or note where they are documented herein
7577:
7578: Here are some examples of using environmental queries:
7579:
7580: @example
7581: s" address-unit-bits" environment? 0=
7582: [IF]
7583: cr .( environmental attribute address-units-bits unknown... ) cr
7584: [THEN]
7585:
7586: s" block" environment? [IF] DROP include block.fs [THEN]
7587:
7588: s" gforth" environment? [IF] 2DROP include compat/vocabulary.fs [THEN]
7589:
7590: s" gforth" environment? [IF] .( Gforth version ) TYPE
7591: [ELSE] .( Not Gforth..) [THEN]
7592: @end example
7593:
7594:
7595: Here is an example of adding a definition to the environment word list:
7596:
7597: @example
7598: get-current environment-wordlist set-current
7599: true constant block
7600: true constant block-ext
7601: set-current
7602: @end example
7603:
7604: You can see what definitions are in the environment word list like this:
7605:
7606: @example
7607: get-order 1+ environment-wordlist swap set-order words previous
7608: @end example
7609:
7610:
7611: @c -------------------------------------------------------------
7612: @node Files, Blocks, Environmental Queries, Words
7613: @section Files
7614: @cindex files
7615: @cindex I/O - file-handling
7616:
7617: Gforth provides facilities for accessing files that are stored in the
7618: host operating system's file-system. Files that are processed by Gforth
7619: can be divided into two categories:
7620:
7621: @itemize @bullet
7622: @item
7623: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
7624: @item
7625: Files that are processed by some other program (@dfn{general files}).
7626: @end itemize
7627:
7628: doc-loadfilename
7629: doc-sourcefilename
7630: doc-sourceline#
7631:
7632: @menu
7633: * Forth source files::
7634: * General files::
7635: * Search Paths::
7636: @end menu
7637:
7638:
7639: @c -------------------------------------------------------------
7640: @node Forth source files, General files, Files, Files
7641: @subsection Forth source files
7642: @cindex including files
7643: @cindex Forth source files
7644:
7645: The simplest way to interpret the contents of a file is to use one of
7646: these two formats:
7647:
7648: @example
7649: include mysource.fs
7650: s" mysource.fs" included
7651: @end example
7652:
7653: Sometimes you want to include a file only if it is not included already
7654: (by, say, another source file). In that case, you can use one of these
7655: three formats:
7656:
7657: @example
7658: require mysource.fs
7659: needs mysource.fs
7660: s" mysource.fs" required
7661: @end example
7662:
7663: @cindex stack effect of included files
7664: @cindex including files, stack effect
7665: It is good practice to write your source files such that interpreting them
7666: does not change the stack. Source files designed in this way can be used with
7667: @code{required} and friends without complications. For example:
7668:
7669: @example
7670: 1 require foo.fs drop
7671: @end example
7672:
7673:
7674: doc-include-file
7675: doc-included
7676: doc-included?
7677: doc-include
7678: doc-required
7679: doc-require
7680: doc-needs
7681: doc-init-included-files
7682:
7683:
7684: A definition in ANS Forth for @code{required} is provided in
7685: @file{compat/required.fs}.
7686:
7687: @c -------------------------------------------------------------
7688: @node General files, Search Paths, Forth source files, Files
7689: @subsection General files
7690: @cindex general files
7691: @cindex file-handling
7692:
7693: Files are opened/created by name and type. The following types are
7694: recognised:
7695:
7696:
7697: doc-r/o
7698: doc-r/w
7699: doc-w/o
7700: doc-bin
7701:
7702:
7703: When a file is opened/created, it returns a file identifier,
7704: @i{wfileid} that is used for all other file commands. All file
7705: commands also return a status value, @i{wior}, that is 0 for a
7706: successful operation and an implementation-defined non-zero value in the
7707: case of an error.
7708:
7709:
7710: doc-open-file
7711: doc-create-file
7712:
7713: doc-close-file
7714: doc-delete-file
7715: doc-rename-file
7716: doc-read-file
7717: doc-read-line
7718: doc-write-file
7719: doc-write-line
7720: doc-emit-file
7721: doc-flush-file
7722:
7723: doc-file-status
7724: doc-file-position
7725: doc-reposition-file
7726: doc-file-size
7727: doc-resize-file
7728:
7729:
7730: @c ---------------------------------------------------------
7731: @node Search Paths, , General files, Files
7732: @subsection Search Paths
7733: @cindex path for @code{included}
7734: @cindex file search path
7735: @cindex @code{include} search path
7736: @cindex search path for files
7737:
7738: If you specify an absolute filename (i.e., a filename starting with
7739: @file{/} or @file{~}, or with @file{:} in the second position (as in
7740: @samp{C:...})) for @code{included} and friends, that file is included
7741: just as you would expect.
7742:
7743: For relative filenames, Gforth uses a search path similar to Forth's
7744: search order (@pxref{Word Lists}). It tries to find the given filename
7745: in the directories present in the path, and includes the first one it
7746: finds. There are separate search paths for Forth source files and
7747: general files.
7748:
7749: If the search path contains the directory @file{.} (as it should), this
7750: refers to the directory that the present file was @code{included}
7751: from. This allows files to include other files relative to their own
7752: position (irrespective of the current working directory or the absolute
7753: position). This feature is essential for libraries consisting of
7754: several files, where a file may include other files from the library.
7755: It corresponds to @code{#include "..."} in C. If the current input
7756: source is not a file, @file{.} refers to the directory of the innermost
7757: file being included, or, if there is no file being included, to the
7758: current working directory.
7759:
7760: Use @file{~+} to refer to the current working directory (as in the
7761: @code{bash}).
7762:
7763: If the filename starts with @file{./}, the search path is not searched
7764: (just as with absolute filenames), and the @file{.} has the same meaning
7765: as described above.
7766:
7767: @menu
7768: * Forth Search Paths::
7769: * General Search Paths::
7770: @end menu
7771:
7772: @c ---------------------------------------------------------
7773: @node Forth Search Paths, General Search Paths, Search Paths, Search Paths
7774: @subsubsection Forth Search Paths
7775: @cindex search path control - Forth
7776:
7777: The search path is initialized when you start Gforth (@pxref{Invoking
7778: Gforth}). You can display it and change it using these words:
7779:
7780:
7781: doc-.fpath
7782: doc-fpath+
7783: doc-fpath=
7784: doc-open-fpath-file
7785:
7786:
7787: @noindent
7788: Here is an example of using @code{fpath} and @code{require}:
7789:
7790: @example
7791: fpath= /usr/lib/forth/|./
7792: require timer.fs
7793: @end example
7794:
7795: @c ---------------------------------------------------------
7796: @node General Search Paths, , Forth Search Paths, Search Paths
7797: @subsubsection General Search Paths
7798: @cindex search path control - for user applications
7799:
7800: Your application may need to search files in several directories, like
7801: @code{included} does. To facilitate this, Gforth allows you to define
7802: and use your own search paths, by providing generic equivalents of the
7803: Forth search path words:
7804:
7805:
7806: doc-.path
7807: doc-path+
7808: doc-path=
7809: doc-open-path-file
7810:
7811:
7812: Here's an example of creating a search path:
7813:
7814: @example
7815: \ Make a buffer for the path:
7816: create mypath 100 chars , \ maximum length (is checked)
7817: 0 , \ real len
7818: 100 chars allot \ space for path
7819: @end example
7820:
7821: @c -------------------------------------------------------------
7822: @node Blocks, Other I/O, Files, Words
7823: @section Blocks
7824: @cindex I/O - blocks
7825: @cindex blocks
7826:
7827: When you run Gforth on a modern desk-top computer, it runs under the
7828: control of an operating system which provides certain services. One of
7829: these services is @var{file services}, which allows Forth source code
7830: and data to be stored in files and read into Gforth (@pxref{Files}).
7831:
7832: Traditionally, Forth has been an important programming language on
7833: systems where it has interfaced directly to the underlying hardware with
7834: no intervening operating system. Forth provides a mechanism, called
7835: @dfn{blocks}, for accessing mass storage on such systems.
7836:
7837: A block is a 1024-byte data area, which can be used to hold data or
7838: Forth source code. No structure is imposed on the contents of the
7839: block. A block is identified by its number; blocks are numbered
7840: contiguously from 1 to an implementation-defined maximum.
7841:
7842: A typical system that used blocks but no operating system might use a
7843: single floppy-disk drive for mass storage, with the disks formatted to
7844: provide 256-byte sectors. Blocks would be implemented by assigning the
7845: first four sectors of the disk to block 1, the second four sectors to
7846: block 2 and so on, up to the limit of the capacity of the disk. The disk
7847: would not contain any file system information, just the set of blocks.
7848:
7849: @cindex blocks file
7850: On systems that do provide file services, blocks are typically
7851: implemented by storing a sequence of blocks within a single @dfn{blocks
7852: file}. The size of the blocks file will be an exact multiple of 1024
7853: bytes, corresponding to the number of blocks it contains. This is the
7854: mechanism that Gforth uses.
7855:
7856: @cindex @file{blocks.fb}
7857: Only 1 blocks file can be open at a time. If you use block words without
7858: having specified a blocks file, Gforth defaults to the blocks file
7859: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
7860: locate a blocks file (@pxref{Forth Search Paths}).
7861:
7862: @cindex block buffers
7863: When you read and write blocks under program control, Gforth uses a
7864: number of @dfn{block buffers} as intermediate storage. These buffers are
7865: not used when you use @code{load} to interpret the contents of a block.
7866:
7867: The behaviour of the block buffers is directly analagous to that of a
7868: cache. Each block buffer has three states:
7869:
7870: @itemize @bullet
7871: @item
7872: Unassigned
7873: @item
7874: Assigned-clean
7875: @item
7876: Assigned-dirty
7877: @end itemize
7878:
7879: Initially, all block buffers are @i{unassigned}. In order to access a
7880: block, the block (specified by its block number) must be assigned to a
7881: block buffer.
7882:
7883: The assignment of a block to a block buffer is performed by @code{block}
7884: or @code{buffer}. Use @code{block} when you wish to modify the existing
7885: contents of a block. Use @code{buffer} when you don't care about the
7886: existing contents of the block@footnote{The ANS Forth definition of
7887: @code{buffer} is intended not to cause disk I/O; if the data associated
7888: with the particular block is already stored in a block buffer due to an
7889: earlier @code{block} command, @code{buffer} will return that block
7890: buffer and the existing contents of the block will be
7891: available. Otherwise, @code{buffer} will simply assign a new, empty
7892: block buffer for the block.}.
7893:
7894: Once a block has been assigned to a block buffer using @code{block} or
7895: @code{buffer}, that block buffer becomes the @i{current block buffer}
7896: and its state changes to @i{assigned-clean}. Data may only be
7897: manipulated (read or written) within the current block buffer.
7898:
7899: When the contents of the current block buffer has been modified it is
7900: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
7901: either abandon the changes (by doing nothing) or commit the changes,
7902: using @code{update}. Using @code{update} does not change the blocks
7903: file; it simply changes a block buffer's state to @i{assigned-dirty}.
7904:
7905: The word @code{flush} causes all @i{assigned-dirty} blocks to be
7906: written back to the blocks file on disk. Leaving Gforth using @code{bye}
7907: also causes a @code{flush} to be performed.
7908:
7909: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
7910: algorithm to assign a block buffer to a block. That means that any
7911: particular block can only be assigned to one specific block buffer,
7912: called (for the particular operation) the @i{victim buffer}. If the
7913: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
7914: the new block immediately. If it is @i{assigned-dirty} its current
7915: contents are written back to the blocks file on disk before it is
7916: allocated to the new block.
7917:
7918: Although no structure is imposed on the contents of a block, it is
7919: traditional to display the contents as 16 lines each of 64 characters. A
7920: block provides a single, continuous stream of input (for example, it
7921: acts as a single parse area) -- there are no end-of-line characters
7922: within a block, and no end-of-file character at the end of a
7923: block. There are two consequences of this:
7924:
7925: @itemize @bullet
7926: @item
7927: The last character of one line wraps straight into the first character
7928: of the following line
7929: @item
7930: The word @code{\} -- comment to end of line -- requires special
7931: treatment; in the context of a block it causes all characters until the
7932: end of the current 64-character ``line'' to be ignored.
7933: @end itemize
7934:
7935: In Gforth, when you use @code{block} with a non-existent block number,
7936: the current blocks file will be extended to the appropriate size and the
7937: block buffer will be initialised with spaces.
7938:
7939: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
7940: for details) but doesn't encourage the use of blocks; the mechanism is
7941: only provided for backward compatibility -- ANS Forth requires blocks to
7942: be available when files are.
7943:
7944: Common techniques that are used when working with blocks include:
7945:
7946: @itemize @bullet
7947: @item
7948: A screen editor that allows you to edit blocks without leaving the Forth
7949: environment.
7950: @item
7951: Shadow screens; where every code block has an associated block
7952: containing comments (for example: code in odd block numbers, comments in
7953: even block numbers). Typically, the block editor provides a convenient
7954: mechanism to toggle between code and comments.
7955: @item
7956: Load blocks; a single block (typically block 1) contains a number of
7957: @code{thru} commands which @code{load} the whole of the application.
7958: @end itemize
7959:
7960: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
7961: integrated into a Forth programming environment.
7962:
7963: @comment TODO what about errors on open-blocks?
7964:
7965: doc-open-blocks
7966: doc-use
7967: doc-get-block-fid
7968: doc-block-position
7969:
7970: doc-scr
7971: doc-list
7972:
7973: doc---gforthman-block
7974: doc-buffer
7975:
7976: doc-update
7977: doc-updated?
7978: doc-save-buffers
7979: doc-empty-buffers
7980: doc-empty-buffer
7981: doc-flush
7982:
7983: doc-load
7984: doc-thru
7985: doc-+load
7986: doc-+thru
7987: doc---gforthman--->
7988: doc-block-included
7989:
7990:
7991: @c -------------------------------------------------------------
7992: @node Other I/O, Programming Tools, Blocks, Words
7993: @section Other I/O
7994: @cindex I/O - keyboard and display
7995:
7996: @menu
7997: * Simple numeric output:: Predefined formats
7998: * Formatted numeric output:: Formatted (pictured) output
7999: * String Formats:: How Forth stores strings in memory
8000: * Displaying characters and strings:: Other stuff
8001: * Input:: Input
8002: @end menu
8003:
8004: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8005: @subsection Simple numeric output
8006: @cindex numeric output - simple/free-format
8007:
8008: The simplest output functions are those that display numbers from the
8009: data or floating-point stacks. Floating-point output is always displayed
8010: using base 10. Numbers displayed from the data stack use the value stored
8011: in @code{base}.
8012:
8013:
8014: doc-.
8015: doc-dec.
8016: doc-hex.
8017: doc-u.
8018: doc-.r
8019: doc-u.r
8020: doc-d.
8021: doc-ud.
8022: doc-d.r
8023: doc-ud.r
8024: doc-f.
8025: doc-fe.
8026: doc-fs.
8027:
8028:
8029: Examples of printing the number 1234.5678E23 in the different floating-point output
8030: formats are shown below:
8031:
8032: @example
8033: f. 123456779999999000000000000.
8034: fe. 123.456779999999E24
8035: fs. 1.23456779999999E26
8036: @end example
8037:
8038:
8039: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8040: @subsection Formatted numeric output
8041: @cindex formatted numeric output
8042: @cindex pictured numeric output
8043: @cindex numeric output - formatted
8044:
8045: Forth traditionally uses a technique called @dfn{pictured numeric
8046: output} for formatted printing of integers. In this technique, digits
8047: are extracted from the number (using the current output radix defined by
8048: @code{base}), converted to ASCII codes and appended to a string that is
8049: built in a scratch-pad area of memory (@pxref{core-idef,
8050: Implementation-defined options, Implementation-defined
8051: options}). Arbitrary characters can be appended to the string during the
8052: extraction process. The completed string is specified by an address
8053: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8054: under program control.
8055:
8056: All of the words described in the previous section for simple numeric
8057: output are implemented in Gforth using pictured numeric output.
8058:
8059: Three important things to remember about pictured numeric output:
8060:
8061: @itemize @bullet
8062: @item
8063: It always operates on double-precision numbers; to display a
8064: single-precision number, convert it first (for ways of doing this
8065: @pxref{Double precision}).
8066: @item
8067: It always treats the double-precision number as though it were
8068: unsigned. The examples below show ways of printing signed numbers.
8069: @item
8070: The string is built up from right to left; least significant digit first.
8071: @end itemize
8072:
8073:
8074: doc-<#
8075: doc-<<#
8076: doc-#
8077: doc-#s
8078: doc-hold
8079: doc-sign
8080: doc-#>
8081: doc-#>>
8082:
8083: doc-represent
8084:
8085:
8086: @noindent
8087: Here are some examples of using pictured numeric output:
8088:
8089: @example
8090: : my-u. ( u -- )
8091: \ Simplest use of pns.. behaves like Standard u.
8092: 0 \ convert to unsigned double
8093: <# \ start conversion
8094: #s \ convert all digits
8095: #> \ complete conversion
8096: TYPE SPACE ; \ display, with trailing space
8097:
8098: : cents-only ( u -- )
8099: 0 \ convert to unsigned double
8100: <# \ start conversion
8101: # # \ convert two least-significant digits
8102: #> \ complete conversion, discard other digits
8103: TYPE SPACE ; \ display, with trailing space
8104:
8105: : dollars-and-cents ( u -- )
8106: 0 \ convert to unsigned double
8107: <# \ start conversion
8108: # # \ convert two least-significant digits
8109: [char] . hold \ insert decimal point
8110: #s \ convert remaining digits
8111: [char] $ hold \ append currency symbol
8112: #> \ complete conversion
8113: TYPE SPACE ; \ display, with trailing space
8114:
8115: : my-. ( n -- )
8116: \ handling negatives.. behaves like Standard .
8117: s>d \ convert to signed double
8118: swap over dabs \ leave sign byte followed by unsigned double
8119: <# \ start conversion
8120: #s \ convert all digits
8121: rot sign \ get at sign byte, append "-" if needed
8122: #> \ complete conversion
8123: TYPE SPACE ; \ display, with trailing space
8124:
8125: : account. ( n -- )
8126: \ accountants don't like minus signs, they use braces
8127: \ for negative numbers
8128: s>d \ convert to signed double
8129: swap over dabs \ leave sign byte followed by unsigned double
8130: <# \ start conversion
8131: 2 pick \ get copy of sign byte
8132: 0< IF [char] ) hold THEN \ right-most character of output
8133: #s \ convert all digits
8134: rot \ get at sign byte
8135: 0< IF [char] ( hold THEN
8136: #> \ complete conversion
8137: TYPE SPACE ; \ display, with trailing space
8138: @end example
8139:
8140: Here are some examples of using these words:
8141:
8142: @example
8143: 1 my-u. 1
8144: hex -1 my-u. decimal FFFFFFFF
8145: 1 cents-only 01
8146: 1234 cents-only 34
8147: 2 dollars-and-cents $0.02
8148: 1234 dollars-and-cents $12.34
8149: 123 my-. 123
8150: -123 my. -123
8151: 123 account. 123
8152: -456 account. (456)
8153: @end example
8154:
8155:
8156: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8157: @subsection String Formats
8158: @cindex strings - see character strings
8159: @cindex character strings - formats
8160: @cindex I/O - see character strings
8161:
8162: Forth commonly uses two different methods for representing character
8163: strings:
8164:
8165: @itemize @bullet
8166: @item
8167: @cindex address of counted string
8168: @cindex counted string
8169: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8170: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8171: string and the string occupies the subsequent @i{n} char addresses in
8172: memory.
8173: @item
8174: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8175: of the string in characters, and @i{c-addr} is the address of the
8176: first byte of the string.
8177: @end itemize
8178:
8179: ANS Forth encourages the use of the second format when representing
8180: strings on the stack, whilst conceeding that the counted string format
8181: remains useful as a way of storing strings in memory.
8182:
8183:
8184: doc-count
8185:
8186:
8187: For words that move, copy and search for strings see @ref{Memory
8188: Blocks}. For words that display characters and strings see
8189: @ref{Displaying characters and strings}.
8190:
8191: @node Displaying characters and strings, Input, String Formats, Other I/O
8192: @subsection Displaying characters and strings
8193: @cindex characters - compiling and displaying
8194: @cindex character strings - compiling and displaying
8195:
8196: This section starts with a glossary of Forth words and ends with a set
8197: of examples.
8198:
8199:
8200: doc-bl
8201: doc-space
8202: doc-spaces
8203: doc-emit
8204: doc-toupper
8205: doc-."
8206: doc-.(
8207: doc-type
8208: doc-typewhite
8209: doc-cr
8210: @cindex cursor control
8211: doc-at-xy
8212: doc-page
8213: doc-s"
8214: doc-c"
8215: doc-char
8216: doc-[char]
8217: doc-sliteral
8218:
8219:
8220: @noindent
8221: As an example, consider the following text, stored in a file @file{test.fs}:
8222:
8223: @example
8224: .( text-1)
8225: : my-word
8226: ." text-2" cr
8227: .( text-3)
8228: ;
8229:
8230: ." text-4"
8231:
8232: : my-char
8233: [char] ALPHABET emit
8234: char emit
8235: ;
8236: @end example
8237:
8238: When you load this code into Gforth, the following output is generated:
8239:
8240: @example
8241: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
8242: @end example
8243:
8244: @itemize @bullet
8245: @item
8246: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
8247: is an immediate word; it behaves in the same way whether it is used inside
8248: or outside a colon definition.
8249: @item
8250: Message @code{text-4} is displayed because of Gforth's added interpretation
8251: semantics for @code{."}.
8252: @item
8253: Message @code{text-2} is @i{not} displayed, because the text interpreter
8254: performs the compilation semantics for @code{."} within the definition of
8255: @code{my-word}.
8256: @end itemize
8257:
8258: Here are some examples of executing @code{my-word} and @code{my-char}:
8259:
8260: @example
8261: @kbd{my-word @key{RET}} text-2
8262: ok
8263: @kbd{my-char fred @key{RET}} Af ok
8264: @kbd{my-char jim @key{RET}} Aj ok
8265: @end example
8266:
8267: @itemize @bullet
8268: @item
8269: Message @code{text-2} is displayed because of the run-time behaviour of
8270: @code{."}.
8271: @item
8272: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
8273: on the stack at run-time. @code{emit} always displays the character
8274: when @code{my-char} is executed.
8275: @item
8276: @code{char} parses a string at run-time and the second @code{emit} displays
8277: the first character of the string.
8278: @item
8279: If you type @code{see my-char} you can see that @code{[char]} discarded
8280: the text ``LPHABET'' and only compiled the display code for ``A'' into the
8281: definition of @code{my-char}.
8282: @end itemize
8283:
8284:
8285:
8286: @node Input, , Displaying characters and strings, Other I/O
8287: @subsection Input
8288: @cindex input
8289: @cindex I/O - see input
8290: @cindex parsing a string
8291:
8292: For ways of storing character strings in memory see @ref{String Formats}.
8293:
8294: @comment TODO examples for >number >float accept key key? pad parse word refill
8295: @comment then index them
8296:
8297:
8298: doc-key
8299: doc-key?
8300: doc-ekey
8301: doc-ekey?
8302: doc-ekey>char
8303: doc->number
8304: doc->float
8305: doc-accept
8306: doc-pad
8307: doc-parse
8308: doc-word
8309: doc-sword
8310: doc-(name)
8311: doc-refill
8312: @comment obsolescent words..
8313: doc-convert
8314: doc-query
8315: doc-expect
8316: doc-span
8317:
8318:
8319:
8320: @c -------------------------------------------------------------
8321: @node Programming Tools, Assembler and Code Words, Other I/O, Words
8322: @section Programming Tools
8323: @cindex programming tools
8324:
8325: @menu
8326: * Debugging:: Simple and quick.
8327: * Assertions:: Making your programs self-checking.
8328: * Singlestep Debugger:: Executing your program word by word.
8329: @end menu
8330:
8331: @node Debugging, Assertions, Programming Tools, Programming Tools
8332: @subsection Debugging
8333: @cindex debugging
8334:
8335: Languages with a slow edit/compile/link/test development loop tend to
8336: require sophisticated tracing/stepping debuggers to facilate
8337: productive debugging.
8338:
8339: A much better (faster) way in fast-compiling languages is to add
8340: printing code at well-selected places, let the program run, look at
8341: the output, see where things went wrong, add more printing code, etc.,
8342: until the bug is found.
8343:
8344: The simple debugging aids provided in @file{debugs.fs}
8345: are meant to support this style of debugging. In addition, there are
8346: words for non-destructively inspecting the stack and memory:
8347:
8348:
8349: doc-.s
8350: doc-f.s
8351:
8352:
8353: There is a word @code{.r} but it does @i{not} display the return
8354: stack! It is used for formatted numeric output.
8355:
8356:
8357: doc-depth
8358: doc-fdepth
8359: doc-clearstack
8360: doc-?
8361: doc-dump
8362:
8363:
8364: The word @code{~~} prints debugging information (by default the source
8365: location and the stack contents). It is easy to insert. If you use Emacs
8366: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
8367: query-replace them with nothing). The deferred words
8368: @code{printdebugdata} and @code{printdebugline} control the output of
8369: @code{~~}. The default source location output format works well with
8370: Emacs' compilation mode, so you can step through the program at the
8371: source level using @kbd{C-x `} (the advantage over a stepping debugger
8372: is that you can step in any direction and you know where the crash has
8373: happened or where the strange data has occurred).
8374:
8375: The default actions of @code{~~} clobber the contents of the pictured
8376: numeric output string, so you should not use @code{~~}, e.g., between
8377: @code{<#} and @code{#>}.
8378:
8379:
8380: doc-~~
8381: doc-printdebugdata
8382: doc-printdebugline
8383:
8384: doc-see
8385: doc-marker
8386:
8387:
8388: Here's an example of using @code{marker} at the start of a source file
8389: that you are debugging; it ensures that you only ever have one copy of
8390: the file's definitions compiled at any time:
8391:
8392: @example
8393: [IFDEF] my-code
8394: my-code
8395: [ENDIF]
8396:
8397: marker my-code
8398: init-included-files
8399:
8400: \ .. definitions start here
8401: \ .
8402: \ .
8403: \ end
8404: @end example
8405:
8406:
8407:
8408: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
8409: @subsection Assertions
8410: @cindex assertions
8411:
8412: It is a good idea to make your programs self-checking, especially if you
8413: make an assumption that may become invalid during maintenance (for
8414: example, that a certain field of a data structure is never zero). Gforth
8415: supports @dfn{assertions} for this purpose. They are used like this:
8416:
8417: @example
8418: assert( @i{flag} )
8419: @end example
8420:
8421: The code between @code{assert(} and @code{)} should compute a flag, that
8422: should be true if everything is alright and false otherwise. It should
8423: not change anything else on the stack. The overall stack effect of the
8424: assertion is @code{( -- )}. E.g.
8425:
8426: @example
8427: assert( 1 1 + 2 = ) \ what we learn in school
8428: assert( dup 0<> ) \ assert that the top of stack is not zero
8429: assert( false ) \ this code should not be reached
8430: @end example
8431:
8432: The need for assertions is different at different times. During
8433: debugging, we want more checking, in production we sometimes care more
8434: for speed. Therefore, assertions can be turned off, i.e., the assertion
8435: becomes a comment. Depending on the importance of an assertion and the
8436: time it takes to check it, you may want to turn off some assertions and
8437: keep others turned on. Gforth provides several levels of assertions for
8438: this purpose:
8439:
8440:
8441: doc-assert0(
8442: doc-assert1(
8443: doc-assert2(
8444: doc-assert3(
8445: doc-assert(
8446: doc-)
8447:
8448:
8449: The variable @code{assert-level} specifies the highest assertions that
8450: are turned on. I.e., at the default @code{assert-level} of one,
8451: @code{assert0(} and @code{assert1(} assertions perform checking, while
8452: @code{assert2(} and @code{assert3(} assertions are treated as comments.
8453:
8454: The value of @code{assert-level} is evaluated at compile-time, not at
8455: run-time. Therefore you cannot turn assertions on or off at run-time;
8456: you have to set the @code{assert-level} appropriately before compiling a
8457: piece of code. You can compile different pieces of code at different
8458: @code{assert-level}s (e.g., a trusted library at level 1 and
8459: newly-written code at level 3).
8460:
8461:
8462: doc-assert-level
8463:
8464:
8465: If an assertion fails, a message compatible with Emacs' compilation mode
8466: is produced and the execution is aborted (currently with @code{ABORT"}.
8467: If there is interest, we will introduce a special throw code. But if you
8468: intend to @code{catch} a specific condition, using @code{throw} is
8469: probably more appropriate than an assertion).
8470:
8471: Definitions in ANS Forth for these assertion words are provided
8472: in @file{compat/assert.fs}.
8473:
8474:
8475: @node Singlestep Debugger, , Assertions, Programming Tools
8476: @subsection Singlestep Debugger
8477: @cindex singlestep Debugger
8478: @cindex debugging Singlestep
8479:
8480: When you create a new word there's often the need to check whether it
8481: behaves correctly or not. You can do this by typing @code{dbg
8482: badword}. A debug session might look like this:
8483:
8484: @example
8485: : badword 0 DO i . LOOP ; ok
8486: 2 dbg badword
8487: : badword
8488: Scanning code...
8489:
8490: Nesting debugger ready!
8491:
8492: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
8493: 400D4740 8049F68 DO -> [ 0 ]
8494: 400D4744 804A0C8 i -> [ 1 ] 00000
8495: 400D4748 400C5E60 . -> 0 [ 0 ]
8496: 400D474C 8049D0C LOOP -> [ 0 ]
8497: 400D4744 804A0C8 i -> [ 1 ] 00001
8498: 400D4748 400C5E60 . -> 1 [ 0 ]
8499: 400D474C 8049D0C LOOP -> [ 0 ]
8500: 400D4758 804B384 ; -> ok
8501: @end example
8502:
8503: Each line displayed is one step. You always have to hit return to
8504: execute the next word that is displayed. If you don't want to execute
8505: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
8506: an overview what keys are available:
8507:
8508: @table @i
8509:
8510: @item @key{RET}
8511: Next; Execute the next word.
8512:
8513: @item n
8514: Nest; Single step through next word.
8515:
8516: @item u
8517: Unnest; Stop debugging and execute rest of word. If we got to this word
8518: with nest, continue debugging with the calling word.
8519:
8520: @item d
8521: Done; Stop debugging and execute rest.
8522:
8523: @item s
8524: Stop; Abort immediately.
8525:
8526: @end table
8527:
8528: Debugging large application with this mechanism is very difficult, because
8529: you have to nest very deeply into the program before the interesting part
8530: begins. This takes a lot of time.
8531:
8532: To do it more directly put a @code{BREAK:} command into your source code.
8533: When program execution reaches @code{BREAK:} the single step debugger is
8534: invoked and you have all the features described above.
8535:
8536: If you have more than one part to debug it is useful to know where the
8537: program has stopped at the moment. You can do this by the
8538: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
8539: string is typed out when the ``breakpoint'' is reached.
8540:
8541:
8542: doc-dbg
8543: doc-break:
8544: doc-break"
8545:
8546:
8547:
8548: @c -------------------------------------------------------------
8549: @node Assembler and Code Words, Threading Words, Programming Tools, Words
8550: @section Assembler and Code Words
8551: @cindex assembler
8552: @cindex code words
8553:
8554: @menu
8555: * Code and ;code::
8556: * Common Assembler:: Assembler Syntax
8557: * Common Disassembler::
8558: * 386 Assembler:: Deviations and special cases
8559: * Alpha Assembler:: Deviations and special cases
8560: * MIPS assembler:: Deviations and special cases
8561: * Other assemblers:: How to write them
8562: @end menu
8563:
8564: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
8565: @subsection @code{Code} and @code{;code}
8566:
8567: Gforth provides some words for defining primitives (words written in
8568: machine code), and for defining the machine-code equivalent of
8569: @code{DOES>}-based defining words. However, the machine-independent
8570: nature of Gforth poses a few problems: First of all, Gforth runs on
8571: several architectures, so it can provide no standard assembler. What's
8572: worse is that the register allocation not only depends on the processor,
8573: but also on the @code{gcc} version and options used.
8574:
8575: The words that Gforth offers encapsulate some system dependences (e.g.,
8576: the header structure), so a system-independent assembler may be used in
8577: Gforth. If you do not have an assembler, you can compile machine code
8578: directly with @code{,} and @code{c,}@footnote{This isn't portable,
8579: because these words emit stuff in @i{data} space; it works because
8580: Gforth has unified code/data spaces. Assembler isn't likely to be
8581: portable anyway.}.
8582:
8583:
8584: doc-assembler
8585: doc-init-asm
8586: doc-code
8587: doc-end-code
8588: doc-;code
8589: doc-flush-icache
8590:
8591:
8592: If @code{flush-icache} does not work correctly, @code{code} words
8593: etc. will not work (reliably), either.
8594:
8595: The typical usage of these @code{code} words can be shown most easily by
8596: analogy to the equivalent high-level defining words:
8597:
8598: @example
8599: : foo code foo
8600: <high-level Forth words> <assembler>
8601: ; end-code
8602:
8603: : bar : bar
8604: <high-level Forth words> <high-level Forth words>
8605: CREATE CREATE
8606: <high-level Forth words> <high-level Forth words>
8607: DOES> ;code
8608: <high-level Forth words> <assembler>
8609: ; end-code
8610: @end example
8611:
8612: @code{flush-icache} is always present. The other words are rarely used
8613: and reside in @code{code.fs}, which is usually not loaded. You can load
8614: it with @code{require code.fs}.
8615:
8616: @cindex registers of the inner interpreter
8617: In the assembly code you will want to refer to the inner interpreter's
8618: registers (e.g., the data stack pointer) and you may want to use other
8619: registers for temporary storage. Unfortunately, the register allocation
8620: is installation-dependent.
8621:
8622: The easiest solution is to use explicit register declarations
8623: (@pxref{Explicit Reg Vars, , Variables in Specified Registers, gcc.info,
8624: GNU C Manual}) for all of the inner interpreter's registers: You have to
8625: compile Gforth with @code{-DFORCE_REG} (configure option
8626: @code{--enable-force-reg}) and the appropriate declarations must be
8627: present in the @code{machine.h} file (see @code{mips.h} for an example;
8628: you can find a full list of all declarable register symbols with
8629: @code{grep register engine.c}). If you give explicit registers to all
8630: variables that are declared at the beginning of @code{engine()}, you
8631: should be able to use the other caller-saved registers for temporary
8632: storage. Alternatively, you can use the @code{gcc} option
8633: @code{-ffixed-REG} (@pxref{Code Gen Options, , Options for Code
8634: Generation Conventions, gcc.info, GNU C Manual}) to reserve a register
8635: (however, this restriction on register allocation may slow Gforth
8636: significantly).
8637:
8638: If this solution is not viable (e.g., because @code{gcc} does not allow
8639: you to explicitly declare all the registers you need), you have to find
8640: out by looking at the code where the inner interpreter's registers
8641: reside and which registers can be used for temporary storage. You can
8642: get an assembly listing of the engine's code with @code{make engine.s}.
8643:
8644: In any case, it is good practice to abstract your assembly code from the
8645: actual register allocation. E.g., if the data stack pointer resides in
8646: register @code{$17}, create an alias for this register called @code{sp},
8647: and use that in your assembly code.
8648:
8649: @cindex code words, portable
8650: Another option for implementing normal and defining words efficiently
8651: is to add the desired functionality to the source of Gforth. For normal
8652: words you just have to edit @file{primitives} (@pxref{Automatic
8653: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
8654: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
8655: @file{prims2x.fs}, and possibly @file{cross.fs}.
8656:
8657: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
8658: @subsection Common Assembler
8659:
8660: The assemblers in Gforth generally use a postfix syntax, i.e., the
8661: instruction name follows the operands.
8662:
8663: The operands are passed in the usual order (the same that is used in the
8664: manual of the architecture). Since they all are Forth words, they have
8665: to be separated by spaces; you can also use Forth words to compute the
8666: operands.
8667:
8668: The instruction names usually end with a @code{,}. This makes it easier
8669: to visually separate instructions if you put several of them on one
8670: line; it also avoids shadowing other Forth words (e.g., @code{and}).
8671:
8672: Registers are usually specified by number; e.g., (decimal) @code{11}
8673: specifies registers R11 and F11 on the Alpha architecture (which one,
8674: depends on the instruction). The usual names are also available, e.g.,
8675: @code{s2} for R11 on Alpha.
8676:
8677: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
8678: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
8679: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
8680: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
8681: conditions are specified in a way specific to each assembler.
8682:
8683: Note that the register assignments of the Gforth engine can change
8684: between Gforth versions, or even between different compilations of the
8685: same Gforth version (e.g., if you use a different GCC version). So if
8686: you want to refer to Gforth's registers (e.g., the stack pointer or
8687: TOS), I recommend defining your own words for refering to these
8688: registers, and using them later on; then you can easily adapt to a
8689: changed register assignment. The stability of the register assignment
8690: is usually better if you build Gforth with @code{--enable-force-reg}.
8691:
8692: In particular, the resturn stack pointer and the instruction pointer are
8693: in memory in @code{gforth}, and usually in registers in
8694: @code{gforth-fast}. The most common use of these registers is to
8695: dispatch to the next word (the @code{next} routine). A portable way to
8696: do this is to jump to @code{' noop >code-address} (of course, this is
8697: less efficient than integrating the @code{next} code and scheduling it
8698: well).
8699:
8700: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
8701: @subsection Common Disassembler
8702:
8703: You can disassemble a @code{code} word with @code{see}
8704: (@pxref{Debugging}). You can disassemble a section of memory with
8705:
8706: doc-disasm
8707:
8708: The disassembler generally produces output that can be fed into the
8709: assembler (i.e., same syntax, etc.). It also includes additional
8710: information in comments. In particular, the address of the instruction
8711: is given in a comment before the instruction.
8712:
8713: @code{See} may display more or less than the actual code of the word,
8714: because the recognition of the end of the code is unreliable. You can
8715: use @code{disasm} if it did not display enough. It may display more, if
8716: the code word is not immediately followed by a named word. If you have
8717: something else there, you can follow the word with @code{align last @ ,}
8718: to ensure that the end is recognized.
8719:
8720: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
8721: @subsection 386 Assembler
8722:
8723: The 386 assembler and disassembler included in Gforth was written by
8724: Andrew McKewan; it is in the public domain.
8725:
8726: The disassembler displays code in prefix Intel syntax.
8727:
8728: The assembler uses an Intel-inspired postfix syntax with reversed
8729: parameters. As usual, a @code{,} is appended to the instruction names
8730: (including @code{rep,} etc.).
8731:
8732: The assembler is somewhat meager, missing a number of instructions
8733: (including FP) and absolute memory addressing modes.
8734:
8735: The registers have their usual names @code{eax} etc. Immediate values
8736: are indicated by postfixing them with @code{#}, e.g., @code{3 #}. Here
8737: are some examples of addressing modes:
8738:
8739: @example
8740: 3 #
8741: eax
8742: 100 [edi]
8743: 4 [ebx] [ecx]
8744: 0 [edi] [eax] *4 \ base register required!
8745: @end example
8746:
8747: Some example of instructions are:
8748:
8749: @example
8750: EAX EBX MOV, \ move ebx,eax
8751: 3 # EAX MOV, \ mov eax,3
8752: 100 [EDI] EAX MOV, \ mov eax,100[edi]
8753: 4 [EBX] [ECX] EAX MOV, \ mov eax,4[ebx][ecx]
8754: 16: EAX EBX MOV, \ mov bx,ax
8755: @end example
8756:
8757: You cannot use the prefix @code{16:} with immediate operands. The
8758: following forms are supported for binary instructions:
8759:
8760: @example
8761: <reg> <reg> <inst>
8762: <n> # <reg> <inst>
8763: <mem> <reg> <inst>
8764: <reg> <mem> <inst>
8765: @end example
8766:
8767: Immediate to memory is not supported. The shift/rotate syntax is:
8768:
8769: @example
8770: <reg/mem> shl,
8771: <reg/mem> 4 shl,
8772: <reg/mem> cl shl,
8773: @end example
8774:
8775: Precede string instructions (@code{movs,} etc.) with @code{byte} to get
8776: the byte version.
8777:
8778: The control structure words @code{if,} @code{until,} etc. must be
8779: preceded by one of these conditions: @code{0= 0< u< u> < > ov ecx0<>}.
8780: You can invert the condition with @code{not} (Note that most of these
8781: words shadow some Forth words when @code{assembler} is before
8782: @code{forth} in the search path, e.g., in code words). Currently the
8783: control structure words use one stack item, so you have to use
8784: @code{roll} instead of @code{cs-roll} to shuffle them (you can also use
8785: @code{swap} etc.).
8786:
8787: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
8788: @subsection Alpha Assembler
8789:
8790: The Alpha assembler and disassembler were originally written by Bernd
8791: Thallner.
8792:
8793: The register names @code{a0}--@code{a5} are not available to avoid
8794: shadowing hex numbers.
8795:
8796: Immediate forms of arithmetic instructions are distinguished by a
8797: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
8798: does not count as arithmetic instruction).
8799:
8800: You have to specify all operands to an instruction, even those that
8801: other assemblers consider optional, e.g., the destination register for
8802: @code{br,}, or the destination register and hint for @code{jmp,}.
8803:
8804: You can specify conditions for @code{if,} by removing the first @code{b}
8805: and the trailing @code{,} from a branch with a corresponding name; e.g.,
8806:
8807: @example
8808: 11 fgt if, \ if F11>0e
8809: ...
8810: endif,
8811: @end example
8812:
8813: @code{fbgt,} gives @code{fgt}.
8814:
8815: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
8816: @subsection MIPS assembler
8817:
8818: The MIPS assembler was originally written by Christian Pirker.
8819:
8820: Currently the assembler and disassembler only cover the MIPS-I
8821: architecture (R3000), and don't support FP instructions.
8822:
8823: The register names @code{$a0}--@code{$a3} are not available to avoid
8824: shadowing hex numbers.
8825:
8826: Because there is no way to distinguish registers from immediate values,
8827: you have to explicitly use the immediate forms of instructions, i.e.,
8828: @code{addiu,}, not just @code{addu,} (@command{as} does this
8829: implicitly).
8830:
8831: If the architecture manual specifies several formats for the instruction
8832: (e.g., for @code{jalr,}), you usually have to use the one with more
8833: arguments (i.e., two for @code{jalr,}). When in doubt, see
8834: @code{arch/mips/testasm.fs} for an example of correct use.
8835:
8836: Branches and jumps in the MIPS architecture have a delay slot. You have
8837: to fill it yourself (the simplest way is to use @code{nop,}), the
8838: assembler does not do it for you (unlike @command{as}). Even
8839: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
8840: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
8841: and @code{then,} just specify branch targets, they are not affected.
8842:
8843: Note that you must not put branches, jumps, or @code{li,} into the delay
8844: slot: @code{li,} may expand to several instructions, and control flow
8845: instructions may not be put into the branch delay slot in any case.
8846:
8847: For branches the argument specifying the target is a relative address;
8848: You have to add the address of the delay slot to get the absolute
8849: address.
8850:
8851: The MIPS architecture also has load delay slots and restrictions on
8852: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
8853: yourself to satisfy these restrictions, the assembler does not do it for
8854: you.
8855:
8856: You can specify the conditions for @code{if,} etc. by taking a
8857: conditional branch and leaving away the @code{b} at the start and the
8858: @code{,} at the end. E.g.,
8859:
8860: @example
8861: 4 5 eq if,
8862: ... \ do something if $4 equals $5
8863: then,
8864: @end example
8865:
8866: @node Other assemblers, , MIPS assembler, Assembler and Code Words
8867: @subsection Other assemblers
8868:
8869: If you want to contribute another assembler/disassembler, please contact
8870: us (@email{bug-gforth@@gnu.org}) to check if we have such an assembler
8871: already. If you are writing them from scratch, please use a similar
8872: syntax style as the one we use (i.e., postfix, commas at the end of the
8873: instruction names, @pxref{Common Assembler}); make the output of the
8874: disassembler be valid input for the assembler, and keep the style
8875: similar to the style we used.
8876:
8877: Hints on implementation: The most important part is to have a good test
8878: suite that contains all instructions. Once you have that, the rest is
8879: easy. For actual coding you can take a look at
8880: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
8881: the assembler and disassembler, avoiding redundancy and some potential
8882: bugs. You can also look at that file (and @pxref{Advanced does> usage})
8883: to get ideas how to factor a disassembler.
8884:
8885: Start with the disassembler, because it's easier to reuse data from the
8886: disassembler for the assembler than the other way round.
8887:
8888: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
8889: how simple it can be.
8890:
8891: @c -------------------------------------------------------------
8892: @node Threading Words, Locals, Assembler and Code Words, Words
8893: @section Threading Words
8894: @cindex threading words
8895:
8896: @cindex code address
8897: These words provide access to code addresses and other threading stuff
8898: in Gforth (and, possibly, other interpretive Forths). It more or less
8899: abstracts away the differences between direct and indirect threading
8900: (and, for direct threading, the machine dependences). However, at
8901: present this wordset is still incomplete. It is also pretty low-level;
8902: some day it will hopefully be made unnecessary by an internals wordset
8903: that abstracts implementation details away completely.
8904:
8905:
8906: doc-threading-method
8907: doc->code-address
8908: doc->does-code
8909: doc-code-address!
8910: doc-does-code!
8911: doc-does-handler!
8912: doc-/does-handler
8913:
8914:
8915: The code addresses produced by various defining words are produced by
8916: the following words:
8917:
8918:
8919: doc-docol:
8920: doc-docon:
8921: doc-dovar:
8922: doc-douser:
8923: doc-dodefer:
8924: doc-dofield:
8925:
8926:
8927: You can recognize words defined by a @code{CREATE}...@code{DOES>} word
8928: with @code{>does-code}. If the word was defined in that way, the value
8929: returned is non-zero and identifies the @code{DOES>} used by the
8930: defining word.
8931: @comment TODO should that be ``identifies the xt of the DOES> ??''
8932:
8933: @c -------------------------------------------------------------
8934: @node Locals, Structures, Threading Words, Words
8935: @section Locals
8936: @cindex locals
8937:
8938: Local variables can make Forth programming more enjoyable and Forth
8939: programs easier to read. Unfortunately, the locals of ANS Forth are
8940: laden with restrictions. Therefore, we provide not only the ANS Forth
8941: locals wordset, but also our own, more powerful locals wordset (we
8942: implemented the ANS Forth locals wordset through our locals wordset).
8943:
8944: The ideas in this section have also been published in the paper
8945: @cite{Automatic Scoping of Local Variables} by M. Anton Ertl, presented
8946: at EuroForth '94; it is available at
8947: @*@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz}.
8948:
8949: @menu
8950: * Gforth locals::
8951: * ANS Forth locals::
8952: @end menu
8953:
8954: @node Gforth locals, ANS Forth locals, Locals, Locals
8955: @subsection Gforth locals
8956: @cindex Gforth locals
8957: @cindex locals, Gforth style
8958:
8959: Locals can be defined with
8960:
8961: @example
8962: @{ local1 local2 ... -- comment @}
8963: @end example
8964: or
8965: @example
8966: @{ local1 local2 ... @}
8967: @end example
8968:
8969: E.g.,
8970: @example
8971: : max @{ n1 n2 -- n3 @}
8972: n1 n2 > if
8973: n1
8974: else
8975: n2
8976: endif ;
8977: @end example
8978:
8979: The similarity of locals definitions with stack comments is intended. A
8980: locals definition often replaces the stack comment of a word. The order
8981: of the locals corresponds to the order in a stack comment and everything
8982: after the @code{--} is really a comment.
8983:
8984: This similarity has one disadvantage: It is too easy to confuse locals
8985: declarations with stack comments, causing bugs and making them hard to
8986: find. However, this problem can be avoided by appropriate coding
8987: conventions: Do not use both notations in the same program. If you do,
8988: they should be distinguished using additional means, e.g. by position.
8989:
8990: @cindex types of locals
8991: @cindex locals types
8992: The name of the local may be preceded by a type specifier, e.g.,
8993: @code{F:} for a floating point value:
8994:
8995: @example
8996: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
8997: \ complex multiplication
8998: Ar Br f* Ai Bi f* f-
8999: Ar Bi f* Ai Br f* f+ ;
9000: @end example
9001:
9002: @cindex flavours of locals
9003: @cindex locals flavours
9004: @cindex value-flavoured locals
9005: @cindex variable-flavoured locals
9006: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9007: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9008: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9009: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9010: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9011: produces its address (which becomes invalid when the variable's scope is
9012: left). E.g., the standard word @code{emit} can be defined in terms of
9013: @code{type} like this:
9014:
9015: @example
9016: : emit @{ C^ char* -- @}
9017: char* 1 type ;
9018: @end example
9019:
9020: @cindex default type of locals
9021: @cindex locals, default type
9022: A local without type specifier is a @code{W:} local. Both flavours of
9023: locals are initialized with values from the data or FP stack.
9024:
9025: Currently there is no way to define locals with user-defined data
9026: structures, but we are working on it.
9027:
9028: Gforth allows defining locals everywhere in a colon definition. This
9029: poses the following questions:
9030:
9031: @menu
9032: * Where are locals visible by name?::
9033: * How long do locals live?::
9034: * Programming Style::
9035: * Implementation::
9036: @end menu
9037:
9038: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9039: @subsubsection Where are locals visible by name?
9040: @cindex locals visibility
9041: @cindex visibility of locals
9042: @cindex scope of locals
9043:
9044: Basically, the answer is that locals are visible where you would expect
9045: it in block-structured languages, and sometimes a little longer. If you
9046: want to restrict the scope of a local, enclose its definition in
9047: @code{SCOPE}...@code{ENDSCOPE}.
9048:
9049:
9050: doc-scope
9051: doc-endscope
9052:
9053:
9054: These words behave like control structure words, so you can use them
9055: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9056: arbitrary ways.
9057:
9058: If you want a more exact answer to the visibility question, here's the
9059: basic principle: A local is visible in all places that can only be
9060: reached through the definition of the local@footnote{In compiler
9061: construction terminology, all places dominated by the definition of the
9062: local.}. In other words, it is not visible in places that can be reached
9063: without going through the definition of the local. E.g., locals defined
9064: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9065: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9066: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
9067:
9068: The reasoning behind this solution is: We want to have the locals
9069: visible as long as it is meaningful. The user can always make the
9070: visibility shorter by using explicit scoping. In a place that can
9071: only be reached through the definition of a local, the meaning of a
9072: local name is clear. In other places it is not: How is the local
9073: initialized at the control flow path that does not contain the
9074: definition? Which local is meant, if the same name is defined twice in
9075: two independent control flow paths?
9076:
9077: This should be enough detail for nearly all users, so you can skip the
9078: rest of this section. If you really must know all the gory details and
9079: options, read on.
9080:
9081: In order to implement this rule, the compiler has to know which places
9082: are unreachable. It knows this automatically after @code{AHEAD},
9083: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9084: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9085: compiler that the control flow never reaches that place. If
9086: @code{UNREACHABLE} is not used where it could, the only consequence is
9087: that the visibility of some locals is more limited than the rule above
9088: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9089: lie to the compiler), buggy code will be produced.
9090:
9091:
9092: doc-unreachable
9093:
9094:
9095: Another problem with this rule is that at @code{BEGIN}, the compiler
9096: does not know which locals will be visible on the incoming
9097: back-edge. All problems discussed in the following are due to this
9098: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9099: loops as examples; the discussion also applies to @code{?DO} and other
9100: loops). Perhaps the most insidious example is:
9101: @example
9102: AHEAD
9103: BEGIN
9104: x
9105: [ 1 CS-ROLL ] THEN
9106: @{ x @}
9107: ...
9108: UNTIL
9109: @end example
9110:
9111: This should be legal according to the visibility rule. The use of
9112: @code{x} can only be reached through the definition; but that appears
9113: textually below the use.
9114:
9115: From this example it is clear that the visibility rules cannot be fully
9116: implemented without major headaches. Our implementation treats common
9117: cases as advertised and the exceptions are treated in a safe way: The
9118: compiler makes a reasonable guess about the locals visible after a
9119: @code{BEGIN}; if it is too pessimistic, the
9120: user will get a spurious error about the local not being defined; if the
9121: compiler is too optimistic, it will notice this later and issue a
9122: warning. In the case above the compiler would complain about @code{x}
9123: being undefined at its use. You can see from the obscure examples in
9124: this section that it takes quite unusual control structures to get the
9125: compiler into trouble, and even then it will often do fine.
9126:
9127: If the @code{BEGIN} is reachable from above, the most optimistic guess
9128: is that all locals visible before the @code{BEGIN} will also be
9129: visible after the @code{BEGIN}. This guess is valid for all loops that
9130: are entered only through the @code{BEGIN}, in particular, for normal
9131: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9132: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9133: compiler. When the branch to the @code{BEGIN} is finally generated by
9134: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9135: warns the user if it was too optimistic:
9136: @example
9137: IF
9138: @{ x @}
9139: BEGIN
9140: \ x ?
9141: [ 1 cs-roll ] THEN
9142: ...
9143: UNTIL
9144: @end example
9145:
9146: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9147: optimistically assumes that it lives until the @code{THEN}. It notices
9148: this difference when it compiles the @code{UNTIL} and issues a
9149: warning. The user can avoid the warning, and make sure that @code{x}
9150: is not used in the wrong area by using explicit scoping:
9151: @example
9152: IF
9153: SCOPE
9154: @{ x @}
9155: ENDSCOPE
9156: BEGIN
9157: [ 1 cs-roll ] THEN
9158: ...
9159: UNTIL
9160: @end example
9161:
9162: Since the guess is optimistic, there will be no spurious error messages
9163: about undefined locals.
9164:
9165: If the @code{BEGIN} is not reachable from above (e.g., after
9166: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9167: optimistic guess, as the locals visible after the @code{BEGIN} may be
9168: defined later. Therefore, the compiler assumes that no locals are
9169: visible after the @code{BEGIN}. However, the user can use
9170: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9171: visible at the BEGIN as at the point where the top control-flow stack
9172: item was created.
9173:
9174:
9175: doc-assume-live
9176:
9177:
9178: @noindent
9179: E.g.,
9180: @example
9181: @{ x @}
9182: AHEAD
9183: ASSUME-LIVE
9184: BEGIN
9185: x
9186: [ 1 CS-ROLL ] THEN
9187: ...
9188: UNTIL
9189: @end example
9190:
9191: Other cases where the locals are defined before the @code{BEGIN} can be
9192: handled by inserting an appropriate @code{CS-ROLL} before the
9193: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9194: behind the @code{ASSUME-LIVE}).
9195:
9196: Cases where locals are defined after the @code{BEGIN} (but should be
9197: visible immediately after the @code{BEGIN}) can only be handled by
9198: rearranging the loop. E.g., the ``most insidious'' example above can be
9199: arranged into:
9200: @example
9201: BEGIN
9202: @{ x @}
9203: ... 0=
9204: WHILE
9205: x
9206: REPEAT
9207: @end example
9208:
9209: @node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
9210: @subsubsection How long do locals live?
9211: @cindex locals lifetime
9212: @cindex lifetime of locals
9213:
9214: The right answer for the lifetime question would be: A local lives at
9215: least as long as it can be accessed. For a value-flavoured local this
9216: means: until the end of its visibility. However, a variable-flavoured
9217: local could be accessed through its address far beyond its visibility
9218: scope. Ultimately, this would mean that such locals would have to be
9219: garbage collected. Since this entails un-Forth-like implementation
9220: complexities, I adopted the same cowardly solution as some other
9221: languages (e.g., C): The local lives only as long as it is visible;
9222: afterwards its address is invalid (and programs that access it
9223: afterwards are erroneous).
9224:
9225: @node Programming Style, Implementation, How long do locals live?, Gforth locals
9226: @subsubsection Programming Style
9227: @cindex locals programming style
9228: @cindex programming style, locals
9229:
9230: The freedom to define locals anywhere has the potential to change
9231: programming styles dramatically. In particular, the need to use the
9232: return stack for intermediate storage vanishes. Moreover, all stack
9233: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9234: determined arguments) can be eliminated: If the stack items are in the
9235: wrong order, just write a locals definition for all of them; then
9236: write the items in the order you want.
9237:
9238: This seems a little far-fetched and eliminating stack manipulations is
9239: unlikely to become a conscious programming objective. Still, the number
9240: of stack manipulations will be reduced dramatically if local variables
9241: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
9242: a traditional implementation of @code{max}).
9243:
9244: This shows one potential benefit of locals: making Forth programs more
9245: readable. Of course, this benefit will only be realized if the
9246: programmers continue to honour the principle of factoring instead of
9247: using the added latitude to make the words longer.
9248:
9249: @cindex single-assignment style for locals
9250: Using @code{TO} can and should be avoided. Without @code{TO},
9251: every value-flavoured local has only a single assignment and many
9252: advantages of functional languages apply to Forth. I.e., programs are
9253: easier to analyse, to optimize and to read: It is clear from the
9254: definition what the local stands for, it does not turn into something
9255: different later.
9256:
9257: E.g., a definition using @code{TO} might look like this:
9258: @example
9259: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9260: u1 u2 min 0
9261: ?do
9262: addr1 c@@ addr2 c@@ -
9263: ?dup-if
9264: unloop exit
9265: then
9266: addr1 char+ TO addr1
9267: addr2 char+ TO addr2
9268: loop
9269: u1 u2 - ;
9270: @end example
9271: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9272: every loop iteration. @code{strcmp} is a typical example of the
9273: readability problems of using @code{TO}. When you start reading
9274: @code{strcmp}, you think that @code{addr1} refers to the start of the
9275: string. Only near the end of the loop you realize that it is something
9276: else.
9277:
9278: This can be avoided by defining two locals at the start of the loop that
9279: are initialized with the right value for the current iteration.
9280: @example
9281: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9282: addr1 addr2
9283: u1 u2 min 0
9284: ?do @{ s1 s2 @}
9285: s1 c@@ s2 c@@ -
9286: ?dup-if
9287: unloop exit
9288: then
9289: s1 char+ s2 char+
9290: loop
9291: 2drop
9292: u1 u2 - ;
9293: @end example
9294: Here it is clear from the start that @code{s1} has a different value
9295: in every loop iteration.
9296:
9297: @node Implementation, , Programming Style, Gforth locals
9298: @subsubsection Implementation
9299: @cindex locals implementation
9300: @cindex implementation of locals
9301:
9302: @cindex locals stack
9303: Gforth uses an extra locals stack. The most compelling reason for
9304: this is that the return stack is not float-aligned; using an extra stack
9305: also eliminates the problems and restrictions of using the return stack
9306: as locals stack. Like the other stacks, the locals stack grows toward
9307: lower addresses. A few primitives allow an efficient implementation:
9308:
9309:
9310: doc-@local#
9311: doc-f@local#
9312: doc-laddr#
9313: doc-lp+!#
9314: doc-lp!
9315: doc->l
9316: doc-f>l
9317:
9318:
9319: In addition to these primitives, some specializations of these
9320: primitives for commonly occurring inline arguments are provided for
9321: efficiency reasons, e.g., @code{@@local0} as specialization of
9322: @code{@@local#} for the inline argument 0. The following compiling words
9323: compile the right specialized version, or the general version, as
9324: appropriate:
9325:
9326:
9327: doc-compile-@local
9328: doc-compile-f@local
9329: doc-compile-lp+!
9330:
9331:
9332: Combinations of conditional branches and @code{lp+!#} like
9333: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9334: is taken) are provided for efficiency and correctness in loops.
9335:
9336: A special area in the dictionary space is reserved for keeping the
9337: local variable names. @code{@{} switches the dictionary pointer to this
9338: area and @code{@}} switches it back and generates the locals
9339: initializing code. @code{W:} etc.@ are normal defining words. This
9340: special area is cleared at the start of every colon definition.
9341:
9342: @cindex word list for defining locals
9343: A special feature of Gforth's dictionary is used to implement the
9344: definition of locals without type specifiers: every word list (aka
9345: vocabulary) has its own methods for searching
9346: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9347: with a special search method: When it is searched for a word, it
9348: actually creates that word using @code{W:}. @code{@{} changes the search
9349: order to first search the word list containing @code{@}}, @code{W:} etc.,
9350: and then the word list for defining locals without type specifiers.
9351:
9352: The lifetime rules support a stack discipline within a colon
9353: definition: The lifetime of a local is either nested with other locals
9354: lifetimes or it does not overlap them.
9355:
9356: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9357: pointer manipulation is generated. Between control structure words
9358: locals definitions can push locals onto the locals stack. @code{AGAIN}
9359: is the simplest of the other three control flow words. It has to
9360: restore the locals stack depth of the corresponding @code{BEGIN}
9361: before branching. The code looks like this:
9362: @format
9363: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9364: @code{branch} <begin>
9365: @end format
9366:
9367: @code{UNTIL} is a little more complicated: If it branches back, it
9368: must adjust the stack just like @code{AGAIN}. But if it falls through,
9369: the locals stack must not be changed. The compiler generates the
9370: following code:
9371: @format
9372: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9373: @end format
9374: The locals stack pointer is only adjusted if the branch is taken.
9375:
9376: @code{THEN} can produce somewhat inefficient code:
9377: @format
9378: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9379: <orig target>:
9380: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9381: @end format
9382: The second @code{lp+!#} adjusts the locals stack pointer from the
9383: level at the @i{orig} point to the level after the @code{THEN}. The
9384: first @code{lp+!#} adjusts the locals stack pointer from the current
9385: level to the level at the orig point, so the complete effect is an
9386: adjustment from the current level to the right level after the
9387: @code{THEN}.
9388:
9389: @cindex locals information on the control-flow stack
9390: @cindex control-flow stack items, locals information
9391: In a conventional Forth implementation a dest control-flow stack entry
9392: is just the target address and an orig entry is just the address to be
9393: patched. Our locals implementation adds a word list to every orig or dest
9394: item. It is the list of locals visible (or assumed visible) at the point
9395: described by the entry. Our implementation also adds a tag to identify
9396: the kind of entry, in particular to differentiate between live and dead
9397: (reachable and unreachable) orig entries.
9398:
9399: A few unusual operations have to be performed on locals word lists:
9400:
9401:
9402: doc-common-list
9403: doc-sub-list?
9404: doc-list-size
9405:
9406:
9407: Several features of our locals word list implementation make these
9408: operations easy to implement: The locals word lists are organised as
9409: linked lists; the tails of these lists are shared, if the lists
9410: contain some of the same locals; and the address of a name is greater
9411: than the address of the names behind it in the list.
9412:
9413: Another important implementation detail is the variable
9414: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9415: determine if they can be reached directly or only through the branch
9416: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9417: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9418: definition, by @code{BEGIN} and usually by @code{THEN}.
9419:
9420: Counted loops are similar to other loops in most respects, but
9421: @code{LEAVE} requires special attention: It performs basically the same
9422: service as @code{AHEAD}, but it does not create a control-flow stack
9423: entry. Therefore the information has to be stored elsewhere;
9424: traditionally, the information was stored in the target fields of the
9425: branches created by the @code{LEAVE}s, by organizing these fields into a
9426: linked list. Unfortunately, this clever trick does not provide enough
9427: space for storing our extended control flow information. Therefore, we
9428: introduce another stack, the leave stack. It contains the control-flow
9429: stack entries for all unresolved @code{LEAVE}s.
9430:
9431: Local names are kept until the end of the colon definition, even if
9432: they are no longer visible in any control-flow path. In a few cases
9433: this may lead to increased space needs for the locals name area, but
9434: usually less than reclaiming this space would cost in code size.
9435:
9436:
9437: @node ANS Forth locals, , Gforth locals, Locals
9438: @subsection ANS Forth locals
9439: @cindex locals, ANS Forth style
9440:
9441: The ANS Forth locals wordset does not define a syntax for locals, but
9442: words that make it possible to define various syntaxes. One of the
9443: possible syntaxes is a subset of the syntax we used in the Gforth locals
9444: wordset, i.e.:
9445:
9446: @example
9447: @{ local1 local2 ... -- comment @}
9448: @end example
9449: @noindent
9450: or
9451: @example
9452: @{ local1 local2 ... @}
9453: @end example
9454:
9455: The order of the locals corresponds to the order in a stack comment. The
9456: restrictions are:
9457:
9458: @itemize @bullet
9459: @item
9460: Locals can only be cell-sized values (no type specifiers are allowed).
9461: @item
9462: Locals can be defined only outside control structures.
9463: @item
9464: Locals can interfere with explicit usage of the return stack. For the
9465: exact (and long) rules, see the standard. If you don't use return stack
9466: accessing words in a definition using locals, you will be all right. The
9467: purpose of this rule is to make locals implementation on the return
9468: stack easier.
9469: @item
9470: The whole definition must be in one line.
9471: @end itemize
9472:
9473: Locals defined in this way behave like @code{VALUE}s
9474: (@pxref{Values}). I.e., they are initialized from the stack. Using their
9475: name produces their value. Their value can be changed using @code{TO}.
9476:
9477: Since this syntax is supported by Gforth directly, you need not do
9478: anything to use it. If you want to port a program using this syntax to
9479: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9480: syntax on the other system.
9481:
9482: Note that a syntax shown in the standard, section A.13 looks
9483: similar, but is quite different in having the order of locals
9484: reversed. Beware!
9485:
9486: The ANS Forth locals wordset itself consists of a word:
9487:
9488:
9489: doc-(local)
9490:
9491:
9492: The ANS Forth locals extension wordset defines a syntax using @code{locals|}, but it is so
9493: awful that we strongly recommend not to use it. We have implemented this
9494: syntax to make porting to Gforth easy, but do not document it here. The
9495: problem with this syntax is that the locals are defined in an order
9496: reversed with respect to the standard stack comment notation, making
9497: programs harder to read, and easier to misread and miswrite. The only
9498: merit of this syntax is that it is easy to implement using the ANS Forth
9499: locals wordset.
9500:
9501:
9502: @c ----------------------------------------------------------
9503: @node Structures, Object-oriented Forth, Locals, Words
9504: @section Structures
9505: @cindex structures
9506: @cindex records
9507:
9508: This section presents the structure package that comes with Gforth. A
9509: version of the package implemented in ANS Forth is available in
9510: @file{compat/struct.fs}. This package was inspired by a posting on
9511: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9512: possibly John Hayes). A version of this section has been published in
9513: ???. Marcel Hendrix provided helpful comments.
9514:
9515: @menu
9516: * Why explicit structure support?::
9517: * Structure Usage::
9518: * Structure Naming Convention::
9519: * Structure Implementation::
9520: * Structure Glossary::
9521: @end menu
9522:
9523: @node Why explicit structure support?, Structure Usage, Structures, Structures
9524: @subsection Why explicit structure support?
9525:
9526: @cindex address arithmetic for structures
9527: @cindex structures using address arithmetic
9528: If we want to use a structure containing several fields, we could simply
9529: reserve memory for it, and access the fields using address arithmetic
9530: (@pxref{Address arithmetic}). As an example, consider a structure with
9531: the following fields
9532:
9533: @table @code
9534: @item a
9535: is a float
9536: @item b
9537: is a cell
9538: @item c
9539: is a float
9540: @end table
9541:
9542: Given the (float-aligned) base address of the structure we get the
9543: address of the field
9544:
9545: @table @code
9546: @item a
9547: without doing anything further.
9548: @item b
9549: with @code{float+}
9550: @item c
9551: with @code{float+ cell+ faligned}
9552: @end table
9553:
9554: It is easy to see that this can become quite tiring.
9555:
9556: Moreover, it is not very readable, because seeing a
9557: @code{cell+} tells us neither which kind of structure is
9558: accessed nor what field is accessed; we have to somehow infer the kind
9559: of structure, and then look up in the documentation, which field of
9560: that structure corresponds to that offset.
9561:
9562: Finally, this kind of address arithmetic also causes maintenance
9563: troubles: If you add or delete a field somewhere in the middle of the
9564: structure, you have to find and change all computations for the fields
9565: afterwards.
9566:
9567: So, instead of using @code{cell+} and friends directly, how
9568: about storing the offsets in constants:
9569:
9570: @example
9571: 0 constant a-offset
9572: 0 float+ constant b-offset
9573: 0 float+ cell+ faligned c-offset
9574: @end example
9575:
9576: Now we can get the address of field @code{x} with @code{x-offset
9577: +}. This is much better in all respects. Of course, you still
9578: have to change all later offset definitions if you add a field. You can
9579: fix this by declaring the offsets in the following way:
9580:
9581: @example
9582: 0 constant a-offset
9583: a-offset float+ constant b-offset
9584: b-offset cell+ faligned constant c-offset
9585: @end example
9586:
9587: Since we always use the offsets with @code{+}, we could use a defining
9588: word @code{cfield} that includes the @code{+} in the action of the
9589: defined word:
9590:
9591: @example
9592: : cfield ( n "name" -- )
9593: create ,
9594: does> ( name execution: addr1 -- addr2 )
9595: @@ + ;
9596:
9597: 0 cfield a
9598: 0 a float+ cfield b
9599: 0 b cell+ faligned cfield c
9600: @end example
9601:
9602: Instead of @code{x-offset +}, we now simply write @code{x}.
9603:
9604: The structure field words now can be used quite nicely. However,
9605: their definition is still a bit cumbersome: We have to repeat the
9606: name, the information about size and alignment is distributed before
9607: and after the field definitions etc. The structure package presented
9608: here addresses these problems.
9609:
9610: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9611: @subsection Structure Usage
9612: @cindex structure usage
9613:
9614: @cindex @code{field} usage
9615: @cindex @code{struct} usage
9616: @cindex @code{end-struct} usage
9617: You can define a structure for a (data-less) linked list with:
9618: @example
9619: struct
9620: cell% field list-next
9621: end-struct list%
9622: @end example
9623:
9624: With the address of the list node on the stack, you can compute the
9625: address of the field that contains the address of the next node with
9626: @code{list-next}. E.g., you can determine the length of a list
9627: with:
9628:
9629: @example
9630: : list-length ( list -- n )
9631: \ "list" is a pointer to the first element of a linked list
9632: \ "n" is the length of the list
9633: 0 BEGIN ( list1 n1 )
9634: over
9635: WHILE ( list1 n1 )
9636: 1+ swap list-next @@ swap
9637: REPEAT
9638: nip ;
9639: @end example
9640:
9641: You can reserve memory for a list node in the dictionary with
9642: @code{list% %allot}, which leaves the address of the list node on the
9643: stack. For the equivalent allocation on the heap you can use @code{list%
9644: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9645: use @code{list% %allocate}). You can get the the size of a list
9646: node with @code{list% %size} and its alignment with @code{list%
9647: %alignment}.
9648:
9649: Note that in ANS Forth the body of a @code{create}d word is
9650: @code{aligned} but not necessarily @code{faligned};
9651: therefore, if you do a:
9652: @example
9653: create @emph{name} foo% %allot
9654: @end example
9655:
9656: @noindent
9657: then the memory alloted for @code{foo%} is
9658: guaranteed to start at the body of @code{@emph{name}} only if
9659: @code{foo%} contains only character, cell and double fields.
9660:
9661: @cindex structures containing structures
9662: You can include a structure @code{foo%} as a field of
9663: another structure, like this:
9664: @example
9665: struct
9666: ...
9667: foo% field ...
9668: ...
9669: end-struct ...
9670: @end example
9671:
9672: @cindex structure extension
9673: @cindex extended records
9674: Instead of starting with an empty structure, you can extend an
9675: existing structure. E.g., a plain linked list without data, as defined
9676: above, is hardly useful; You can extend it to a linked list of integers,
9677: like this:@footnote{This feature is also known as @emph{extended
9678: records}. It is the main innovation in the Oberon language; in other
9679: words, adding this feature to Modula-2 led Wirth to create a new
9680: language, write a new compiler etc. Adding this feature to Forth just
9681: required a few lines of code.}
9682:
9683: @example
9684: list%
9685: cell% field intlist-int
9686: end-struct intlist%
9687: @end example
9688:
9689: @code{intlist%} is a structure with two fields:
9690: @code{list-next} and @code{intlist-int}.
9691:
9692: @cindex structures containing arrays
9693: You can specify an array type containing @emph{n} elements of
9694: type @code{foo%} like this:
9695:
9696: @example
9697: foo% @emph{n} *
9698: @end example
9699:
9700: You can use this array type in any place where you can use a normal
9701: type, e.g., when defining a @code{field}, or with
9702: @code{%allot}.
9703:
9704: @cindex first field optimization
9705: The first field is at the base address of a structure and the word
9706: for this field (e.g., @code{list-next}) actually does not change
9707: the address on the stack. You may be tempted to leave it away in the
9708: interest of run-time and space efficiency. This is not necessary,
9709: because the structure package optimizes this case and compiling such
9710: words does not generate any code. So, in the interest of readability
9711: and maintainability you should include the word for the field when
9712: accessing the field.
9713:
9714: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
9715: @subsection Structure Naming Convention
9716: @cindex structure naming convention
9717:
9718: The field names that come to (my) mind are often quite generic, and,
9719: if used, would cause frequent name clashes. E.g., many structures
9720: probably contain a @code{counter} field. The structure names
9721: that come to (my) mind are often also the logical choice for the names
9722: of words that create such a structure.
9723:
9724: Therefore, I have adopted the following naming conventions:
9725:
9726: @itemize @bullet
9727: @cindex field naming convention
9728: @item
9729: The names of fields are of the form
9730: @code{@emph{struct}-@emph{field}}, where
9731: @code{@emph{struct}} is the basic name of the structure, and
9732: @code{@emph{field}} is the basic name of the field. You can
9733: think of field words as converting the (address of the)
9734: structure into the (address of the) field.
9735:
9736: @cindex structure naming convention
9737: @item
9738: The names of structures are of the form
9739: @code{@emph{struct}%}, where
9740: @code{@emph{struct}} is the basic name of the structure.
9741: @end itemize
9742:
9743: This naming convention does not work that well for fields of extended
9744: structures; e.g., the integer list structure has a field
9745: @code{intlist-int}, but has @code{list-next}, not
9746: @code{intlist-next}.
9747:
9748: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
9749: @subsection Structure Implementation
9750: @cindex structure implementation
9751: @cindex implementation of structures
9752:
9753: The central idea in the implementation is to pass the data about the
9754: structure being built on the stack, not in some global
9755: variable. Everything else falls into place naturally once this design
9756: decision is made.
9757:
9758: The type description on the stack is of the form @emph{align
9759: size}. Keeping the size on the top-of-stack makes dealing with arrays
9760: very simple.
9761:
9762: @code{field} is a defining word that uses @code{Create}
9763: and @code{DOES>}. The body of the field contains the offset
9764: of the field, and the normal @code{DOES>} action is simply:
9765:
9766: @example
9767: @@ +
9768: @end example
9769:
9770: @noindent
9771: i.e., add the offset to the address, giving the stack effect
9772: @i{addr1 -- addr2} for a field.
9773:
9774: @cindex first field optimization, implementation
9775: This simple structure is slightly complicated by the optimization
9776: for fields with offset 0, which requires a different
9777: @code{DOES>}-part (because we cannot rely on there being
9778: something on the stack if such a field is invoked during
9779: compilation). Therefore, we put the different @code{DOES>}-parts
9780: in separate words, and decide which one to invoke based on the
9781: offset. For a zero offset, the field is basically a noop; it is
9782: immediate, and therefore no code is generated when it is compiled.
9783:
9784: @node Structure Glossary, , Structure Implementation, Structures
9785: @subsection Structure Glossary
9786: @cindex structure glossary
9787:
9788:
9789: doc-%align
9790: doc-%alignment
9791: doc-%alloc
9792: doc-%allocate
9793: doc-%allot
9794: doc-cell%
9795: doc-char%
9796: doc-dfloat%
9797: doc-double%
9798: doc-end-struct
9799: doc-field
9800: doc-float%
9801: doc-naligned
9802: doc-sfloat%
9803: doc-%size
9804: doc-struct
9805:
9806:
9807: @c -------------------------------------------------------------
9808: @node Object-oriented Forth, Passing Commands to the OS, Structures, Words
9809: @section Object-oriented Forth
9810:
9811: Gforth comes with three packages for object-oriented programming:
9812: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
9813: is preloaded, so you have to @code{include} them before use. The most
9814: important differences between these packages (and others) are discussed
9815: in @ref{Comparison with other object models}. All packages are written
9816: in ANS Forth and can be used with any other ANS Forth.
9817:
9818: @menu
9819: * Why object-oriented programming?::
9820: * Object-Oriented Terminology::
9821: * Objects::
9822: * OOF::
9823: * Mini-OOF::
9824: * Comparison with other object models::
9825: @end menu
9826:
9827: @c ----------------------------------------------------------------
9828: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
9829: @subsection Why object-oriented programming?
9830: @cindex object-oriented programming motivation
9831: @cindex motivation for object-oriented programming
9832:
9833: Often we have to deal with several data structures (@emph{objects}),
9834: that have to be treated similarly in some respects, but differently in
9835: others. Graphical objects are the textbook example: circles, triangles,
9836: dinosaurs, icons, and others, and we may want to add more during program
9837: development. We want to apply some operations to any graphical object,
9838: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
9839: has to do something different for every kind of object.
9840: @comment TODO add some other operations eg perimeter, area
9841: @comment and tie in to concrete examples later..
9842:
9843: We could implement @code{draw} as a big @code{CASE}
9844: control structure that executes the appropriate code depending on the
9845: kind of object to be drawn. This would be not be very elegant, and,
9846: moreover, we would have to change @code{draw} every time we add
9847: a new kind of graphical object (say, a spaceship).
9848:
9849: What we would rather do is: When defining spaceships, we would tell
9850: the system: ``Here's how you @code{draw} a spaceship; you figure
9851: out the rest''.
9852:
9853: This is the problem that all systems solve that (rightfully) call
9854: themselves object-oriented; the object-oriented packages presented here
9855: solve this problem (and not much else).
9856: @comment TODO ?list properties of oo systems.. oo vs o-based?
9857:
9858: @c ------------------------------------------------------------------------
9859: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
9860: @subsection Object-Oriented Terminology
9861: @cindex object-oriented terminology
9862: @cindex terminology for object-oriented programming
9863:
9864: This section is mainly for reference, so you don't have to understand
9865: all of it right away. The terminology is mainly Smalltalk-inspired. In
9866: short:
9867:
9868: @table @emph
9869: @cindex class
9870: @item class
9871: a data structure definition with some extras.
9872:
9873: @cindex object
9874: @item object
9875: an instance of the data structure described by the class definition.
9876:
9877: @cindex instance variables
9878: @item instance variables
9879: fields of the data structure.
9880:
9881: @cindex selector
9882: @cindex method selector
9883: @cindex virtual function
9884: @item selector
9885: (or @emph{method selector}) a word (e.g.,
9886: @code{draw}) that performs an operation on a variety of data
9887: structures (classes). A selector describes @emph{what} operation to
9888: perform. In C++ terminology: a (pure) virtual function.
9889:
9890: @cindex method
9891: @item method
9892: the concrete definition that performs the operation
9893: described by the selector for a specific class. A method specifies
9894: @emph{how} the operation is performed for a specific class.
9895:
9896: @cindex selector invocation
9897: @cindex message send
9898: @cindex invoking a selector
9899: @item selector invocation
9900: a call of a selector. One argument of the call (the TOS (top-of-stack))
9901: is used for determining which method is used. In Smalltalk terminology:
9902: a message (consisting of the selector and the other arguments) is sent
9903: to the object.
9904:
9905: @cindex receiving object
9906: @item receiving object
9907: the object used for determining the method executed by a selector
9908: invocation. In the @file{objects.fs} model, it is the object that is on
9909: the TOS when the selector is invoked. (@emph{Receiving} comes from
9910: the Smalltalk @emph{message} terminology.)
9911:
9912: @cindex child class
9913: @cindex parent class
9914: @cindex inheritance
9915: @item child class
9916: a class that has (@emph{inherits}) all properties (instance variables,
9917: selectors, methods) from a @emph{parent class}. In Smalltalk
9918: terminology: The subclass inherits from the superclass. In C++
9919: terminology: The derived class inherits from the base class.
9920:
9921: @end table
9922:
9923: @c If you wonder about the message sending terminology, it comes from
9924: @c a time when each object had it's own task and objects communicated via
9925: @c message passing; eventually the Smalltalk developers realized that
9926: @c they can do most things through simple (indirect) calls. They kept the
9927: @c terminology.
9928:
9929: @c --------------------------------------------------------------
9930: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
9931: @subsection The @file{objects.fs} model
9932: @cindex objects
9933: @cindex object-oriented programming
9934:
9935: @cindex @file{objects.fs}
9936: @cindex @file{oof.fs}
9937:
9938: This section describes the @file{objects.fs} package. This material also
9939: has been published in @cite{Yet Another Forth Objects Package} by Anton
9940: Ertl and appeared in Forth Dimensions 19(2), pages 37--43
9941: (@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html}).
9942: @c McKewan's and Zsoter's packages
9943:
9944: This section assumes that you have read @ref{Structures}.
9945:
9946: The techniques on which this model is based have been used to implement
9947: the parser generator, Gray, and have also been used in Gforth for
9948: implementing the various flavours of word lists (hashed or not,
9949: case-sensitive or not, special-purpose word lists for locals etc.).
9950:
9951:
9952: @menu
9953: * Properties of the Objects model::
9954: * Basic Objects Usage::
9955: * The Objects base class::
9956: * Creating objects::
9957: * Object-Oriented Programming Style::
9958: * Class Binding::
9959: * Method conveniences::
9960: * Classes and Scoping::
9961: * Dividing classes::
9962: * Object Interfaces::
9963: * Objects Implementation::
9964: * Objects Glossary::
9965: @end menu
9966:
9967: Marcel Hendrix provided helpful comments on this section. Andras Zsoter
9968: and Bernd Paysan helped me with the related works section.
9969:
9970: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
9971: @subsubsection Properties of the @file{objects.fs} model
9972: @cindex @file{objects.fs} properties
9973:
9974: @itemize @bullet
9975: @item
9976: It is straightforward to pass objects on the stack. Passing
9977: selectors on the stack is a little less convenient, but possible.
9978:
9979: @item
9980: Objects are just data structures in memory, and are referenced by their
9981: address. You can create words for objects with normal defining words
9982: like @code{constant}. Likewise, there is no difference between instance
9983: variables that contain objects and those that contain other data.
9984:
9985: @item
9986: Late binding is efficient and easy to use.
9987:
9988: @item
9989: It avoids parsing, and thus avoids problems with state-smartness
9990: and reduced extensibility; for convenience there are a few parsing
9991: words, but they have non-parsing counterparts. There are also a few
9992: defining words that parse. This is hard to avoid, because all standard
9993: defining words parse (except @code{:noname}); however, such
9994: words are not as bad as many other parsing words, because they are not
9995: state-smart.
9996:
9997: @item
9998: It does not try to incorporate everything. It does a few things and does
9999: them well (IMO). In particular, this model was not designed to support
10000: information hiding (although it has features that may help); you can use
10001: a separate package for achieving this.
10002:
10003: @item
10004: It is layered; you don't have to learn and use all features to use this
10005: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10006: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
10007: are optional and independent of each other.
10008:
10009: @item
10010: An implementation in ANS Forth is available.
10011:
10012: @end itemize
10013:
10014:
10015: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10016: @subsubsection Basic @file{objects.fs} Usage
10017: @cindex basic objects usage
10018: @cindex objects, basic usage
10019:
10020: You can define a class for graphical objects like this:
10021:
10022: @cindex @code{class} usage
10023: @cindex @code{end-class} usage
10024: @cindex @code{selector} usage
10025: @example
10026: object class \ "object" is the parent class
10027: selector draw ( x y graphical -- )
10028: end-class graphical
10029: @end example
10030:
10031: This code defines a class @code{graphical} with an
10032: operation @code{draw}. We can perform the operation
10033: @code{draw} on any @code{graphical} object, e.g.:
10034:
10035: @example
10036: 100 100 t-rex draw
10037: @end example
10038:
10039: @noindent
10040: where @code{t-rex} is a word (say, a constant) that produces a
10041: graphical object.
10042:
10043: @comment TODO add a 2nd operation eg perimeter.. and use for
10044: @comment a concrete example
10045:
10046: @cindex abstract class
10047: How do we create a graphical object? With the present definitions,
10048: we cannot create a useful graphical object. The class
10049: @code{graphical} describes graphical objects in general, but not
10050: any concrete graphical object type (C++ users would call it an
10051: @emph{abstract class}); e.g., there is no method for the selector
10052: @code{draw} in the class @code{graphical}.
10053:
10054: For concrete graphical objects, we define child classes of the
10055: class @code{graphical}, e.g.:
10056:
10057: @cindex @code{overrides} usage
10058: @cindex @code{field} usage in class definition
10059: @example
10060: graphical class \ "graphical" is the parent class
10061: cell% field circle-radius
10062:
10063: :noname ( x y circle -- )
10064: circle-radius @@ draw-circle ;
10065: overrides draw
10066:
10067: :noname ( n-radius circle -- )
10068: circle-radius ! ;
10069: overrides construct
10070:
10071: end-class circle
10072: @end example
10073:
10074: Here we define a class @code{circle} as a child of @code{graphical},
10075: with field @code{circle-radius} (which behaves just like a field
10076: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10077: for the selectors @code{draw} and @code{construct} (@code{construct} is
10078: defined in @code{object}, the parent class of @code{graphical}).
10079:
10080: Now we can create a circle on the heap (i.e.,
10081: @code{allocate}d memory) with:
10082:
10083: @cindex @code{heap-new} usage
10084: @example
10085: 50 circle heap-new constant my-circle
10086: @end example
10087:
10088: @noindent
10089: @code{heap-new} invokes @code{construct}, thus
10090: initializing the field @code{circle-radius} with 50. We can draw
10091: this new circle at (100,100) with:
10092:
10093: @example
10094: 100 100 my-circle draw
10095: @end example
10096:
10097: @cindex selector invocation, restrictions
10098: @cindex class definition, restrictions
10099: Note: You can only invoke a selector if the object on the TOS
10100: (the receiving object) belongs to the class where the selector was
10101: defined or one of its descendents; e.g., you can invoke
10102: @code{draw} only for objects belonging to @code{graphical}
10103: or its descendents (e.g., @code{circle}). Immediately before
10104: @code{end-class}, the search order has to be the same as
10105: immediately after @code{class}.
10106:
10107: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10108: @subsubsection The @file{object.fs} base class
10109: @cindex @code{object} class
10110:
10111: When you define a class, you have to specify a parent class. So how do
10112: you start defining classes? There is one class available from the start:
10113: @code{object}. It is ancestor for all classes and so is the
10114: only class that has no parent. It has two selectors: @code{construct}
10115: and @code{print}.
10116:
10117: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10118: @subsubsection Creating objects
10119: @cindex creating objects
10120: @cindex object creation
10121: @cindex object allocation options
10122:
10123: @cindex @code{heap-new} discussion
10124: @cindex @code{dict-new} discussion
10125: @cindex @code{construct} discussion
10126: You can create and initialize an object of a class on the heap with
10127: @code{heap-new} ( ... class -- object ) and in the dictionary
10128: (allocation with @code{allot}) with @code{dict-new} (
10129: ... class -- object ). Both words invoke @code{construct}, which
10130: consumes the stack items indicated by "..." above.
10131:
10132: @cindex @code{init-object} discussion
10133: @cindex @code{class-inst-size} discussion
10134: If you want to allocate memory for an object yourself, you can get its
10135: alignment and size with @code{class-inst-size 2@@} ( class --
10136: align size ). Once you have memory for an object, you can initialize
10137: it with @code{init-object} ( ... class object -- );
10138: @code{construct} does only a part of the necessary work.
10139:
10140: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10141: @subsubsection Object-Oriented Programming Style
10142: @cindex object-oriented programming style
10143: @cindex programming style, object-oriented
10144:
10145: This section is not exhaustive.
10146:
10147: @cindex stack effects of selectors
10148: @cindex selectors and stack effects
10149: In general, it is a good idea to ensure that all methods for the
10150: same selector have the same stack effect: when you invoke a selector,
10151: you often have no idea which method will be invoked, so, unless all
10152: methods have the same stack effect, you will not know the stack effect
10153: of the selector invocation.
10154:
10155: One exception to this rule is methods for the selector
10156: @code{construct}. We know which method is invoked, because we
10157: specify the class to be constructed at the same place. Actually, I
10158: defined @code{construct} as a selector only to give the users a
10159: convenient way to specify initialization. The way it is used, a
10160: mechanism different from selector invocation would be more natural
10161: (but probably would take more code and more space to explain).
10162:
10163: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10164: @subsubsection Class Binding
10165: @cindex class binding
10166: @cindex early binding
10167:
10168: @cindex late binding
10169: Normal selector invocations determine the method at run-time depending
10170: on the class of the receiving object. This run-time selection is called
10171: @i{late binding}.
10172:
10173: Sometimes it's preferable to invoke a different method. For example,
10174: you might want to use the simple method for @code{print}ing
10175: @code{object}s instead of the possibly long-winded @code{print} method
10176: of the receiver class. You can achieve this by replacing the invocation
10177: of @code{print} with:
10178:
10179: @cindex @code{[bind]} usage
10180: @example
10181: [bind] object print
10182: @end example
10183:
10184: @noindent
10185: in compiled code or:
10186:
10187: @cindex @code{bind} usage
10188: @example
10189: bind object print
10190: @end example
10191:
10192: @cindex class binding, alternative to
10193: @noindent
10194: in interpreted code. Alternatively, you can define the method with a
10195: name (e.g., @code{print-object}), and then invoke it through the
10196: name. Class binding is just a (often more convenient) way to achieve
10197: the same effect; it avoids name clutter and allows you to invoke
10198: methods directly without naming them first.
10199:
10200: @cindex superclass binding
10201: @cindex parent class binding
10202: A frequent use of class binding is this: When we define a method
10203: for a selector, we often want the method to do what the selector does
10204: in the parent class, and a little more. There is a special word for
10205: this purpose: @code{[parent]}; @code{[parent]
10206: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10207: selector}}, where @code{@emph{parent}} is the parent
10208: class of the current class. E.g., a method definition might look like:
10209:
10210: @cindex @code{[parent]} usage
10211: @example
10212: :noname
10213: dup [parent] foo \ do parent's foo on the receiving object
10214: ... \ do some more
10215: ; overrides foo
10216: @end example
10217:
10218: @cindex class binding as optimization
10219: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10220: March 1997), Andrew McKewan presents class binding as an optimization
10221: technique. I recommend not using it for this purpose unless you are in
10222: an emergency. Late binding is pretty fast with this model anyway, so the
10223: benefit of using class binding is small; the cost of using class binding
10224: where it is not appropriate is reduced maintainability.
10225:
10226: While we are at programming style questions: You should bind
10227: selectors only to ancestor classes of the receiving object. E.g., say,
10228: you know that the receiving object is of class @code{foo} or its
10229: descendents; then you should bind only to @code{foo} and its
10230: ancestors.
10231:
10232: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10233: @subsubsection Method conveniences
10234: @cindex method conveniences
10235:
10236: In a method you usually access the receiving object pretty often. If
10237: you define the method as a plain colon definition (e.g., with
10238: @code{:noname}), you may have to do a lot of stack
10239: gymnastics. To avoid this, you can define the method with @code{m:
10240: ... ;m}. E.g., you could define the method for
10241: @code{draw}ing a @code{circle} with
10242:
10243: @cindex @code{this} usage
10244: @cindex @code{m:} usage
10245: @cindex @code{;m} usage
10246: @example
10247: m: ( x y circle -- )
10248: ( x y ) this circle-radius @@ draw-circle ;m
10249: @end example
10250:
10251: @cindex @code{exit} in @code{m: ... ;m}
10252: @cindex @code{exitm} discussion
10253: @cindex @code{catch} in @code{m: ... ;m}
10254: When this method is executed, the receiver object is removed from the
10255: stack; you can access it with @code{this} (admittedly, in this
10256: example the use of @code{m: ... ;m} offers no advantage). Note
10257: that I specify the stack effect for the whole method (i.e. including
10258: the receiver object), not just for the code between @code{m:}
10259: and @code{;m}. You cannot use @code{exit} in
10260: @code{m:...;m}; instead, use
10261: @code{exitm}.@footnote{Moreover, for any word that calls
10262: @code{catch} and was defined before loading
10263: @code{objects.fs}, you have to redefine it like I redefined
10264: @code{catch}: @code{: catch this >r catch r> to-this ;}}
10265:
10266: @cindex @code{inst-var} usage
10267: You will frequently use sequences of the form @code{this
10268: @emph{field}} (in the example above: @code{this
10269: circle-radius}). If you use the field only in this way, you can
10270: define it with @code{inst-var} and eliminate the
10271: @code{this} before the field name. E.g., the @code{circle}
10272: class above could also be defined with:
10273:
10274: @example
10275: graphical class
10276: cell% inst-var radius
10277:
10278: m: ( x y circle -- )
10279: radius @@ draw-circle ;m
10280: overrides draw
10281:
10282: m: ( n-radius circle -- )
10283: radius ! ;m
10284: overrides construct
10285:
10286: end-class circle
10287: @end example
10288:
10289: @code{radius} can only be used in @code{circle} and its
10290: descendent classes and inside @code{m:...;m}.
10291:
10292: @cindex @code{inst-value} usage
10293: You can also define fields with @code{inst-value}, which is
10294: to @code{inst-var} what @code{value} is to
10295: @code{variable}. You can change the value of such a field with
10296: @code{[to-inst]}. E.g., we could also define the class
10297: @code{circle} like this:
10298:
10299: @example
10300: graphical class
10301: inst-value radius
10302:
10303: m: ( x y circle -- )
10304: radius draw-circle ;m
10305: overrides draw
10306:
10307: m: ( n-radius circle -- )
10308: [to-inst] radius ;m
10309: overrides construct
10310:
10311: end-class circle
10312: @end example
10313:
10314: Finally, you can define named methods with @code{:m}. One use of this
10315: feature is the definition of words that occur only in one class and are
10316: not intended to be overridden, but which still need method context
10317: (e.g., for accessing @code{inst-var}s). Another use is for methods that
10318: would be bound frequently, if defined anonymously.
10319:
10320:
10321: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
10322: @subsubsection Classes and Scoping
10323: @cindex classes and scoping
10324: @cindex scoping and classes
10325:
10326: Inheritance is frequent, unlike structure extension. This exacerbates
10327: the problem with the field name convention (@pxref{Structure Naming
10328: Convention}): One always has to remember in which class the field was
10329: originally defined; changing a part of the class structure would require
10330: changes for renaming in otherwise unaffected code.
10331:
10332: @cindex @code{inst-var} visibility
10333: @cindex @code{inst-value} visibility
10334: To solve this problem, I added a scoping mechanism (which was not in my
10335: original charter): A field defined with @code{inst-var} (or
10336: @code{inst-value}) is visible only in the class where it is defined and in
10337: the descendent classes of this class. Using such fields only makes
10338: sense in @code{m:}-defined methods in these classes anyway.
10339:
10340: This scoping mechanism allows us to use the unadorned field name,
10341: because name clashes with unrelated words become much less likely.
10342:
10343: @cindex @code{protected} discussion
10344: @cindex @code{private} discussion
10345: Once we have this mechanism, we can also use it for controlling the
10346: visibility of other words: All words defined after
10347: @code{protected} are visible only in the current class and its
10348: descendents. @code{public} restores the compilation
10349: (i.e. @code{current}) word list that was in effect before. If you
10350: have several @code{protected}s without an intervening
10351: @code{public} or @code{set-current}, @code{public}
10352: will restore the compilation word list in effect before the first of
10353: these @code{protected}s.
10354:
10355: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10356: @subsubsection Dividing classes
10357: @cindex Dividing classes
10358: @cindex @code{methods}...@code{end-methods}
10359:
10360: You may want to do the definition of methods separate from the
10361: definition of the class, its selectors, fields, and instance variables,
10362: i.e., separate the implementation from the definition. You can do this
10363: in the following way:
10364:
10365: @example
10366: graphical class
10367: inst-value radius
10368: end-class circle
10369:
10370: ... \ do some other stuff
10371:
10372: circle methods \ now we are ready
10373:
10374: m: ( x y circle -- )
10375: radius draw-circle ;m
10376: overrides draw
10377:
10378: m: ( n-radius circle -- )
10379: [to-inst] radius ;m
10380: overrides construct
10381:
10382: end-methods
10383: @end example
10384:
10385: You can use several @code{methods}...@code{end-methods} sections. The
10386: only things you can do to the class in these sections are: defining
10387: methods, and overriding the class's selectors. You must not define new
10388: selectors or fields.
10389:
10390: Note that you often have to override a selector before using it. In
10391: particular, you usually have to override @code{construct} with a new
10392: method before you can invoke @code{heap-new} and friends. E.g., you
10393: must not create a circle before the @code{overrides construct} sequence
10394: in the example above.
10395:
10396: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
10397: @subsubsection Object Interfaces
10398: @cindex object interfaces
10399: @cindex interfaces for objects
10400:
10401: In this model you can only call selectors defined in the class of the
10402: receiving objects or in one of its ancestors. If you call a selector
10403: with a receiving object that is not in one of these classes, the
10404: result is undefined; if you are lucky, the program crashes
10405: immediately.
10406:
10407: @cindex selectors common to hardly-related classes
10408: Now consider the case when you want to have a selector (or several)
10409: available in two classes: You would have to add the selector to a
10410: common ancestor class, in the worst case to @code{object}. You
10411: may not want to do this, e.g., because someone else is responsible for
10412: this ancestor class.
10413:
10414: The solution for this problem is interfaces. An interface is a
10415: collection of selectors. If a class implements an interface, the
10416: selectors become available to the class and its descendents. A class
10417: can implement an unlimited number of interfaces. For the problem
10418: discussed above, we would define an interface for the selector(s), and
10419: both classes would implement the interface.
10420:
10421: As an example, consider an interface @code{storage} for
10422: writing objects to disk and getting them back, and a class
10423: @code{foo} that implements it. The code would look like this:
10424:
10425: @cindex @code{interface} usage
10426: @cindex @code{end-interface} usage
10427: @cindex @code{implementation} usage
10428: @example
10429: interface
10430: selector write ( file object -- )
10431: selector read1 ( file object -- )
10432: end-interface storage
10433:
10434: bar class
10435: storage implementation
10436:
10437: ... overrides write
10438: ... overrides read1
10439: ...
10440: end-class foo
10441: @end example
10442:
10443: @noindent
10444: (I would add a word @code{read} @i{( file -- object )} that uses
10445: @code{read1} internally, but that's beyond the point illustrated
10446: here.)
10447:
10448: Note that you cannot use @code{protected} in an interface; and
10449: of course you cannot define fields.
10450:
10451: In the Neon model, all selectors are available for all classes;
10452: therefore it does not need interfaces. The price you pay in this model
10453: is slower late binding, and therefore, added complexity to avoid late
10454: binding.
10455:
10456: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10457: @subsubsection @file{objects.fs} Implementation
10458: @cindex @file{objects.fs} implementation
10459:
10460: @cindex @code{object-map} discussion
10461: An object is a piece of memory, like one of the data structures
10462: described with @code{struct...end-struct}. It has a field
10463: @code{object-map} that points to the method map for the object's
10464: class.
10465:
10466: @cindex method map
10467: @cindex virtual function table
10468: The @emph{method map}@footnote{This is Self terminology; in C++
10469: terminology: virtual function table.} is an array that contains the
10470: execution tokens (@i{xt}s) of the methods for the object's class. Each
10471: selector contains an offset into a method map.
10472:
10473: @cindex @code{selector} implementation, class
10474: @code{selector} is a defining word that uses
10475: @code{CREATE} and @code{DOES>}. The body of the
10476: selector contains the offset; the @code{DOES>} action for a
10477: class selector is, basically:
10478:
10479: @example
10480: ( object addr ) @@ over object-map @@ + @@ execute
10481: @end example
10482:
10483: Since @code{object-map} is the first field of the object, it
10484: does not generate any code. As you can see, calling a selector has a
10485: small, constant cost.
10486:
10487: @cindex @code{current-interface} discussion
10488: @cindex class implementation and representation
10489: A class is basically a @code{struct} combined with a method
10490: map. During the class definition the alignment and size of the class
10491: are passed on the stack, just as with @code{struct}s, so
10492: @code{field} can also be used for defining class
10493: fields. However, passing more items on the stack would be
10494: inconvenient, so @code{class} builds a data structure in memory,
10495: which is accessed through the variable
10496: @code{current-interface}. After its definition is complete, the
10497: class is represented on the stack by a pointer (e.g., as parameter for
10498: a child class definition).
10499:
10500: A new class starts off with the alignment and size of its parent,
10501: and a copy of the parent's method map. Defining new fields extends the
10502: size and alignment; likewise, defining new selectors extends the
10503: method map. @code{overrides} just stores a new @i{xt} in the method
10504: map at the offset given by the selector.
10505:
10506: @cindex class binding, implementation
10507: Class binding just gets the @i{xt} at the offset given by the selector
10508: from the class's method map and @code{compile,}s (in the case of
10509: @code{[bind]}) it.
10510:
10511: @cindex @code{this} implementation
10512: @cindex @code{catch} and @code{this}
10513: @cindex @code{this} and @code{catch}
10514: I implemented @code{this} as a @code{value}. At the
10515: start of an @code{m:...;m} method the old @code{this} is
10516: stored to the return stack and restored at the end; and the object on
10517: the TOS is stored @code{TO this}. This technique has one
10518: disadvantage: If the user does not leave the method via
10519: @code{;m}, but via @code{throw} or @code{exit},
10520: @code{this} is not restored (and @code{exit} may
10521: crash). To deal with the @code{throw} problem, I have redefined
10522: @code{catch} to save and restore @code{this}; the same
10523: should be done with any word that can catch an exception. As for
10524: @code{exit}, I simply forbid it (as a replacement, there is
10525: @code{exitm}).
10526:
10527: @cindex @code{inst-var} implementation
10528: @code{inst-var} is just the same as @code{field}, with
10529: a different @code{DOES>} action:
10530: @example
10531: @@ this +
10532: @end example
10533: Similar for @code{inst-value}.
10534:
10535: @cindex class scoping implementation
10536: Each class also has a word list that contains the words defined with
10537: @code{inst-var} and @code{inst-value}, and its protected
10538: words. It also has a pointer to its parent. @code{class} pushes
10539: the word lists of the class and all its ancestors onto the search order stack,
10540: and @code{end-class} drops them.
10541:
10542: @cindex interface implementation
10543: An interface is like a class without fields, parent and protected
10544: words; i.e., it just has a method map. If a class implements an
10545: interface, its method map contains a pointer to the method map of the
10546: interface. The positive offsets in the map are reserved for class
10547: methods, therefore interface map pointers have negative
10548: offsets. Interfaces have offsets that are unique throughout the
10549: system, unlike class selectors, whose offsets are only unique for the
10550: classes where the selector is available (invokable).
10551:
10552: This structure means that interface selectors have to perform one
10553: indirection more than class selectors to find their method. Their body
10554: contains the interface map pointer offset in the class method map, and
10555: the method offset in the interface method map. The
10556: @code{does>} action for an interface selector is, basically:
10557:
10558: @example
10559: ( object selector-body )
10560: 2dup selector-interface @@ ( object selector-body object interface-offset )
10561: swap object-map @@ + @@ ( object selector-body map )
10562: swap selector-offset @@ + @@ execute
10563: @end example
10564:
10565: where @code{object-map} and @code{selector-offset} are
10566: first fields and generate no code.
10567:
10568: As a concrete example, consider the following code:
10569:
10570: @example
10571: interface
10572: selector if1sel1
10573: selector if1sel2
10574: end-interface if1
10575:
10576: object class
10577: if1 implementation
10578: selector cl1sel1
10579: cell% inst-var cl1iv1
10580:
10581: ' m1 overrides construct
10582: ' m2 overrides if1sel1
10583: ' m3 overrides if1sel2
10584: ' m4 overrides cl1sel2
10585: end-class cl1
10586:
10587: create obj1 object dict-new drop
10588: create obj2 cl1 dict-new drop
10589: @end example
10590:
10591: The data structure created by this code (including the data structure
10592: for @code{object}) is shown in the <a
10593: href="objects-implementation.eps">figure</a>, assuming a cell size of 4.
10594: @comment TODO add this diagram..
10595:
10596: @node Objects Glossary, , Objects Implementation, Objects
10597: @subsubsection @file{objects.fs} Glossary
10598: @cindex @file{objects.fs} Glossary
10599:
10600:
10601: doc---objects-bind
10602: doc---objects-<bind>
10603: doc---objects-bind'
10604: doc---objects-[bind]
10605: doc---objects-class
10606: doc---objects-class->map
10607: doc---objects-class-inst-size
10608: doc---objects-class-override!
10609: doc---objects-construct
10610: doc---objects-current'
10611: doc---objects-[current]
10612: doc---objects-current-interface
10613: doc---objects-dict-new
10614: doc---objects-drop-order
10615: doc---objects-end-class
10616: doc---objects-end-class-noname
10617: doc---objects-end-interface
10618: doc---objects-end-interface-noname
10619: doc---objects-end-methods
10620: doc---objects-exitm
10621: doc---objects-heap-new
10622: doc---objects-implementation
10623: doc---objects-init-object
10624: doc---objects-inst-value
10625: doc---objects-inst-var
10626: doc---objects-interface
10627: doc---objects-m:
10628: doc---objects-:m
10629: doc---objects-;m
10630: doc---objects-method
10631: doc---objects-methods
10632: doc---objects-object
10633: doc---objects-overrides
10634: doc---objects-[parent]
10635: doc---objects-print
10636: doc---objects-protected
10637: doc---objects-public
10638: doc---objects-push-order
10639: doc---objects-selector
10640: doc---objects-this
10641: doc---objects-<to-inst>
10642: doc---objects-[to-inst]
10643: doc---objects-to-this
10644: doc---objects-xt-new
10645:
10646:
10647: @c -------------------------------------------------------------
10648: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10649: @subsection The @file{oof.fs} model
10650: @cindex oof
10651: @cindex object-oriented programming
10652:
10653: @cindex @file{objects.fs}
10654: @cindex @file{oof.fs}
10655:
10656: This section describes the @file{oof.fs} package.
10657:
10658: The package described in this section has been used in bigFORTH since 1991, and
10659: used for two large applications: a chromatographic system used to
10660: create new medicaments, and a graphic user interface library (MINOS).
10661:
10662: You can find a description (in German) of @file{oof.fs} in @cite{Object
10663: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10664: 10(2), 1994.
10665:
10666: @menu
10667: * Properties of the OOF model::
10668: * Basic OOF Usage::
10669: * The OOF base class::
10670: * Class Declaration::
10671: * Class Implementation::
10672: @end menu
10673:
10674: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
10675: @subsubsection Properties of the @file{oof.fs} model
10676: @cindex @file{oof.fs} properties
10677:
10678: @itemize @bullet
10679: @item
10680: This model combines object oriented programming with information
10681: hiding. It helps you writing large application, where scoping is
10682: necessary, because it provides class-oriented scoping.
10683:
10684: @item
10685: Named objects, object pointers, and object arrays can be created,
10686: selector invocation uses the ``object selector'' syntax. Selector invocation
10687: to objects and/or selectors on the stack is a bit less convenient, but
10688: possible.
10689:
10690: @item
10691: Selector invocation and instance variable usage of the active object is
10692: straightforward, since both make use of the active object.
10693:
10694: @item
10695: Late binding is efficient and easy to use.
10696:
10697: @item
10698: State-smart objects parse selectors. However, extensibility is provided
10699: using a (parsing) selector @code{postpone} and a selector @code{'}.
10700:
10701: @item
10702: An implementation in ANS Forth is available.
10703:
10704: @end itemize
10705:
10706:
10707: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
10708: @subsubsection Basic @file{oof.fs} Usage
10709: @cindex @file{oof.fs} usage
10710:
10711: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
10712:
10713: You can define a class for graphical objects like this:
10714:
10715: @cindex @code{class} usage
10716: @cindex @code{class;} usage
10717: @cindex @code{method} usage
10718: @example
10719: object class graphical \ "object" is the parent class
10720: method draw ( x y graphical -- )
10721: class;
10722: @end example
10723:
10724: This code defines a class @code{graphical} with an
10725: operation @code{draw}. We can perform the operation
10726: @code{draw} on any @code{graphical} object, e.g.:
10727:
10728: @example
10729: 100 100 t-rex draw
10730: @end example
10731:
10732: @noindent
10733: where @code{t-rex} is an object or object pointer, created with e.g.
10734: @code{graphical : t-rex}.
10735:
10736: @cindex abstract class
10737: How do we create a graphical object? With the present definitions,
10738: we cannot create a useful graphical object. The class
10739: @code{graphical} describes graphical objects in general, but not
10740: any concrete graphical object type (C++ users would call it an
10741: @emph{abstract class}); e.g., there is no method for the selector
10742: @code{draw} in the class @code{graphical}.
10743:
10744: For concrete graphical objects, we define child classes of the
10745: class @code{graphical}, e.g.:
10746:
10747: @example
10748: graphical class circle \ "graphical" is the parent class
10749: cell var circle-radius
10750: how:
10751: : draw ( x y -- )
10752: circle-radius @@ draw-circle ;
10753:
10754: : init ( n-radius -- (
10755: circle-radius ! ;
10756: class;
10757: @end example
10758:
10759: Here we define a class @code{circle} as a child of @code{graphical},
10760: with a field @code{circle-radius}; it defines new methods for the
10761: selectors @code{draw} and @code{init} (@code{init} is defined in
10762: @code{object}, the parent class of @code{graphical}).
10763:
10764: Now we can create a circle in the dictionary with:
10765:
10766: @example
10767: 50 circle : my-circle
10768: @end example
10769:
10770: @noindent
10771: @code{:} invokes @code{init}, thus initializing the field
10772: @code{circle-radius} with 50. We can draw this new circle at (100,100)
10773: with:
10774:
10775: @example
10776: 100 100 my-circle draw
10777: @end example
10778:
10779: @cindex selector invocation, restrictions
10780: @cindex class definition, restrictions
10781: Note: You can only invoke a selector if the receiving object belongs to
10782: the class where the selector was defined or one of its descendents;
10783: e.g., you can invoke @code{draw} only for objects belonging to
10784: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
10785: mechanism will check if you try to invoke a selector that is not
10786: defined in this class hierarchy, so you'll get an error at compilation
10787: time.
10788:
10789:
10790: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
10791: @subsubsection The @file{oof.fs} base class
10792: @cindex @file{oof.fs} base class
10793:
10794: When you define a class, you have to specify a parent class. So how do
10795: you start defining classes? There is one class available from the start:
10796: @code{object}. You have to use it as ancestor for all classes. It is the
10797: only class that has no parent. Classes are also objects, except that
10798: they don't have instance variables; class manipulation such as
10799: inheritance or changing definitions of a class is handled through
10800: selectors of the class @code{object}.
10801:
10802: @code{object} provides a number of selectors:
10803:
10804: @itemize @bullet
10805: @item
10806: @code{class} for subclassing, @code{definitions} to add definitions
10807: later on, and @code{class?} to get type informations (is the class a
10808: subclass of the class passed on the stack?).
10809:
10810: doc---object-class
10811: doc---object-definitions
10812: doc---object-class?
10813:
10814:
10815: @item
10816: @code{init} and @code{dispose} as constructor and destructor of the
10817: object. @code{init} is invocated after the object's memory is allocated,
10818: while @code{dispose} also handles deallocation. Thus if you redefine
10819: @code{dispose}, you have to call the parent's dispose with @code{super
10820: dispose}, too.
10821:
10822: doc---object-init
10823: doc---object-dispose
10824:
10825:
10826: @item
10827: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
10828: @code{[]} to create named and unnamed objects and object arrays or
10829: object pointers.
10830:
10831: doc---object-new
10832: doc---object-new[]
10833: doc---object-:
10834: doc---object-ptr
10835: doc---object-asptr
10836: doc---object-[]
10837:
10838:
10839: @item
10840: @code{::} and @code{super} for explicit scoping. You should use explicit
10841: scoping only for super classes or classes with the same set of instance
10842: variables. Explicitly-scoped selectors use early binding.
10843:
10844: doc---object-::
10845: doc---object-super
10846:
10847:
10848: @item
10849: @code{self} to get the address of the object
10850:
10851: doc---object-self
10852:
10853:
10854: @item
10855: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
10856: pointers and instance defers.
10857:
10858: doc---object-bind
10859: doc---object-bound
10860: doc---object-link
10861: doc---object-is
10862:
10863:
10864: @item
10865: @code{'} to obtain selector tokens, @code{send} to invocate selectors
10866: form the stack, and @code{postpone} to generate selector invocation code.
10867:
10868: doc---object-'
10869: doc---object-postpone
10870:
10871:
10872: @item
10873: @code{with} and @code{endwith} to select the active object from the
10874: stack, and enable its scope. Using @code{with} and @code{endwith}
10875: also allows you to create code using selector @code{postpone} without being
10876: trapped by the state-smart objects.
10877:
10878: doc---object-with
10879: doc---object-endwith
10880:
10881:
10882: @end itemize
10883:
10884: @node Class Declaration, Class Implementation, The OOF base class, OOF
10885: @subsubsection Class Declaration
10886: @cindex class declaration
10887:
10888: @itemize @bullet
10889: @item
10890: Instance variables
10891:
10892: doc---oof-var
10893:
10894:
10895: @item
10896: Object pointers
10897:
10898: doc---oof-ptr
10899: doc---oof-asptr
10900:
10901:
10902: @item
10903: Instance defers
10904:
10905: doc---oof-defer
10906:
10907:
10908: @item
10909: Method selectors
10910:
10911: doc---oof-early
10912: doc---oof-method
10913:
10914:
10915: @item
10916: Class-wide variables
10917:
10918: doc---oof-static
10919:
10920:
10921: @item
10922: End declaration
10923:
10924: doc---oof-how:
10925: doc---oof-class;
10926:
10927:
10928: @end itemize
10929:
10930: @c -------------------------------------------------------------
10931: @node Class Implementation, , Class Declaration, OOF
10932: @subsubsection Class Implementation
10933: @cindex class implementation
10934:
10935: @c -------------------------------------------------------------
10936: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
10937: @subsection The @file{mini-oof.fs} model
10938: @cindex mini-oof
10939:
10940: Gforth's third object oriented Forth package is a 12-liner. It uses a
10941: mixture of the @file{object.fs} and the @file{oof.fs} syntax,
10942: and reduces to the bare minimum of features. This is based on a posting
10943: of Bernd Paysan in comp.arch.
10944:
10945: @menu
10946: * Basic Mini-OOF Usage::
10947: * Mini-OOF Example::
10948: * Mini-OOF Implementation::
10949: * Comparison with other object models::
10950: @end menu
10951:
10952: @c -------------------------------------------------------------
10953: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
10954: @subsubsection Basic @file{mini-oof.fs} Usage
10955: @cindex mini-oof usage
10956:
10957: There is a base class (@code{class}, which allocates one cell for the
10958: object pointer) plus seven other words: to define a method, a variable,
10959: a class; to end a class, to resolve binding, to allocate an object and
10960: to compile a class method.
10961: @comment TODO better description of the last one
10962:
10963:
10964: doc-object
10965: doc-method
10966: doc-var
10967: doc-class
10968: doc-end-class
10969: doc-defines
10970: doc-new
10971: doc-::
10972:
10973:
10974:
10975: @c -------------------------------------------------------------
10976: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
10977: @subsubsection Mini-OOF Example
10978: @cindex mini-oof example
10979:
10980: A short example shows how to use this package. This example, in slightly
10981: extended form, is supplied as @file{moof-exm.fs}
10982: @comment TODO could flesh this out with some comments from the Forthwrite article
10983:
10984: @example
10985: object class
10986: method init
10987: method draw
10988: end-class graphical
10989: @end example
10990:
10991: This code defines a class @code{graphical} with an
10992: operation @code{draw}. We can perform the operation
10993: @code{draw} on any @code{graphical} object, e.g.:
10994:
10995: @example
10996: 100 100 t-rex draw
10997: @end example
10998:
10999: where @code{t-rex} is an object or object pointer, created with e.g.
11000: @code{graphical new Constant t-rex}.
11001:
11002: For concrete graphical objects, we define child classes of the
11003: class @code{graphical}, e.g.:
11004:
11005: @example
11006: graphical class
11007: cell var circle-radius
11008: end-class circle \ "graphical" is the parent class
11009:
11010: :noname ( x y -- )
11011: circle-radius @@ draw-circle ; circle defines draw
11012: :noname ( r -- )
11013: circle-radius ! ; circle defines init
11014: @end example
11015:
11016: There is no implicit init method, so we have to define one. The creation
11017: code of the object now has to call init explicitely.
11018:
11019: @example
11020: circle new Constant my-circle
11021: 50 my-circle init
11022: @end example
11023:
11024: It is also possible to add a function to create named objects with
11025: automatic call of @code{init}, given that all objects have @code{init}
11026: on the same place:
11027:
11028: @example
11029: : new: ( .. o "name" -- )
11030: new dup Constant init ;
11031: 80 circle new: large-circle
11032: @end example
11033:
11034: We can draw this new circle at (100,100) with:
11035:
11036: @example
11037: 100 100 my-circle draw
11038: @end example
11039:
11040: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
11041: @subsubsection @file{mini-oof.fs} Implementation
11042:
11043: Object-oriented systems with late binding typically use a
11044: ``vtable''-approach: the first variable in each object is a pointer to a
11045: table, which contains the methods as function pointers. The vtable
11046: may also contain other information.
11047:
11048: So first, let's declare methods:
11049:
11050: @example
11051: : method ( m v -- m' v ) Create over , swap cell+ swap
11052: DOES> ( ... o -- ... ) @ over @ + @ execute ;
11053: @end example
11054:
11055: During method declaration, the number of methods and instance
11056: variables is on the stack (in address units). @code{method} creates
11057: one method and increments the method number. To execute a method, it
11058: takes the object, fetches the vtable pointer, adds the offset, and
11059: executes the @i{xt} stored there. Each method takes the object it is
11060: invoked from as top of stack parameter. The method itself should
11061: consume that object.
11062:
11063: Now, we also have to declare instance variables
11064:
11065: @example
11066: : var ( m v size -- m v' ) Create over , +
11067: DOES> ( o -- addr ) @ + ;
11068: @end example
11069:
11070: As before, a word is created with the current offset. Instance
11071: variables can have different sizes (cells, floats, doubles, chars), so
11072: all we do is take the size and add it to the offset. If your machine
11073: has alignment restrictions, put the proper @code{aligned} or
11074: @code{faligned} before the variable, to adjust the variable
11075: offset. That's why it is on the top of stack.
11076:
11077: We need a starting point (the base object) and some syntactic sugar:
11078:
11079: @example
11080: Create object 1 cells , 2 cells ,
11081: : class ( class -- class methods vars ) dup 2@ ;
11082: @end example
11083:
11084: For inheritance, the vtable of the parent object has to be
11085: copied when a new, derived class is declared. This gives all the
11086: methods of the parent class, which can be overridden, though.
11087:
11088: @example
11089: : end-class ( class methods vars -- )
11090: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
11091: cell+ dup cell+ r> rot @ 2 cells /string move ;
11092: @end example
11093:
11094: The first line creates the vtable, initialized with
11095: @code{noop}s. The second line is the inheritance mechanism, it
11096: copies the xts from the parent vtable.
11097:
11098: We still have no way to define new methods, let's do that now:
11099:
11100: @example
11101: : defines ( xt class -- ) ' >body @ + ! ;
11102: @end example
11103:
11104: To allocate a new object, we need a word, too:
11105:
11106: @example
11107: : new ( class -- o ) here over @ allot swap over ! ;
11108: @end example
11109:
11110: Sometimes derived classes want to access the method of the
11111: parent object. There are two ways to achieve this with Mini-OOF:
11112: first, you could use named words, and second, you could look up the
11113: vtable of the parent object.
11114:
11115: @example
11116: : :: ( class "name" -- ) ' >body @ + @ compile, ;
11117: @end example
11118:
11119:
11120: Nothing can be more confusing than a good example, so here is
11121: one. First let's declare a text object (called
11122: @code{button}), that stores text and position:
11123:
11124: @example
11125: object class
11126: cell var text
11127: cell var len
11128: cell var x
11129: cell var y
11130: method init
11131: method draw
11132: end-class button
11133: @end example
11134:
11135: @noindent
11136: Now, implement the two methods, @code{draw} and @code{init}:
11137:
11138: @example
11139: :noname ( o -- )
11140: >r r@ x @ r@ y @ at-xy r@ text @ r> len @ type ;
11141: button defines draw
11142: :noname ( addr u o -- )
11143: >r 0 r@ x ! 0 r@ y ! r@ len ! r> text ! ;
11144: button defines init
11145: @end example
11146:
11147: @noindent
11148: To demonstrate inheritance, we define a class @code{bold-button}, with no
11149: new data and no new methods:
11150:
11151: @example
11152: button class
11153: end-class bold-button
11154:
11155: : bold 27 emit ." [1m" ;
11156: : normal 27 emit ." [0m" ;
11157: @end example
11158:
11159: @noindent
11160: The class @code{bold-button} has a different draw method to
11161: @code{button}, but the new method is defined in terms of the draw method
11162: for @code{button}:
11163:
11164: @example
11165: :noname bold [ button :: draw ] normal ; bold-button defines draw
11166: @end example
11167:
11168: @noindent
11169: Finally, create two objects and apply methods:
11170:
11171: @example
11172: button new Constant foo
11173: s" thin foo" foo init
11174: page
11175: foo draw
11176: bold-button new Constant bar
11177: s" fat bar" bar init
11178: 1 bar y !
11179: bar draw
11180: @end example
11181:
11182:
11183: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11184: @subsection Comparison with other object models
11185: @cindex comparison of object models
11186: @cindex object models, comparison
11187:
11188: Many object-oriented Forth extensions have been proposed (@cite{A survey
11189: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11190: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11191: relation of the object models described here to two well-known and two
11192: closely-related (by the use of method maps) models.
11193:
11194: @cindex Neon model
11195: The most popular model currently seems to be the Neon model (see
11196: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11197: 1997) by Andrew McKewan) but this model has a number of limitations
11198: @footnote{A longer version of this critique can be
11199: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11200: Dimensions, May 1997) by Anton Ertl.}:
11201:
11202: @itemize @bullet
11203: @item
11204: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11205: to pass objects on the stack.
11206:
11207: @item
11208: It requires that the selector parses the input stream (at
11209: compile time); this leads to reduced extensibility and to bugs that are+
11210: hard to find.
11211:
11212: @item
11213: It allows using every selector to every object;
11214: this eliminates the need for classes, but makes it harder to create
11215: efficient implementations.
11216: @end itemize
11217:
11218: @cindex Pountain's object-oriented model
11219: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11220: Press, London, 1987) by Dick Pountain. However, it is not really about
11221: object-oriented programming, because it hardly deals with late
11222: binding. Instead, it focuses on features like information hiding and
11223: overloading that are characteristic of modular languages like Ada (83).
11224:
11225: @cindex Zsoter's object-oriented model
11226: In @cite{Does late binding have to be slow?} (Forth Dimensions 18(1)
11227: 1996, pages 31-35) Andras Zsoter describes a model that makes heavy use
11228: of an active object (like @code{this} in @file{objects.fs}): The active
11229: object is not only used for accessing all fields, but also specifies the
11230: receiving object of every selector invocation; you have to change the
11231: active object explicitly with @code{@{ ... @}}, whereas in
11232: @file{objects.fs} it changes more or less implicitly at @code{m:
11233: ... ;m}. Such a change at the method entry point is unnecessary with the
11234: Zsoter's model, because the receiving object is the active object
11235: already. On the other hand, the explicit change is absolutely necessary
11236: in that model, because otherwise no one could ever change the active
11237: object. An ANS Forth implementation of this model is available at
11238: @uref{http://www.forth.org/fig/oopf.html}.
11239:
11240: @cindex @file{oof.fs}, differences to other models
11241: The @file{oof.fs} model combines information hiding and overloading
11242: resolution (by keeping names in various word lists) with object-oriented
11243: programming. It sets the active object implicitly on method entry, but
11244: also allows explicit changing (with @code{>o...o>} or with
11245: @code{with...endwith}). It uses parsing and state-smart objects and
11246: classes for resolving overloading and for early binding: the object or
11247: class parses the selector and determines the method from this. If the
11248: selector is not parsed by an object or class, it performs a call to the
11249: selector for the active object (late binding), like Zsoter's model.
11250: Fields are always accessed through the active object. The big
11251: disadvantage of this model is the parsing and the state-smartness, which
11252: reduces extensibility and increases the opportunities for subtle bugs;
11253: essentially, you are only safe if you never tick or @code{postpone} an
11254: object or class (Bernd disagrees, but I (Anton) am not convinced).
11255:
11256: @cindex @file{mini-oof.fs}, differences to other models
11257: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11258: version of the @file{objects.fs} model, but syntactically it is a
11259: mixture of the @file{objects.fs} and @file{oof.fs} models.
11260:
11261: @c -------------------------------------------------------------
11262: @node Passing Commands to the OS, Keeping track of Time, Object-oriented Forth, Words
11263: @section Passing Commands to the Operating System
11264: @cindex operating system - passing commands
11265: @cindex shell commands
11266:
11267: Gforth allows you to pass an arbitrary string to the host operating
11268: system shell (if such a thing exists) for execution.
11269:
11270:
11271: doc-sh
11272: doc-system
11273: doc-$?
11274: doc-getenv
11275:
11276:
11277: @c -------------------------------------------------------------
11278: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
11279: @section Keeping track of Time
11280: @cindex time-related words
11281:
11282: Gforth implements time-related operations by making calls to the C
11283: library function, @code{gettimeofday}.
11284:
11285: doc-ms
11286: doc-time&date
11287:
11288:
11289:
11290: @c -------------------------------------------------------------
11291: @node Miscellaneous Words, , Keeping track of Time, Words
11292: @section Miscellaneous Words
11293: @cindex miscellaneous words
11294:
11295: @comment TODO find homes for these
11296:
11297: These section lists the ANS Forth words that are not documented
11298: elsewhere in this manual. Ultimately, they all need proper homes.
11299:
11300: doc-[compile]
11301:
11302:
11303: The following ANS Forth words are not currently supported by Gforth
11304: (@pxref{ANS conformance}):
11305:
11306: @code{EDITOR}
11307: @code{EMIT?}
11308: @code{FORGET}
11309:
11310: @c ******************************************************************
11311: @node Error messages, Tools, Words, Top
11312: @chapter Error messages
11313: @cindex error messages
11314: @cindex backtrace
11315:
11316: A typical Gforth error message looks like this:
11317:
11318: @example
11319: in file included from :-1
11320: in file included from ./yyy.fs:1
11321: ./xxx.fs:4: Invalid memory address
11322: bar
11323: ^^^
11324: $400E664C @@
11325: $400E6664 foo
11326: @end example
11327:
11328: The message identifying the error is @code{Invalid memory address}. The
11329: error happened when text-interpreting line 4 of the file
11330: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
11331: word on the line where the error happened, is pointed out (with
11332: @code{^^^}).
11333:
11334: The file containing the error was included in line 1 of @file{./yyy.fs},
11335: and @file{yyy.fs} was included from a non-file (in this case, by giving
11336: @file{yyy.fs} as command-line parameter to Gforth).
11337:
11338: At the end of the error message you find a return stack dump that can be
11339: interpreted as a backtrace (possibly empty). On top you find the top of
11340: the return stack when the @code{throw} happened, and at the bottom you
11341: find the return stack entry just above the return stack of the topmost
11342: text interpreter.
11343:
11344: To the right of most return stack entries you see a guess for the word
11345: that pushed that return stack entry as its return address. This gives a
11346: backtrace. In our case we see that @code{bar} called @code{foo}, and
11347: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
11348: address} exception).
11349:
11350: Note that the backtrace is not perfect: We don't know which return stack
11351: entries are return addresses (so we may get false positives); and in
11352: some cases (e.g., for @code{abort"}) we cannot determine from the return
11353: address the word that pushed the return address, so for some return
11354: addresses you see no names in the return stack dump.
11355:
11356: @cindex @code{catch} and backtraces
11357: The return stack dump represents the return stack at the time when a
11358: specific @code{throw} was executed. In programs that make use of
11359: @code{catch}, it is not necessarily clear which @code{throw} should be
11360: used for the return stack dump (e.g., consider one @code{throw} that
11361: indicates an error, which is caught, and during recovery another error
11362: happens; which @code{throw} should be used for the stack dump?). Gforth
11363: presents the return stack dump for the first @code{throw} after the last
11364: executed (not returned-to) @code{catch}; this works well in the usual
11365: case.
11366:
11367: @cindex @code{gforth-fast} and backtraces
11368: @cindex @code{gforth-fast}, difference from @code{gforth}
11369: @cindex backtraces with @code{gforth-fast}
11370: @cindex return stack dump with @code{gforth-fast}
11371: @code{gforth} is able to do a return stack dump for throws generated
11372: from primitives (e.g., invalid memory address, stack empty etc.);
11373: @code{gforth-fast} is only able to do a return stack dump from a
11374: directly called @code{throw} (including @code{abort} etc.). This is the
11375: only difference (apart from a speed factor of between 1.15 (K6-2) and
11376: 1.6 (21164A)) between @code{gforth} and @code{gforth-fast}. Given an
11377: exception caused by a primitive in @code{gforth-fast}, you will
11378: typically see no return stack dump at all; however, if the exception is
11379: caught by @code{catch} (e.g., for restoring some state), and then
11380: @code{throw}n again, the return stack dump will be for the first such
11381: @code{throw}.
11382:
11383: @c ******************************************************************
11384: @node Tools, ANS conformance, Error messages, Top
11385: @chapter Tools
11386:
11387: @menu
11388: * ANS Report:: Report the words used, sorted by wordset.
11389: @end menu
11390:
11391: See also @ref{Emacs and Gforth}.
11392:
11393: @node ANS Report, , Tools, Tools
11394: @section @file{ans-report.fs}: Report the words used, sorted by wordset
11395: @cindex @file{ans-report.fs}
11396: @cindex report the words used in your program
11397: @cindex words used in your program
11398:
11399: If you want to label a Forth program as ANS Forth Program, you must
11400: document which wordsets the program uses; for extension wordsets, it is
11401: helpful to list the words the program requires from these wordsets
11402: (because Forth systems are allowed to provide only some words of them).
11403:
11404: The @file{ans-report.fs} tool makes it easy for you to determine which
11405: words from which wordset and which non-ANS words your application
11406: uses. You simply have to include @file{ans-report.fs} before loading the
11407: program you want to check. After loading your program, you can get the
11408: report with @code{print-ans-report}. A typical use is to run this as
11409: batch job like this:
11410: @example
11411: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
11412: @end example
11413:
11414: The output looks like this (for @file{compat/control.fs}):
11415: @example
11416: The program uses the following words
11417: from CORE :
11418: : POSTPONE THEN ; immediate ?dup IF 0=
11419: from BLOCK-EXT :
11420: \
11421: from FILE :
11422: (
11423: @end example
11424:
11425: @subsection Caveats
11426:
11427: Note that @file{ans-report.fs} just checks which words are used, not whether
11428: they are used in an ANS Forth conforming way!
11429:
11430: Some words are defined in several wordsets in the
11431: standard. @file{ans-report.fs} reports them for only one of the
11432: wordsets, and not necessarily the one you expect. It depends on usage
11433: which wordset is the right one to specify. E.g., if you only use the
11434: compilation semantics of @code{S"}, it is a Core word; if you also use
11435: its interpretation semantics, it is a File word.
11436:
11437: @c ******************************************************************
11438: @node ANS conformance, Model, Tools, Top
11439: @chapter ANS conformance
11440: @cindex ANS conformance of Gforth
11441:
11442: To the best of our knowledge, Gforth is an
11443:
11444: ANS Forth System
11445: @itemize @bullet
11446: @item providing the Core Extensions word set
11447: @item providing the Block word set
11448: @item providing the Block Extensions word set
11449: @item providing the Double-Number word set
11450: @item providing the Double-Number Extensions word set
11451: @item providing the Exception word set
11452: @item providing the Exception Extensions word set
11453: @item providing the Facility word set
11454: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
11455: @item providing the File Access word set
11456: @item providing the File Access Extensions word set
11457: @item providing the Floating-Point word set
11458: @item providing the Floating-Point Extensions word set
11459: @item providing the Locals word set
11460: @item providing the Locals Extensions word set
11461: @item providing the Memory-Allocation word set
11462: @item providing the Memory-Allocation Extensions word set (that one's easy)
11463: @item providing the Programming-Tools word set
11464: @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
11465: @item providing the Search-Order word set
11466: @item providing the Search-Order Extensions word set
11467: @item providing the String word set
11468: @item providing the String Extensions word set (another easy one)
11469: @end itemize
11470:
11471: @cindex system documentation
11472: In addition, ANS Forth systems are required to document certain
11473: implementation choices. This chapter tries to meet these
11474: requirements. In many cases it gives a way to ask the system for the
11475: information instead of providing the information directly, in
11476: particular, if the information depends on the processor, the operating
11477: system or the installation options chosen, or if they are likely to
11478: change during the maintenance of Gforth.
11479:
11480: @comment The framework for the rest has been taken from pfe.
11481:
11482: @menu
11483: * The Core Words::
11484: * The optional Block word set::
11485: * The optional Double Number word set::
11486: * The optional Exception word set::
11487: * The optional Facility word set::
11488: * The optional File-Access word set::
11489: * The optional Floating-Point word set::
11490: * The optional Locals word set::
11491: * The optional Memory-Allocation word set::
11492: * The optional Programming-Tools word set::
11493: * The optional Search-Order word set::
11494: @end menu
11495:
11496:
11497: @c =====================================================================
11498: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
11499: @comment node-name, next, previous, up
11500: @section The Core Words
11501: @c =====================================================================
11502: @cindex core words, system documentation
11503: @cindex system documentation, core words
11504:
11505: @menu
11506: * core-idef:: Implementation Defined Options
11507: * core-ambcond:: Ambiguous Conditions
11508: * core-other:: Other System Documentation
11509: @end menu
11510:
11511: @c ---------------------------------------------------------------------
11512: @node core-idef, core-ambcond, The Core Words, The Core Words
11513: @subsection Implementation Defined Options
11514: @c ---------------------------------------------------------------------
11515: @cindex core words, implementation-defined options
11516: @cindex implementation-defined options, core words
11517:
11518:
11519: @table @i
11520: @item (Cell) aligned addresses:
11521: @cindex cell-aligned addresses
11522: @cindex aligned addresses
11523: processor-dependent. Gforth's alignment words perform natural alignment
11524: (e.g., an address aligned for a datum of size 8 is divisible by
11525: 8). Unaligned accesses usually result in a @code{-23 THROW}.
11526:
11527: @item @code{EMIT} and non-graphic characters:
11528: @cindex @code{EMIT} and non-graphic characters
11529: @cindex non-graphic characters and @code{EMIT}
11530: The character is output using the C library function (actually, macro)
11531: @code{putc}.
11532:
11533: @item character editing of @code{ACCEPT} and @code{EXPECT}:
11534: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
11535: @cindex editing in @code{ACCEPT} and @code{EXPECT}
11536: @cindex @code{ACCEPT}, editing
11537: @cindex @code{EXPECT}, editing
11538: This is modeled on the GNU readline library (@pxref{Readline
11539: Interaction, , Command Line Editing, readline, The GNU Readline
11540: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
11541: producing a full word completion every time you type it (instead of
11542: producing the common prefix of all completions). @xref{Command-line editing}.
11543:
11544: @item character set:
11545: @cindex character set
11546: The character set of your computer and display device. Gforth is
11547: 8-bit-clean (but some other component in your system may make trouble).
11548:
11549: @item Character-aligned address requirements:
11550: @cindex character-aligned address requirements
11551: installation-dependent. Currently a character is represented by a C
11552: @code{unsigned char}; in the future we might switch to @code{wchar_t}
11553: (Comments on that requested).
11554:
11555: @item character-set extensions and matching of names:
11556: @cindex character-set extensions and matching of names
11557: @cindex case-sensitivity for name lookup
11558: @cindex name lookup, case-sensitivity
11559: @cindex locale and case-sensitivity
11560: Any character except the ASCII NUL character can be used in a
11561: name. Matching is case-insensitive (except in @code{TABLE}s). The
11562: matching is performed using the C library function @code{strncasecmp}, whose
11563: function is probably influenced by the locale. E.g., the @code{C} locale
11564: does not know about accents and umlauts, so they are matched
11565: case-sensitively in that locale. For portability reasons it is best to
11566: write programs such that they work in the @code{C} locale. Then one can
11567: use libraries written by a Polish programmer (who might use words
11568: containing ISO Latin-2 encoded characters) and by a French programmer
11569: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
11570: funny results for some of the words (which ones, depends on the font you
11571: are using)). Also, the locale you prefer may not be available in other
11572: operating systems. Hopefully, Unicode will solve these problems one day.
11573:
11574: @item conditions under which control characters match a space delimiter:
11575: @cindex space delimiters
11576: @cindex control characters as delimiters
11577: If @code{WORD} is called with the space character as a delimiter, all
11578: white-space characters (as identified by the C macro @code{isspace()})
11579: are delimiters. @code{PARSE}, on the other hand, treats space like other
11580: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
11581: like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
11582: interpreter (aka text interpreter) by default, treats all white-space
11583: characters as delimiters.
11584:
11585: @item format of the control-flow stack:
11586: @cindex control-flow stack, format
11587: The data stack is used as control-flow stack. The size of a control-flow
11588: stack item in cells is given by the constant @code{cs-item-size}. At the
11589: time of this writing, an item consists of a (pointer to a) locals list
11590: (third), an address in the code (second), and a tag for identifying the
11591: item (TOS). The following tags are used: @code{defstart},
11592: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
11593: @code{scopestart}.
11594:
11595: @item conversion of digits > 35
11596: @cindex digits > 35
11597: The characters @code{[\]^_'} are the digits with the decimal value
11598: 36@minus{}41. There is no way to input many of the larger digits.
11599:
11600: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
11601: @cindex @code{EXPECT}, display after end of input
11602: @cindex @code{ACCEPT}, display after end of input
11603: The cursor is moved to the end of the entered string. If the input is
11604: terminated using the @kbd{Return} key, a space is typed.
11605:
11606: @item exception abort sequence of @code{ABORT"}:
11607: @cindex exception abort sequence of @code{ABORT"}
11608: @cindex @code{ABORT"}, exception abort sequence
11609: The error string is stored into the variable @code{"error} and a
11610: @code{-2 throw} is performed.
11611:
11612: @item input line terminator:
11613: @cindex input line terminator
11614: @cindex line terminator on input
11615: @cindex newline character on input
11616: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
11617: lines. One of these characters is typically produced when you type the
11618: @kbd{Enter} or @kbd{Return} key.
11619:
11620: @item maximum size of a counted string:
11621: @cindex maximum size of a counted string
11622: @cindex counted string, maximum size
11623: @code{s" /counted-string" environment? drop .}. Currently 255 characters
11624: on all ports, but this may change.
11625:
11626: @item maximum size of a parsed string:
11627: @cindex maximum size of a parsed string
11628: @cindex parsed string, maximum size
11629: Given by the constant @code{/line}. Currently 255 characters.
11630:
11631: @item maximum size of a definition name, in characters:
11632: @cindex maximum size of a definition name, in characters
11633: @cindex name, maximum length
11634: 31
11635:
11636: @item maximum string length for @code{ENVIRONMENT?}, in characters:
11637: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
11638: @cindex @code{ENVIRONMENT?} string length, maximum
11639: 31
11640:
11641: @item method of selecting the user input device:
11642: @cindex user input device, method of selecting
11643: The user input device is the standard input. There is currently no way to
11644: change it from within Gforth. However, the input can typically be
11645: redirected in the command line that starts Gforth.
11646:
11647: @item method of selecting the user output device:
11648: @cindex user output device, method of selecting
11649: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
11650: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
11651: output when the user output device is a terminal, otherwise the output
11652: is buffered.
11653:
11654: @item methods of dictionary compilation:
11655: What are we expected to document here?
11656:
11657: @item number of bits in one address unit:
11658: @cindex number of bits in one address unit
11659: @cindex address unit, size in bits
11660: @code{s" address-units-bits" environment? drop .}. 8 in all current
11661: ports.
11662:
11663: @item number representation and arithmetic:
11664: @cindex number representation and arithmetic
11665: Processor-dependent. Binary two's complement on all current ports.
11666:
11667: @item ranges for integer types:
11668: @cindex ranges for integer types
11669: @cindex integer types, ranges
11670: Installation-dependent. Make environmental queries for @code{MAX-N},
11671: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
11672: unsigned (and positive) types is 0. The lower bound for signed types on
11673: two's complement and one's complement machines machines can be computed
11674: by adding 1 to the upper bound.
11675:
11676: @item read-only data space regions:
11677: @cindex read-only data space regions
11678: @cindex data-space, read-only regions
11679: The whole Forth data space is writable.
11680:
11681: @item size of buffer at @code{WORD}:
11682: @cindex size of buffer at @code{WORD}
11683: @cindex @code{WORD} buffer size
11684: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
11685: shared with the pictured numeric output string. If overwriting
11686: @code{PAD} is acceptable, it is as large as the remaining dictionary
11687: space, although only as much can be sensibly used as fits in a counted
11688: string.
11689:
11690: @item size of one cell in address units:
11691: @cindex cell size
11692: @code{1 cells .}.
11693:
11694: @item size of one character in address units:
11695: @cindex char size
11696: @code{1 chars .}. 1 on all current ports.
11697:
11698: @item size of the keyboard terminal buffer:
11699: @cindex size of the keyboard terminal buffer
11700: @cindex terminal buffer, size
11701: Varies. You can determine the size at a specific time using @code{lp@@
11702: tib - .}. It is shared with the locals stack and TIBs of files that
11703: include the current file. You can change the amount of space for TIBs
11704: and locals stack at Gforth startup with the command line option
11705: @code{-l}.
11706:
11707: @item size of the pictured numeric output buffer:
11708: @cindex size of the pictured numeric output buffer
11709: @cindex pictured numeric output buffer, size
11710: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
11711: shared with @code{WORD}.
11712:
11713: @item size of the scratch area returned by @code{PAD}:
11714: @cindex size of the scratch area returned by @code{PAD}
11715: @cindex @code{PAD} size
11716: The remainder of dictionary space. @code{unused pad here - - .}.
11717:
11718: @item system case-sensitivity characteristics:
11719: @cindex case-sensitivity characteristics
11720: Dictionary searches are case-insensitive (except in
11721: @code{TABLE}s). However, as explained above under @i{character-set
11722: extensions}, the matching for non-ASCII characters is determined by the
11723: locale you are using. In the default @code{C} locale all non-ASCII
11724: characters are matched case-sensitively.
11725:
11726: @item system prompt:
11727: @cindex system prompt
11728: @cindex prompt
11729: @code{ ok} in interpret state, @code{ compiled} in compile state.
11730:
11731: @item division rounding:
11732: @cindex division rounding
11733: installation dependent. @code{s" floored" environment? drop .}. We leave
11734: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
11735: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
11736:
11737: @item values of @code{STATE} when true:
11738: @cindex @code{STATE} values
11739: -1.
11740:
11741: @item values returned after arithmetic overflow:
11742: On two's complement machines, arithmetic is performed modulo
11743: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
11744: arithmetic (with appropriate mapping for signed types). Division by zero
11745: typically results in a @code{-55 throw} (Floating-point unidentified
11746: fault), although a @code{-10 throw} (divide by zero) would be more
11747: appropriate.
11748:
11749: @item whether the current definition can be found after @t{DOES>}:
11750: @cindex @t{DOES>}, visibility of current definition
11751: No.
11752:
11753: @end table
11754:
11755: @c ---------------------------------------------------------------------
11756: @node core-ambcond, core-other, core-idef, The Core Words
11757: @subsection Ambiguous conditions
11758: @c ---------------------------------------------------------------------
11759: @cindex core words, ambiguous conditions
11760: @cindex ambiguous conditions, core words
11761:
11762: @table @i
11763:
11764: @item a name is neither a word nor a number:
11765: @cindex name not found
11766: @cindex undefined word
11767: @code{-13 throw} (Undefined word). Actually, @code{-13 bounce}, which
11768: preserves the data and FP stack, so you don't lose more work than
11769: necessary.
11770:
11771: @item a definition name exceeds the maximum length allowed:
11772: @cindex word name too long
11773: @code{-19 throw} (Word name too long)
11774:
11775: @item addressing a region not inside the various data spaces of the forth system:
11776: @cindex Invalid memory address
11777: The stacks, code space and header space are accessible. Machine code space is
11778: typically readable. Accessing other addresses gives results dependent on
11779: the operating system. On decent systems: @code{-9 throw} (Invalid memory
11780: address).
11781:
11782: @item argument type incompatible with parameter:
11783: @cindex argument type mismatch
11784: This is usually not caught. Some words perform checks, e.g., the control
11785: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
11786: mismatch).
11787:
11788: @item attempting to obtain the execution token of a word with undefined execution semantics:
11789: @cindex Interpreting a compile-only word, for @code{'} etc.
11790: @cindex execution token of words with undefined execution semantics
11791: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
11792: get an execution token for @code{compile-only-error} (which performs a
11793: @code{-14 throw} when executed).
11794:
11795: @item dividing by zero:
11796: @cindex dividing by zero
11797: @cindex floating point unidentified fault, integer division
11798: On better platforms, this produces a @code{-10 throw} (Division by
11799: zero); on other systems, this typically results in a @code{-55 throw}
11800: (Floating-point unidentified fault).
11801:
11802: @item insufficient data stack or return stack space:
11803: @cindex insufficient data stack or return stack space
11804: @cindex stack overflow
11805: @cindex address alignment exception, stack overflow
11806: @cindex Invalid memory address, stack overflow
11807: Depending on the operating system, the installation, and the invocation
11808: of Gforth, this is either checked by the memory management hardware, or
11809: it is not checked. If it is checked, you typically get a @code{-3 throw}
11810: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
11811: throw} (Invalid memory address) (depending on the platform and how you
11812: achieved the overflow) as soon as the overflow happens. If it is not
11813: checked, overflows typically result in mysterious illegal memory
11814: accesses, producing @code{-9 throw} (Invalid memory address) or
11815: @code{-23 throw} (Address alignment exception); they might also destroy
11816: the internal data structure of @code{ALLOCATE} and friends, resulting in
11817: various errors in these words.
11818:
11819: @item insufficient space for loop control parameters:
11820: @cindex insufficient space for loop control parameters
11821: like other return stack overflows.
11822:
11823: @item insufficient space in the dictionary:
11824: @cindex insufficient space in the dictionary
11825: @cindex dictionary overflow
11826: If you try to allot (either directly with @code{allot}, or indirectly
11827: with @code{,}, @code{create} etc.) more memory than available in the
11828: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
11829: to access memory beyond the end of the dictionary, the results are
11830: similar to stack overflows.
11831:
11832: @item interpreting a word with undefined interpretation semantics:
11833: @cindex interpreting a word with undefined interpretation semantics
11834: @cindex Interpreting a compile-only word
11835: For some words, we have defined interpretation semantics. For the
11836: others: @code{-14 throw} (Interpreting a compile-only word).
11837:
11838: @item modifying the contents of the input buffer or a string literal:
11839: @cindex modifying the contents of the input buffer or a string literal
11840: These are located in writable memory and can be modified.
11841:
11842: @item overflow of the pictured numeric output string:
11843: @cindex overflow of the pictured numeric output string
11844: @cindex pictured numeric output string, overflow
11845: @code{-17 throw} (Pictured numeric ouput string overflow).
11846:
11847: @item parsed string overflow:
11848: @cindex parsed string overflow
11849: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
11850:
11851: @item producing a result out of range:
11852: @cindex result out of range
11853: On two's complement machines, arithmetic is performed modulo
11854: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
11855: arithmetic (with appropriate mapping for signed types). Division by zero
11856: typically results in a @code{-10 throw} (divide by zero) or @code{-55
11857: throw} (floating point unidentified fault). @code{convert} and
11858: @code{>number} currently overflow silently.
11859:
11860: @item reading from an empty data or return stack:
11861: @cindex stack empty
11862: @cindex stack underflow
11863: @cindex return stack underflow
11864: The data stack is checked by the outer (aka text) interpreter after
11865: every word executed. If it has underflowed, a @code{-4 throw} (Stack
11866: underflow) is performed. Apart from that, stacks may be checked or not,
11867: depending on operating system, installation, and invocation. If they are
11868: caught by a check, they typically result in @code{-4 throw} (Stack
11869: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
11870: (Invalid memory address), depending on the platform and which stack
11871: underflows and by how much. Note that even if the system uses checking
11872: (through the MMU), your program may have to underflow by a significant
11873: number of stack items to trigger the reaction (the reason for this is
11874: that the MMU, and therefore the checking, works with a page-size
11875: granularity). If there is no checking, the symptoms resulting from an
11876: underflow are similar to those from an overflow. Unbalanced return
11877: stack errors result in a variaty of symptoms, including @code{-9 throw}
11878: (Invalid memory address) and Illegal Instruction (typically @code{-260
11879: throw}).
11880:
11881: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
11882: @cindex unexpected end of the input buffer
11883: @cindex zero-length string as a name
11884: @cindex Attempt to use zero-length string as a name
11885: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
11886: use zero-length string as a name). Words like @code{'} probably will not
11887: find what they search. Note that it is possible to create zero-length
11888: names with @code{nextname} (should it not?).
11889:
11890: @item @code{>IN} greater than input buffer:
11891: @cindex @code{>IN} greater than input buffer
11892: The next invocation of a parsing word returns a string with length 0.
11893:
11894: @item @code{RECURSE} appears after @code{DOES>}:
11895: @cindex @code{RECURSE} appears after @code{DOES>}
11896: Compiles a recursive call to the defining word, not to the defined word.
11897:
11898: @item argument input source different than current input source for @code{RESTORE-INPUT}:
11899: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
11900: @cindex argument type mismatch, @code{RESTORE-INPUT}
11901: @cindex @code{RESTORE-INPUT}, Argument type mismatch
11902: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
11903: the end of the file was reached), its source-id may be
11904: reused. Therefore, restoring an input source specification referencing a
11905: closed file may lead to unpredictable results instead of a @code{-12
11906: THROW}.
11907:
11908: In the future, Gforth may be able to restore input source specifications
11909: from other than the current input source.
11910:
11911: @item data space containing definitions gets de-allocated:
11912: @cindex data space containing definitions gets de-allocated
11913: Deallocation with @code{allot} is not checked. This typically results in
11914: memory access faults or execution of illegal instructions.
11915:
11916: @item data space read/write with incorrect alignment:
11917: @cindex data space read/write with incorrect alignment
11918: @cindex alignment faults
11919: @cindex address alignment exception
11920: Processor-dependent. Typically results in a @code{-23 throw} (Address
11921: alignment exception). Under Linux-Intel on a 486 or later processor with
11922: alignment turned on, incorrect alignment results in a @code{-9 throw}
11923: (Invalid memory address). There are reportedly some processors with
11924: alignment restrictions that do not report violations.
11925:
11926: @item data space pointer not properly aligned, @code{,}, @code{C,}:
11927: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
11928: Like other alignment errors.
11929:
11930: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
11931: Like other stack underflows.
11932:
11933: @item loop control parameters not available:
11934: @cindex loop control parameters not available
11935: Not checked. The counted loop words simply assume that the top of return
11936: stack items are loop control parameters and behave accordingly.
11937:
11938: @item most recent definition does not have a name (@code{IMMEDIATE}):
11939: @cindex most recent definition does not have a name (@code{IMMEDIATE})
11940: @cindex last word was headerless
11941: @code{abort" last word was headerless"}.
11942:
11943: @item name not defined by @code{VALUE} used by @code{TO}:
11944: @cindex name not defined by @code{VALUE} used by @code{TO}
11945: @cindex @code{TO} on non-@code{VALUE}s
11946: @cindex Invalid name argument, @code{TO}
11947: @code{-32 throw} (Invalid name argument) (unless name is a local or was
11948: defined by @code{CONSTANT}; in the latter case it just changes the constant).
11949:
11950: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
11951: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
11952: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
11953: @code{-13 throw} (Undefined word)
11954:
11955: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
11956: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
11957: Gforth behaves as if they were of the same type. I.e., you can predict
11958: the behaviour by interpreting all parameters as, e.g., signed.
11959:
11960: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
11961: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
11962: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
11963: compilation semantics of @code{TO}.
11964:
11965: @item String longer than a counted string returned by @code{WORD}:
11966: @cindex string longer than a counted string returned by @code{WORD}
11967: @cindex @code{WORD}, string overflow
11968: Not checked. The string will be ok, but the count will, of course,
11969: contain only the least significant bits of the length.
11970:
11971: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
11972: @cindex @code{LSHIFT}, large shift counts
11973: @cindex @code{RSHIFT}, large shift counts
11974: Processor-dependent. Typical behaviours are returning 0 and using only
11975: the low bits of the shift count.
11976:
11977: @item word not defined via @code{CREATE}:
11978: @cindex @code{>BODY} of non-@code{CREATE}d words
11979: @code{>BODY} produces the PFA of the word no matter how it was defined.
11980:
11981: @cindex @code{DOES>} of non-@code{CREATE}d words
11982: @code{DOES>} changes the execution semantics of the last defined word no
11983: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
11984: @code{CREATE , DOES>}.
11985:
11986: @item words improperly used outside @code{<#} and @code{#>}:
11987: Not checked. As usual, you can expect memory faults.
11988:
11989: @end table
11990:
11991:
11992: @c ---------------------------------------------------------------------
11993: @node core-other, , core-ambcond, The Core Words
11994: @subsection Other system documentation
11995: @c ---------------------------------------------------------------------
11996: @cindex other system documentation, core words
11997: @cindex core words, other system documentation
11998:
11999: @table @i
12000: @item nonstandard words using @code{PAD}:
12001: @cindex @code{PAD} use by nonstandard words
12002: None.
12003:
12004: @item operator's terminal facilities available:
12005: @cindex operator's terminal facilities available
12006: After processing the command line, Gforth goes into interactive mode,
12007: and you can give commands to Gforth interactively. The actual facilities
12008: available depend on how you invoke Gforth.
12009:
12010: @item program data space available:
12011: @cindex program data space available
12012: @cindex data space available
12013: @code{UNUSED .} gives the remaining dictionary space. The total
12014: dictionary space can be specified with the @code{-m} switch
12015: (@pxref{Invoking Gforth}) when Gforth starts up.
12016:
12017: @item return stack space available:
12018: @cindex return stack space available
12019: You can compute the total return stack space in cells with
12020: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
12021: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
12022:
12023: @item stack space available:
12024: @cindex stack space available
12025: You can compute the total data stack space in cells with
12026: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
12027: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
12028:
12029: @item system dictionary space required, in address units:
12030: @cindex system dictionary space required, in address units
12031: Type @code{here forthstart - .} after startup. At the time of this
12032: writing, this gives 80080 (bytes) on a 32-bit system.
12033: @end table
12034:
12035:
12036: @c =====================================================================
12037: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
12038: @section The optional Block word set
12039: @c =====================================================================
12040: @cindex system documentation, block words
12041: @cindex block words, system documentation
12042:
12043: @menu
12044: * block-idef:: Implementation Defined Options
12045: * block-ambcond:: Ambiguous Conditions
12046: * block-other:: Other System Documentation
12047: @end menu
12048:
12049:
12050: @c ---------------------------------------------------------------------
12051: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
12052: @subsection Implementation Defined Options
12053: @c ---------------------------------------------------------------------
12054: @cindex implementation-defined options, block words
12055: @cindex block words, implementation-defined options
12056:
12057: @table @i
12058: @item the format for display by @code{LIST}:
12059: @cindex @code{LIST} display format
12060: First the screen number is displayed, then 16 lines of 64 characters,
12061: each line preceded by the line number.
12062:
12063: @item the length of a line affected by @code{\}:
12064: @cindex length of a line affected by @code{\}
12065: @cindex @code{\}, line length in blocks
12066: 64 characters.
12067: @end table
12068:
12069:
12070: @c ---------------------------------------------------------------------
12071: @node block-ambcond, block-other, block-idef, The optional Block word set
12072: @subsection Ambiguous conditions
12073: @c ---------------------------------------------------------------------
12074: @cindex block words, ambiguous conditions
12075: @cindex ambiguous conditions, block words
12076:
12077: @table @i
12078: @item correct block read was not possible:
12079: @cindex block read not possible
12080: Typically results in a @code{throw} of some OS-derived value (between
12081: -512 and -2048). If the blocks file was just not long enough, blanks are
12082: supplied for the missing portion.
12083:
12084: @item I/O exception in block transfer:
12085: @cindex I/O exception in block transfer
12086: @cindex block transfer, I/O exception
12087: Typically results in a @code{throw} of some OS-derived value (between
12088: -512 and -2048).
12089:
12090: @item invalid block number:
12091: @cindex invalid block number
12092: @cindex block number invalid
12093: @code{-35 throw} (Invalid block number)
12094:
12095: @item a program directly alters the contents of @code{BLK}:
12096: @cindex @code{BLK}, altering @code{BLK}
12097: The input stream is switched to that other block, at the same
12098: position. If the storing to @code{BLK} happens when interpreting
12099: non-block input, the system will get quite confused when the block ends.
12100:
12101: @item no current block buffer for @code{UPDATE}:
12102: @cindex @code{UPDATE}, no current block buffer
12103: @code{UPDATE} has no effect.
12104:
12105: @end table
12106:
12107: @c ---------------------------------------------------------------------
12108: @node block-other, , block-ambcond, The optional Block word set
12109: @subsection Other system documentation
12110: @c ---------------------------------------------------------------------
12111: @cindex other system documentation, block words
12112: @cindex block words, other system documentation
12113:
12114: @table @i
12115: @item any restrictions a multiprogramming system places on the use of buffer addresses:
12116: No restrictions (yet).
12117:
12118: @item the number of blocks available for source and data:
12119: depends on your disk space.
12120:
12121: @end table
12122:
12123:
12124: @c =====================================================================
12125: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
12126: @section The optional Double Number word set
12127: @c =====================================================================
12128: @cindex system documentation, double words
12129: @cindex double words, system documentation
12130:
12131: @menu
12132: * double-ambcond:: Ambiguous Conditions
12133: @end menu
12134:
12135:
12136: @c ---------------------------------------------------------------------
12137: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
12138: @subsection Ambiguous conditions
12139: @c ---------------------------------------------------------------------
12140: @cindex double words, ambiguous conditions
12141: @cindex ambiguous conditions, double words
12142:
12143: @table @i
12144: @item @i{d} outside of range of @i{n} in @code{D>S}:
12145: @cindex @code{D>S}, @i{d} out of range of @i{n}
12146: The least significant cell of @i{d} is produced.
12147:
12148: @end table
12149:
12150:
12151: @c =====================================================================
12152: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
12153: @section The optional Exception word set
12154: @c =====================================================================
12155: @cindex system documentation, exception words
12156: @cindex exception words, system documentation
12157:
12158: @menu
12159: * exception-idef:: Implementation Defined Options
12160: @end menu
12161:
12162:
12163: @c ---------------------------------------------------------------------
12164: @node exception-idef, , The optional Exception word set, The optional Exception word set
12165: @subsection Implementation Defined Options
12166: @c ---------------------------------------------------------------------
12167: @cindex implementation-defined options, exception words
12168: @cindex exception words, implementation-defined options
12169:
12170: @table @i
12171: @item @code{THROW}-codes used in the system:
12172: @cindex @code{THROW}-codes used in the system
12173: The codes -256@minus{}-511 are used for reporting signals. The mapping
12174: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
12175: codes -512@minus{}-2047 are used for OS errors (for file and memory
12176: allocation operations). The mapping from OS error numbers to throw codes
12177: is -512@minus{}@code{errno}. One side effect of this mapping is that
12178: undefined OS errors produce a message with a strange number; e.g.,
12179: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
12180: @end table
12181:
12182: @c =====================================================================
12183: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
12184: @section The optional Facility word set
12185: @c =====================================================================
12186: @cindex system documentation, facility words
12187: @cindex facility words, system documentation
12188:
12189: @menu
12190: * facility-idef:: Implementation Defined Options
12191: * facility-ambcond:: Ambiguous Conditions
12192: @end menu
12193:
12194:
12195: @c ---------------------------------------------------------------------
12196: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
12197: @subsection Implementation Defined Options
12198: @c ---------------------------------------------------------------------
12199: @cindex implementation-defined options, facility words
12200: @cindex facility words, implementation-defined options
12201:
12202: @table @i
12203: @item encoding of keyboard events (@code{EKEY}):
12204: @cindex keyboard events, encoding in @code{EKEY}
12205: @cindex @code{EKEY}, encoding of keyboard events
12206: Keys corresponding to ASCII characters are encoded as ASCII characters.
12207: Other keys are encoded with the constants @code{k-left}, @code{k-right},
12208: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
12209: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
12210: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
12211:
12212:
12213: @item duration of a system clock tick:
12214: @cindex duration of a system clock tick
12215: @cindex clock tick duration
12216: System dependent. With respect to @code{MS}, the time is specified in
12217: microseconds. How well the OS and the hardware implement this, is
12218: another question.
12219:
12220: @item repeatability to be expected from the execution of @code{MS}:
12221: @cindex repeatability to be expected from the execution of @code{MS}
12222: @cindex @code{MS}, repeatability to be expected
12223: System dependent. On Unix, a lot depends on load. If the system is
12224: lightly loaded, and the delay is short enough that Gforth does not get
12225: swapped out, the performance should be acceptable. Under MS-DOS and
12226: other single-tasking systems, it should be good.
12227:
12228: @end table
12229:
12230:
12231: @c ---------------------------------------------------------------------
12232: @node facility-ambcond, , facility-idef, The optional Facility word set
12233: @subsection Ambiguous conditions
12234: @c ---------------------------------------------------------------------
12235: @cindex facility words, ambiguous conditions
12236: @cindex ambiguous conditions, facility words
12237:
12238: @table @i
12239: @item @code{AT-XY} can't be performed on user output device:
12240: @cindex @code{AT-XY} can't be performed on user output device
12241: Largely terminal dependent. No range checks are done on the arguments.
12242: No errors are reported. You may see some garbage appearing, you may see
12243: simply nothing happen.
12244:
12245: @end table
12246:
12247:
12248: @c =====================================================================
12249: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
12250: @section The optional File-Access word set
12251: @c =====================================================================
12252: @cindex system documentation, file words
12253: @cindex file words, system documentation
12254:
12255: @menu
12256: * file-idef:: Implementation Defined Options
12257: * file-ambcond:: Ambiguous Conditions
12258: @end menu
12259:
12260: @c ---------------------------------------------------------------------
12261: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
12262: @subsection Implementation Defined Options
12263: @c ---------------------------------------------------------------------
12264: @cindex implementation-defined options, file words
12265: @cindex file words, implementation-defined options
12266:
12267: @table @i
12268: @item file access methods used:
12269: @cindex file access methods used
12270: @code{R/O}, @code{R/W} and @code{BIN} work as you would
12271: expect. @code{W/O} translates into the C file opening mode @code{w} (or
12272: @code{wb}): The file is cleared, if it exists, and created, if it does
12273: not (with both @code{open-file} and @code{create-file}). Under Unix
12274: @code{create-file} creates a file with 666 permissions modified by your
12275: umask.
12276:
12277: @item file exceptions:
12278: @cindex file exceptions
12279: The file words do not raise exceptions (except, perhaps, memory access
12280: faults when you pass illegal addresses or file-ids).
12281:
12282: @item file line terminator:
12283: @cindex file line terminator
12284: System-dependent. Gforth uses C's newline character as line
12285: terminator. What the actual character code(s) of this are is
12286: system-dependent.
12287:
12288: @item file name format:
12289: @cindex file name format
12290: System dependent. Gforth just uses the file name format of your OS.
12291:
12292: @item information returned by @code{FILE-STATUS}:
12293: @cindex @code{FILE-STATUS}, returned information
12294: @code{FILE-STATUS} returns the most powerful file access mode allowed
12295: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
12296: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
12297: along with the returned mode.
12298:
12299: @item input file state after an exception when including source:
12300: @cindex exception when including source
12301: All files that are left via the exception are closed.
12302:
12303: @item @i{ior} values and meaning:
12304: @cindex @i{ior} values and meaning
12305: The @i{ior}s returned by the file and memory allocation words are
12306: intended as throw codes. They typically are in the range
12307: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
12308: @i{ior}s is -512@minus{}@i{errno}.
12309:
12310: @item maximum depth of file input nesting:
12311: @cindex maximum depth of file input nesting
12312: @cindex file input nesting, maximum depth
12313: limited by the amount of return stack, locals/TIB stack, and the number
12314: of open files available. This should not give you troubles.
12315:
12316: @item maximum size of input line:
12317: @cindex maximum size of input line
12318: @cindex input line size, maximum
12319: @code{/line}. Currently 255.
12320:
12321: @item methods of mapping block ranges to files:
12322: @cindex mapping block ranges to files
12323: @cindex files containing blocks
12324: @cindex blocks in files
12325: By default, blocks are accessed in the file @file{blocks.fb} in the
12326: current working directory. The file can be switched with @code{USE}.
12327:
12328: @item number of string buffers provided by @code{S"}:
12329: @cindex @code{S"}, number of string buffers
12330: 1
12331:
12332: @item size of string buffer used by @code{S"}:
12333: @cindex @code{S"}, size of string buffer
12334: @code{/line}. currently 255.
12335:
12336: @end table
12337:
12338: @c ---------------------------------------------------------------------
12339: @node file-ambcond, , file-idef, The optional File-Access word set
12340: @subsection Ambiguous conditions
12341: @c ---------------------------------------------------------------------
12342: @cindex file words, ambiguous conditions
12343: @cindex ambiguous conditions, file words
12344:
12345: @table @i
12346: @item attempting to position a file outside its boundaries:
12347: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
12348: @code{REPOSITION-FILE} is performed as usual: Afterwards,
12349: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
12350:
12351: @item attempting to read from file positions not yet written:
12352: @cindex reading from file positions not yet written
12353: End-of-file, i.e., zero characters are read and no error is reported.
12354:
12355: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
12356: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
12357: An appropriate exception may be thrown, but a memory fault or other
12358: problem is more probable.
12359:
12360: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
12361: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
12362: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
12363: The @i{ior} produced by the operation, that discovered the problem, is
12364: thrown.
12365:
12366: @item named file cannot be opened (@code{INCLUDED}):
12367: @cindex @code{INCLUDED}, named file cannot be opened
12368: The @i{ior} produced by @code{open-file} is thrown.
12369:
12370: @item requesting an unmapped block number:
12371: @cindex unmapped block numbers
12372: There are no unmapped legal block numbers. On some operating systems,
12373: writing a block with a large number may overflow the file system and
12374: have an error message as consequence.
12375:
12376: @item using @code{source-id} when @code{blk} is non-zero:
12377: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
12378: @code{source-id} performs its function. Typically it will give the id of
12379: the source which loaded the block. (Better ideas?)
12380:
12381: @end table
12382:
12383:
12384: @c =====================================================================
12385: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
12386: @section The optional Floating-Point word set
12387: @c =====================================================================
12388: @cindex system documentation, floating-point words
12389: @cindex floating-point words, system documentation
12390:
12391: @menu
12392: * floating-idef:: Implementation Defined Options
12393: * floating-ambcond:: Ambiguous Conditions
12394: @end menu
12395:
12396:
12397: @c ---------------------------------------------------------------------
12398: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
12399: @subsection Implementation Defined Options
12400: @c ---------------------------------------------------------------------
12401: @cindex implementation-defined options, floating-point words
12402: @cindex floating-point words, implementation-defined options
12403:
12404: @table @i
12405: @item format and range of floating point numbers:
12406: @cindex format and range of floating point numbers
12407: @cindex floating point numbers, format and range
12408: System-dependent; the @code{double} type of C.
12409:
12410: @item results of @code{REPRESENT} when @i{float} is out of range:
12411: @cindex @code{REPRESENT}, results when @i{float} is out of range
12412: System dependent; @code{REPRESENT} is implemented using the C library
12413: function @code{ecvt()} and inherits its behaviour in this respect.
12414:
12415: @item rounding or truncation of floating-point numbers:
12416: @cindex rounding of floating-point numbers
12417: @cindex truncation of floating-point numbers
12418: @cindex floating-point numbers, rounding or truncation
12419: System dependent; the rounding behaviour is inherited from the hosting C
12420: compiler. IEEE-FP-based (i.e., most) systems by default round to
12421: nearest, and break ties by rounding to even (i.e., such that the last
12422: bit of the mantissa is 0).
12423:
12424: @item size of floating-point stack:
12425: @cindex floating-point stack size
12426: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
12427: the floating-point stack (in floats). You can specify this on startup
12428: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
12429:
12430: @item width of floating-point stack:
12431: @cindex floating-point stack width
12432: @code{1 floats}.
12433:
12434: @end table
12435:
12436:
12437: @c ---------------------------------------------------------------------
12438: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
12439: @subsection Ambiguous conditions
12440: @c ---------------------------------------------------------------------
12441: @cindex floating-point words, ambiguous conditions
12442: @cindex ambiguous conditions, floating-point words
12443:
12444: @table @i
12445: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
12446: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
12447: System-dependent. Typically results in a @code{-23 THROW} like other
12448: alignment violations.
12449:
12450: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
12451: @cindex @code{f@@} used with an address that is not float aligned
12452: @cindex @code{f!} used with an address that is not float aligned
12453: System-dependent. Typically results in a @code{-23 THROW} like other
12454: alignment violations.
12455:
12456: @item floating-point result out of range:
12457: @cindex floating-point result out of range
12458: System-dependent. Can result in a @code{-55 THROW} (Floating-point
12459: unidentified fault), or can produce a special value representing, e.g.,
12460: Infinity.
12461:
12462: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
12463: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
12464: System-dependent. Typically results in an alignment fault like other
12465: alignment violations.
12466:
12467: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
12468: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
12469: The floating-point number is converted into decimal nonetheless.
12470:
12471: @item Both arguments are equal to zero (@code{FATAN2}):
12472: @cindex @code{FATAN2}, both arguments are equal to zero
12473: System-dependent. @code{FATAN2} is implemented using the C library
12474: function @code{atan2()}.
12475:
12476: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
12477: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
12478: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
12479: because of small errors and the tan will be a very large (or very small)
12480: but finite number.
12481:
12482: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
12483: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
12484: The result is rounded to the nearest float.
12485:
12486: @item dividing by zero:
12487: @cindex dividing by zero, floating-point
12488: @cindex floating-point dividing by zero
12489: @cindex floating-point unidentified fault, FP divide-by-zero
12490: @code{-55 throw} (Floating-point unidentified fault)
12491:
12492: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
12493: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
12494: System dependent. On IEEE-FP based systems the number is converted into
12495: an infinity.
12496:
12497: @item @i{float}<1 (@code{FACOSH}):
12498: @cindex @code{FACOSH}, @i{float}<1
12499: @cindex floating-point unidentified fault, @code{FACOSH}
12500: @code{-55 throw} (Floating-point unidentified fault)
12501:
12502: @item @i{float}=<-1 (@code{FLNP1}):
12503: @cindex @code{FLNP1}, @i{float}=<-1
12504: @cindex floating-point unidentified fault, @code{FLNP1}
12505: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
12506: negative infinity is typically produced for @i{float}=-1.
12507:
12508: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
12509: @cindex @code{FLN}, @i{float}=<0
12510: @cindex @code{FLOG}, @i{float}=<0
12511: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
12512: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
12513: negative infinity is typically produced for @i{float}=0.
12514:
12515: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
12516: @cindex @code{FASINH}, @i{float}<0
12517: @cindex @code{FSQRT}, @i{float}<0
12518: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
12519: @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
12520: produces values for these inputs on my Linux box (Bug in the C library?)
12521:
12522: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
12523: @cindex @code{FACOS}, |@i{float}|>1
12524: @cindex @code{FASIN}, |@i{float}|>1
12525: @cindex @code{FATANH}, |@i{float}|>1
12526: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
12527: @code{-55 throw} (Floating-point unidentified fault).
12528:
12529: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
12530: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
12531: @cindex floating-point unidentified fault, @code{F>D}
12532: @code{-55 throw} (Floating-point unidentified fault).
12533:
12534: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
12535: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
12536: This does not happen.
12537: @end table
12538:
12539: @c =====================================================================
12540: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
12541: @section The optional Locals word set
12542: @c =====================================================================
12543: @cindex system documentation, locals words
12544: @cindex locals words, system documentation
12545:
12546: @menu
12547: * locals-idef:: Implementation Defined Options
12548: * locals-ambcond:: Ambiguous Conditions
12549: @end menu
12550:
12551:
12552: @c ---------------------------------------------------------------------
12553: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
12554: @subsection Implementation Defined Options
12555: @c ---------------------------------------------------------------------
12556: @cindex implementation-defined options, locals words
12557: @cindex locals words, implementation-defined options
12558:
12559: @table @i
12560: @item maximum number of locals in a definition:
12561: @cindex maximum number of locals in a definition
12562: @cindex locals, maximum number in a definition
12563: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
12564: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
12565: characters. The number of locals in a definition is bounded by the size
12566: of locals-buffer, which contains the names of the locals.
12567:
12568: @end table
12569:
12570:
12571: @c ---------------------------------------------------------------------
12572: @node locals-ambcond, , locals-idef, The optional Locals word set
12573: @subsection Ambiguous conditions
12574: @c ---------------------------------------------------------------------
12575: @cindex locals words, ambiguous conditions
12576: @cindex ambiguous conditions, locals words
12577:
12578: @table @i
12579: @item executing a named local in interpretation state:
12580: @cindex local in interpretation state
12581: @cindex Interpreting a compile-only word, for a local
12582: Locals have no interpretation semantics. If you try to perform the
12583: interpretation semantics, you will get a @code{-14 throw} somewhere
12584: (Interpreting a compile-only word). If you perform the compilation
12585: semantics, the locals access will be compiled (irrespective of state).
12586:
12587: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
12588: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
12589: @cindex @code{TO} on non-@code{VALUE}s and non-locals
12590: @cindex Invalid name argument, @code{TO}
12591: @code{-32 throw} (Invalid name argument)
12592:
12593: @end table
12594:
12595:
12596: @c =====================================================================
12597: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
12598: @section The optional Memory-Allocation word set
12599: @c =====================================================================
12600: @cindex system documentation, memory-allocation words
12601: @cindex memory-allocation words, system documentation
12602:
12603: @menu
12604: * memory-idef:: Implementation Defined Options
12605: @end menu
12606:
12607:
12608: @c ---------------------------------------------------------------------
12609: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
12610: @subsection Implementation Defined Options
12611: @c ---------------------------------------------------------------------
12612: @cindex implementation-defined options, memory-allocation words
12613: @cindex memory-allocation words, implementation-defined options
12614:
12615: @table @i
12616: @item values and meaning of @i{ior}:
12617: @cindex @i{ior} values and meaning
12618: The @i{ior}s returned by the file and memory allocation words are
12619: intended as throw codes. They typically are in the range
12620: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
12621: @i{ior}s is -512@minus{}@i{errno}.
12622:
12623: @end table
12624:
12625: @c =====================================================================
12626: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
12627: @section The optional Programming-Tools word set
12628: @c =====================================================================
12629: @cindex system documentation, programming-tools words
12630: @cindex programming-tools words, system documentation
12631:
12632: @menu
12633: * programming-idef:: Implementation Defined Options
12634: * programming-ambcond:: Ambiguous Conditions
12635: @end menu
12636:
12637:
12638: @c ---------------------------------------------------------------------
12639: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
12640: @subsection Implementation Defined Options
12641: @c ---------------------------------------------------------------------
12642: @cindex implementation-defined options, programming-tools words
12643: @cindex programming-tools words, implementation-defined options
12644:
12645: @table @i
12646: @item ending sequence for input following @code{;CODE} and @code{CODE}:
12647: @cindex @code{;CODE} ending sequence
12648: @cindex @code{CODE} ending sequence
12649: @code{END-CODE}
12650:
12651: @item manner of processing input following @code{;CODE} and @code{CODE}:
12652: @cindex @code{;CODE}, processing input
12653: @cindex @code{CODE}, processing input
12654: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
12655: the input is processed by the text interpreter, (starting) in interpret
12656: state.
12657:
12658: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
12659: @cindex @code{ASSEMBLER}, search order capability
12660: The ANS Forth search order word set.
12661:
12662: @item source and format of display by @code{SEE}:
12663: @cindex @code{SEE}, source and format of output
12664: The source for @code{see} is the intermediate code used by the inner
12665: interpreter. The current @code{see} tries to output Forth source code
12666: as well as possible.
12667:
12668: @end table
12669:
12670: @c ---------------------------------------------------------------------
12671: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
12672: @subsection Ambiguous conditions
12673: @c ---------------------------------------------------------------------
12674: @cindex programming-tools words, ambiguous conditions
12675: @cindex ambiguous conditions, programming-tools words
12676:
12677: @table @i
12678:
12679: @item deleting the compilation word list (@code{FORGET}):
12680: @cindex @code{FORGET}, deleting the compilation word list
12681: Not implemented (yet).
12682:
12683: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
12684: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
12685: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
12686: @cindex control-flow stack underflow
12687: This typically results in an @code{abort"} with a descriptive error
12688: message (may change into a @code{-22 throw} (Control structure mismatch)
12689: in the future). You may also get a memory access error. If you are
12690: unlucky, this ambiguous condition is not caught.
12691:
12692: @item @i{name} can't be found (@code{FORGET}):
12693: @cindex @code{FORGET}, @i{name} can't be found
12694: Not implemented (yet).
12695:
12696: @item @i{name} not defined via @code{CREATE}:
12697: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
12698: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
12699: the execution semantics of the last defined word no matter how it was
12700: defined.
12701:
12702: @item @code{POSTPONE} applied to @code{[IF]}:
12703: @cindex @code{POSTPONE} applied to @code{[IF]}
12704: @cindex @code{[IF]} and @code{POSTPONE}
12705: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
12706: equivalent to @code{[IF]}.
12707:
12708: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
12709: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
12710: Continue in the same state of conditional compilation in the next outer
12711: input source. Currently there is no warning to the user about this.
12712:
12713: @item removing a needed definition (@code{FORGET}):
12714: @cindex @code{FORGET}, removing a needed definition
12715: Not implemented (yet).
12716:
12717: @end table
12718:
12719:
12720: @c =====================================================================
12721: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
12722: @section The optional Search-Order word set
12723: @c =====================================================================
12724: @cindex system documentation, search-order words
12725: @cindex search-order words, system documentation
12726:
12727: @menu
12728: * search-idef:: Implementation Defined Options
12729: * search-ambcond:: Ambiguous Conditions
12730: @end menu
12731:
12732:
12733: @c ---------------------------------------------------------------------
12734: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
12735: @subsection Implementation Defined Options
12736: @c ---------------------------------------------------------------------
12737: @cindex implementation-defined options, search-order words
12738: @cindex search-order words, implementation-defined options
12739:
12740: @table @i
12741: @item maximum number of word lists in search order:
12742: @cindex maximum number of word lists in search order
12743: @cindex search order, maximum depth
12744: @code{s" wordlists" environment? drop .}. Currently 16.
12745:
12746: @item minimum search order:
12747: @cindex minimum search order
12748: @cindex search order, minimum
12749: @code{root root}.
12750:
12751: @end table
12752:
12753: @c ---------------------------------------------------------------------
12754: @node search-ambcond, , search-idef, The optional Search-Order word set
12755: @subsection Ambiguous conditions
12756: @c ---------------------------------------------------------------------
12757: @cindex search-order words, ambiguous conditions
12758: @cindex ambiguous conditions, search-order words
12759:
12760: @table @i
12761: @item changing the compilation word list (during compilation):
12762: @cindex changing the compilation word list (during compilation)
12763: @cindex compilation word list, change before definition ends
12764: The word is entered into the word list that was the compilation word list
12765: at the start of the definition. Any changes to the name field (e.g.,
12766: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
12767: are applied to the latest defined word (as reported by @code{last} or
12768: @code{lastxt}), if possible, irrespective of the compilation word list.
12769:
12770: @item search order empty (@code{previous}):
12771: @cindex @code{previous}, search order empty
12772: @cindex vocstack empty, @code{previous}
12773: @code{abort" Vocstack empty"}.
12774:
12775: @item too many word lists in search order (@code{also}):
12776: @cindex @code{also}, too many word lists in search order
12777: @cindex vocstack full, @code{also}
12778: @code{abort" Vocstack full"}.
12779:
12780: @end table
12781:
12782: @c ***************************************************************
12783: @node Model, Integrating Gforth, ANS conformance, Top
12784: @chapter Model
12785:
12786: This chapter has yet to be written. It will contain information, on
12787: which internal structures you can rely.
12788:
12789: @c ***************************************************************
12790: @node Integrating Gforth, Emacs and Gforth, Model, Top
12791: @chapter Integrating Gforth into C programs
12792:
12793: This is not yet implemented.
12794:
12795: Several people like to use Forth as scripting language for applications
12796: that are otherwise written in C, C++, or some other language.
12797:
12798: The Forth system ATLAST provides facilities for embedding it into
12799: applications; unfortunately it has several disadvantages: most
12800: importantly, it is not based on ANS Forth, and it is apparently dead
12801: (i.e., not developed further and not supported). The facilities
12802: provided by Gforth in this area are inspired by ATLAST's facilities, so
12803: making the switch should not be hard.
12804:
12805: We also tried to design the interface such that it can easily be
12806: implemented by other Forth systems, so that we may one day arrive at a
12807: standardized interface. Such a standard interface would allow you to
12808: replace the Forth system without having to rewrite C code.
12809:
12810: You embed the Gforth interpreter by linking with the library
12811: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
12812: global symbols in this library that belong to the interface, have the
12813: prefix @code{forth_}. (Global symbols that are used internally have the
12814: prefix @code{gforth_}).
12815:
12816: You can include the declarations of Forth types and the functions and
12817: variables of the interface with @code{#include <forth.h>}.
12818:
12819: Types.
12820:
12821: Variables.
12822:
12823: Data and FP Stack pointer. Area sizes.
12824:
12825: functions.
12826:
12827: forth_init(imagefile)
12828: forth_evaluate(string) exceptions?
12829: forth_goto(address) (or forth_execute(xt)?)
12830: forth_continue() (a corountining mechanism)
12831:
12832: Adding primitives.
12833:
12834: No checking.
12835:
12836: Signals?
12837:
12838: Accessing the Stacks
12839:
12840: @c ******************************************************************
12841: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
12842: @chapter Emacs and Gforth
12843: @cindex Emacs and Gforth
12844:
12845: @cindex @file{gforth.el}
12846: @cindex @file{forth.el}
12847: @cindex Rydqvist, Goran
12848: @cindex comment editing commands
12849: @cindex @code{\}, editing with Emacs
12850: @cindex debug tracer editing commands
12851: @cindex @code{~~}, removal with Emacs
12852: @cindex Forth mode in Emacs
12853: Gforth comes with @file{gforth.el}, an improved version of
12854: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
12855: improvements are:
12856:
12857: @itemize @bullet
12858: @item
12859: A better (but still not perfect) handling of indentation.
12860: @item
12861: Comment paragraph filling (@kbd{M-q})
12862: @item
12863: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
12864: @item
12865: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
12866: @item
12867: Support of the @code{info-lookup} feature for looking up the
12868: documentation of a word.
12869: @end itemize
12870:
12871: I left the stuff I do not use alone, even though some of it only makes
12872: sense for TILE. To get a description of these features, enter Forth mode
12873: and type @kbd{C-h m}.
12874:
12875: @cindex source location of error or debugging output in Emacs
12876: @cindex error output, finding the source location in Emacs
12877: @cindex debugging output, finding the source location in Emacs
12878: In addition, Gforth supports Emacs quite well: The source code locations
12879: given in error messages, debugging output (from @code{~~}) and failed
12880: assertion messages are in the right format for Emacs' compilation mode
12881: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
12882: Manual}) so the source location corresponding to an error or other
12883: message is only a few keystrokes away (@kbd{C-x `} for the next error,
12884: @kbd{C-c C-c} for the error under the cursor).
12885:
12886: @cindex @file{TAGS} file
12887: @cindex @file{etags.fs}
12888: @cindex viewing the source of a word in Emacs
12889: @cindex @code{require}, placement in files
12890: @cindex @code{include}, placement in files
12891: Also, if you @code{require} @file{etags.fs}, a new @file{TAGS} file will
12892: be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
12893: contains the definitions of all words defined afterwards. You can then
12894: find the source for a word using @kbd{M-.}. Note that emacs can use
12895: several tags files at the same time (e.g., one for the Gforth sources
12896: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
12897: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
12898: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
12899: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
12900: with @file{etags.fs}, you should avoid putting definitions both before
12901: and after @code{require} etc., otherwise you will see the same file
12902: visited several times by commands like @code{tags-search}.
12903:
12904: @cindex viewing the documentation of a word in Emacs
12905: @cindex context-sensitive help
12906: Moreover, for words documented in this manual, you can look up the
12907: glossary entry quickly by using @kbd{C-h TAB}
12908: (@code{info-lookup-symbol}, see @pxref{Documentation, ,Documentation
12909: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
12910: later and does not work for words containing @code{:}.
12911:
12912:
12913: @cindex @file{.emacs}
12914: To get all these benefits, add the following lines to your @file{.emacs}
12915: file:
12916:
12917: @example
12918: (autoload 'forth-mode "gforth.el")
12919: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
12920: @end example
12921:
12922: @c ******************************************************************
12923: @node Image Files, Engine, Emacs and Gforth, Top
12924: @chapter Image Files
12925: @cindex image file
12926: @cindex @file{.fi} files
12927: @cindex precompiled Forth code
12928: @cindex dictionary in persistent form
12929: @cindex persistent form of dictionary
12930:
12931: An image file is a file containing an image of the Forth dictionary,
12932: i.e., compiled Forth code and data residing in the dictionary. By
12933: convention, we use the extension @code{.fi} for image files.
12934:
12935: @menu
12936: * Image Licensing Issues:: Distribution terms for images.
12937: * Image File Background:: Why have image files?
12938: * Non-Relocatable Image Files:: don't always work.
12939: * Data-Relocatable Image Files:: are better.
12940: * Fully Relocatable Image Files:: better yet.
12941: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
12942: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
12943: * Modifying the Startup Sequence:: and turnkey applications.
12944: @end menu
12945:
12946: @node Image Licensing Issues, Image File Background, Image Files, Image Files
12947: @section Image Licensing Issues
12948: @cindex license for images
12949: @cindex image license
12950:
12951: An image created with @code{gforthmi} (@pxref{gforthmi}) or
12952: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
12953: original image; i.e., according to copyright law it is a derived work of
12954: the original image.
12955:
12956: Since Gforth is distributed under the GNU GPL, the newly created image
12957: falls under the GNU GPL, too. In particular, this means that if you
12958: distribute the image, you have to make all of the sources for the image
12959: available, including those you wrote. For details see @ref{License, ,
12960: GNU General Public License (Section 3)}.
12961:
12962: If you create an image with @code{cross} (@pxref{cross.fs}), the image
12963: contains only code compiled from the sources you gave it; if none of
12964: these sources is under the GPL, the terms discussed above do not apply
12965: to the image. However, if your image needs an engine (a gforth binary)
12966: that is under the GPL, you should make sure that you distribute both in
12967: a way that is at most a @emph{mere aggregation}, if you don't want the
12968: terms of the GPL to apply to the image.
12969:
12970: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
12971: @section Image File Background
12972: @cindex image file background
12973:
12974: Our Forth system consists not only of primitives, but also of
12975: definitions written in Forth. Since the Forth compiler itself belongs to
12976: those definitions, it is not possible to start the system with the
12977: primitives and the Forth source alone. Therefore we provide the Forth
12978: code as an image file in nearly executable form. When Gforth starts up,
12979: a C routine loads the image file into memory, optionally relocates the
12980: addresses, then sets up the memory (stacks etc.) according to
12981: information in the image file, and (finally) starts executing Forth
12982: code.
12983:
12984: The image file variants represent different compromises between the
12985: goals of making it easy to generate image files and making them
12986: portable.
12987:
12988: @cindex relocation at run-time
12989: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
12990: run-time. This avoids many of the complications discussed below (image
12991: files are data relocatable without further ado), but costs performance
12992: (one addition per memory access).
12993:
12994: @cindex relocation at load-time
12995: By contrast, the Gforth loader performs relocation at image load time. The
12996: loader also has to replace tokens that represent primitive calls with the
12997: appropriate code-field addresses (or code addresses in the case of
12998: direct threading).
12999:
13000: There are three kinds of image files, with different degrees of
13001: relocatability: non-relocatable, data-relocatable, and fully relocatable
13002: image files.
13003:
13004: @cindex image file loader
13005: @cindex relocating loader
13006: @cindex loader for image files
13007: These image file variants have several restrictions in common; they are
13008: caused by the design of the image file loader:
13009:
13010: @itemize @bullet
13011: @item
13012: There is only one segment; in particular, this means, that an image file
13013: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
13014: them). The contents of the stacks are not represented, either.
13015:
13016: @item
13017: The only kinds of relocation supported are: adding the same offset to
13018: all cells that represent data addresses; and replacing special tokens
13019: with code addresses or with pieces of machine code.
13020:
13021: If any complex computations involving addresses are performed, the
13022: results cannot be represented in the image file. Several applications that
13023: use such computations come to mind:
13024: @itemize @minus
13025: @item
13026: Hashing addresses (or data structures which contain addresses) for table
13027: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
13028: purpose, you will have no problem, because the hash tables are
13029: recomputed automatically when the system is started. If you use your own
13030: hash tables, you will have to do something similar.
13031:
13032: @item
13033: There's a cute implementation of doubly-linked lists that uses
13034: @code{XOR}ed addresses. You could represent such lists as singly-linked
13035: in the image file, and restore the doubly-linked representation on
13036: startup.@footnote{In my opinion, though, you should think thrice before
13037: using a doubly-linked list (whatever implementation).}
13038:
13039: @item
13040: The code addresses of run-time routines like @code{docol:} cannot be
13041: represented in the image file (because their tokens would be replaced by
13042: machine code in direct threaded implementations). As a workaround,
13043: compute these addresses at run-time with @code{>code-address} from the
13044: executions tokens of appropriate words (see the definitions of
13045: @code{docol:} and friends in @file{kernel.fs}).
13046:
13047: @item
13048: On many architectures addresses are represented in machine code in some
13049: shifted or mangled form. You cannot put @code{CODE} words that contain
13050: absolute addresses in this form in a relocatable image file. Workarounds
13051: are representing the address in some relative form (e.g., relative to
13052: the CFA, which is present in some register), or loading the address from
13053: a place where it is stored in a non-mangled form.
13054: @end itemize
13055: @end itemize
13056:
13057: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
13058: @section Non-Relocatable Image Files
13059: @cindex non-relocatable image files
13060: @cindex image file, non-relocatable
13061:
13062: These files are simple memory dumps of the dictionary. They are specific
13063: to the executable (i.e., @file{gforth} file) they were created
13064: with. What's worse, they are specific to the place on which the
13065: dictionary resided when the image was created. Now, there is no
13066: guarantee that the dictionary will reside at the same place the next
13067: time you start Gforth, so there's no guarantee that a non-relocatable
13068: image will work the next time (Gforth will complain instead of crashing,
13069: though).
13070:
13071: You can create a non-relocatable image file with
13072:
13073:
13074: doc-savesystem
13075:
13076:
13077: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
13078: @section Data-Relocatable Image Files
13079: @cindex data-relocatable image files
13080: @cindex image file, data-relocatable
13081:
13082: These files contain relocatable data addresses, but fixed code addresses
13083: (instead of tokens). They are specific to the executable (i.e.,
13084: @file{gforth} file) they were created with. For direct threading on some
13085: architectures (e.g., the i386), data-relocatable images do not work. You
13086: get a data-relocatable image, if you use @file{gforthmi} with a
13087: Gforth binary that is not doubly indirect threaded (@pxref{Fully
13088: Relocatable Image Files}).
13089:
13090: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
13091: @section Fully Relocatable Image Files
13092: @cindex fully relocatable image files
13093: @cindex image file, fully relocatable
13094:
13095: @cindex @file{kern*.fi}, relocatability
13096: @cindex @file{gforth.fi}, relocatability
13097: These image files have relocatable data addresses, and tokens for code
13098: addresses. They can be used with different binaries (e.g., with and
13099: without debugging) on the same machine, and even across machines with
13100: the same data formats (byte order, cell size, floating point
13101: format). However, they are usually specific to the version of Gforth
13102: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
13103: are fully relocatable.
13104:
13105: There are two ways to create a fully relocatable image file:
13106:
13107: @menu
13108: * gforthmi:: The normal way
13109: * cross.fs:: The hard way
13110: @end menu
13111:
13112: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
13113: @subsection @file{gforthmi}
13114: @cindex @file{comp-i.fs}
13115: @cindex @file{gforthmi}
13116:
13117: You will usually use @file{gforthmi}. If you want to create an
13118: image @i{file} that contains everything you would load by invoking
13119: Gforth with @code{gforth @i{options}}, you simply say:
13120: @example
13121: gforthmi @i{file} @i{options}
13122: @end example
13123:
13124: E.g., if you want to create an image @file{asm.fi} that has the file
13125: @file{asm.fs} loaded in addition to the usual stuff, you could do it
13126: like this:
13127:
13128: @example
13129: gforthmi asm.fi asm.fs
13130: @end example
13131:
13132: @file{gforthmi} is implemented as a sh script and works like this: It
13133: produces two non-relocatable images for different addresses and then
13134: compares them. Its output reflects this: first you see the output (if
13135: any) of the two Gforth invocations that produce the nonrelocatable image
13136: files, then you see the output of the comparing program: It displays the
13137: offset used for data addresses and the offset used for code addresses;
13138: moreover, for each cell that cannot be represented correctly in the
13139: image files, it displays a line like this:
13140:
13141: @example
13142: 78DC BFFFFA50 BFFFFA40
13143: @end example
13144:
13145: This means that at offset $78dc from @code{forthstart}, one input image
13146: contains $bffffa50, and the other contains $bffffa40. Since these cells
13147: cannot be represented correctly in the output image, you should examine
13148: these places in the dictionary and verify that these cells are dead
13149: (i.e., not read before they are written).
13150:
13151: @cindex --application, @code{gforthmi} option
13152: If you insert the option @code{--application} in front of the image file
13153: name, you will get an image that uses the @code{--appl-image} option
13154: instead of the @code{--image-file} option (@pxref{Invoking
13155: Gforth}). When you execute such an image on Unix (by typing the image
13156: name as command), the Gforth engine will pass all options to the image
13157: instead of trying to interpret them as engine options.
13158:
13159: If you type @file{gforthmi} with no arguments, it prints some usage
13160: instructions.
13161:
13162: @cindex @code{savesystem} during @file{gforthmi}
13163: @cindex @code{bye} during @file{gforthmi}
13164: @cindex doubly indirect threaded code
13165: @cindex environment variables
13166: @cindex @code{GFORTHD} -- environment variable
13167: @cindex @code{GFORTH} -- environment variable
13168: @cindex @code{gforth-ditc}
13169: There are a few wrinkles: After processing the passed @i{options}, the
13170: words @code{savesystem} and @code{bye} must be visible. A special doubly
13171: indirect threaded version of the @file{gforth} executable is used for
13172: creating the nonrelocatable images; you can pass the exact filename of
13173: this executable through the environment variable @code{GFORTHD}
13174: (default: @file{gforth-ditc}); if you pass a version that is not doubly
13175: indirect threaded, you will not get a fully relocatable image, but a
13176: data-relocatable image (because there is no code address offset). The
13177: normal @file{gforth} executable is used for creating the relocatable
13178: image; you can pass the exact filename of this executable through the
13179: environment variable @code{GFORTH}.
13180:
13181: @node cross.fs, , gforthmi, Fully Relocatable Image Files
13182: @subsection @file{cross.fs}
13183: @cindex @file{cross.fs}
13184: @cindex cross-compiler
13185: @cindex metacompiler
13186: @cindex target compiler
13187:
13188: You can also use @code{cross}, a batch compiler that accepts a Forth-like
13189: programming language (@pxref{Cross Compiler}).
13190:
13191: @code{cross} allows you to create image files for machines with
13192: different data sizes and data formats than the one used for generating
13193: the image file. You can also use it to create an application image that
13194: does not contain a Forth compiler. These features are bought with
13195: restrictions and inconveniences in programming. E.g., addresses have to
13196: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
13197: order to make the code relocatable.
13198:
13199:
13200: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
13201: @section Stack and Dictionary Sizes
13202: @cindex image file, stack and dictionary sizes
13203: @cindex dictionary size default
13204: @cindex stack size default
13205:
13206: If you invoke Gforth with a command line flag for the size
13207: (@pxref{Invoking Gforth}), the size you specify is stored in the
13208: dictionary. If you save the dictionary with @code{savesystem} or create
13209: an image with @file{gforthmi}, this size will become the default
13210: for the resulting image file. E.g., the following will create a
13211: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
13212:
13213: @example
13214: gforthmi gforth.fi -m 1M
13215: @end example
13216:
13217: In other words, if you want to set the default size for the dictionary
13218: and the stacks of an image, just invoke @file{gforthmi} with the
13219: appropriate options when creating the image.
13220:
13221: @cindex stack size, cache-friendly
13222: Note: For cache-friendly behaviour (i.e., good performance), you should
13223: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
13224: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
13225: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
13226:
13227: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
13228: @section Running Image Files
13229: @cindex running image files
13230: @cindex invoking image files
13231: @cindex image file invocation
13232:
13233: @cindex -i, invoke image file
13234: @cindex --image file, invoke image file
13235: You can invoke Gforth with an image file @i{image} instead of the
13236: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
13237: @example
13238: gforth -i @i{image}
13239: @end example
13240:
13241: @cindex executable image file
13242: @cindex image file, executable
13243: If your operating system supports starting scripts with a line of the
13244: form @code{#! ...}, you just have to type the image file name to start
13245: Gforth with this image file (note that the file extension @code{.fi} is
13246: just a convention). I.e., to run Gforth with the image file @i{image},
13247: you can just type @i{image} instead of @code{gforth -i @i{image}}.
13248: This works because every @code{.fi} file starts with a line of this
13249: format:
13250:
13251: @example
13252: #! /usr/local/bin/gforth-0.4.0 -i
13253: @end example
13254:
13255: The file and pathname for the Gforth engine specified on this line is
13256: the specific Gforth executable that it was built against; i.e. the value
13257: of the environment variable @code{GFORTH} at the time that
13258: @file{gforthmi} was executed.
13259:
13260: You can make use of the same shell capability to make a Forth source
13261: file into an executable. For example, if you place this text in a file:
13262:
13263: @example
13264: #! /usr/local/bin/gforth
13265:
13266: ." Hello, world" CR
13267: bye
13268: @end example
13269:
13270: @noindent
13271: and then make the file executable (chmod +x in Unix), you can run it
13272: directly from the command line. The sequence @code{#!} is used in two
13273: ways; firstly, it is recognised as a ``magic sequence'' by the operating
13274: system@footnote{The Unix kernel actually recognises two types of files:
13275: executable files and files of data, where the data is processed by an
13276: interpreter that is specified on the ``interpreter line'' -- the first
13277: line of the file, starting with the sequence #!. There may be a small
13278: limit (e.g., 32) on the number of characters that may be specified on
13279: the interpreter line.} secondly it is treated as a comment character by
13280: Gforth. Because of the second usage, a space is required between
13281: @code{#!} and the path to the executable.
13282:
13283: The disadvantage of this latter technique, compared with using
13284: @file{gforthmi}, is that it is slower; the Forth source code is compiled
13285: on-the-fly, each time the program is invoked.
13286:
13287:
13288: doc-#!
13289:
13290:
13291: @node Modifying the Startup Sequence, , Running Image Files, Image Files
13292: @section Modifying the Startup Sequence
13293: @cindex startup sequence for image file
13294: @cindex image file initialization sequence
13295: @cindex initialization sequence of image file
13296:
13297: You can add your own initialization to the startup sequence through the
13298: deferred word @code{'cold}. @code{'cold} is invoked just before the
13299: image-specific command line processing (by default, loading files and
13300: evaluating (@code{-e}) strings) starts.
13301:
13302: A sequence for adding your initialization usually looks like this:
13303:
13304: @example
13305: :noname
13306: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
13307: ... \ your stuff
13308: ; IS 'cold
13309: @end example
13310:
13311: @cindex turnkey image files
13312: @cindex image file, turnkey applications
13313: You can make a turnkey image by letting @code{'cold} execute a word
13314: (your turnkey application) that never returns; instead, it exits Gforth
13315: via @code{bye} or @code{throw}.
13316:
13317: @cindex command-line arguments, access
13318: @cindex arguments on the command line, access
13319: You can access the (image-specific) command-line arguments through the
13320: variables @code{argc} and @code{argv}. @code{arg} provides convenient
13321: access to @code{argv}.
13322:
13323: If @code{'cold} exits normally, Gforth processes the command-line
13324: arguments as files to be loaded and strings to be evaluated. Therefore,
13325: @code{'cold} should remove the arguments it has used in this case.
13326:
13327:
13328:
13329: doc-'cold
13330: doc-argc
13331: doc-argv
13332: doc-arg
13333:
13334:
13335:
13336: @c ******************************************************************
13337: @node Engine, Binding to System Library, Image Files, Top
13338: @chapter Engine
13339: @cindex engine
13340: @cindex virtual machine
13341:
13342: Reading this chapter is not necessary for programming with Gforth. It
13343: may be helpful for finding your way in the Gforth sources.
13344:
13345: The ideas in this section have also been published in the papers
13346: @cite{ANS fig/GNU/??? Forth} (in German) by Bernd Paysan, presented at
13347: the Forth-Tagung '93 and @cite{A Portable Forth Engine} by M. Anton
13348: Ertl, presented at EuroForth '93; the latter is available at
13349: @*@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z}.
13350:
13351: @menu
13352: * Portability::
13353: * Threading::
13354: * Primitives::
13355: * Performance::
13356: @end menu
13357:
13358: @node Portability, Threading, Engine, Engine
13359: @section Portability
13360: @cindex engine portability
13361:
13362: An important goal of the Gforth Project is availability across a wide
13363: range of personal machines. fig-Forth, and, to a lesser extent, F83,
13364: achieved this goal by manually coding the engine in assembly language
13365: for several then-popular processors. This approach is very
13366: labor-intensive and the results are short-lived due to progress in
13367: computer architecture.
13368:
13369: @cindex C, using C for the engine
13370: Others have avoided this problem by coding in C, e.g., Mitch Bradley
13371: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
13372: particularly popular for UNIX-based Forths due to the large variety of
13373: architectures of UNIX machines. Unfortunately an implementation in C
13374: does not mix well with the goals of efficiency and with using
13375: traditional techniques: Indirect or direct threading cannot be expressed
13376: in C, and switch threading, the fastest technique available in C, is
13377: significantly slower. Another problem with C is that it is very
13378: cumbersome to express double integer arithmetic.
13379:
13380: @cindex GNU C for the engine
13381: @cindex long long
13382: Fortunately, there is a portable language that does not have these
13383: limitations: GNU C, the version of C processed by the GNU C compiler
13384: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
13385: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
13386: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
13387: threading possible, its @code{long long} type (@pxref{Long Long, ,
13388: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
13389: double numbers@footnote{Unfortunately, long longs are not implemented
13390: properly on all machines (e.g., on alpha-osf1, long longs are only 64
13391: bits, the same size as longs (and pointers), but they should be twice as
13392: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
13393: C Manual}). So, we had to implement doubles in C after all. Still, on
13394: most machines we can use long longs and achieve better performance than
13395: with the emulation package.}. GNU C is available for free on all
13396: important (and many unimportant) UNIX machines, VMS, 80386s running
13397: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
13398: on all these machines.
13399:
13400: Writing in a portable language has the reputation of producing code that
13401: is slower than assembly. For our Forth engine we repeatedly looked at
13402: the code produced by the compiler and eliminated most compiler-induced
13403: inefficiencies by appropriate changes in the source code.
13404:
13405: @cindex explicit register declarations
13406: @cindex --enable-force-reg, configuration flag
13407: @cindex -DFORCE_REG
13408: However, register allocation cannot be portably influenced by the
13409: programmer, leading to some inefficiencies on register-starved
13410: machines. We use explicit register declarations (@pxref{Explicit Reg
13411: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
13412: improve the speed on some machines. They are turned on by using the
13413: configuration flag @code{--enable-force-reg} (@code{gcc} switch
13414: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
13415: machine, but also on the compiler version: On some machines some
13416: compiler versions produce incorrect code when certain explicit register
13417: declarations are used. So by default @code{-DFORCE_REG} is not used.
13418:
13419: @node Threading, Primitives, Portability, Engine
13420: @section Threading
13421: @cindex inner interpreter implementation
13422: @cindex threaded code implementation
13423:
13424: @cindex labels as values
13425: GNU C's labels as values extension (available since @code{gcc-2.0},
13426: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
13427: makes it possible to take the address of @i{label} by writing
13428: @code{&&@i{label}}. This address can then be used in a statement like
13429: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
13430: @code{goto x}.
13431:
13432: @cindex @code{NEXT}, indirect threaded
13433: @cindex indirect threaded inner interpreter
13434: @cindex inner interpreter, indirect threaded
13435: With this feature an indirect threaded @code{NEXT} looks like:
13436: @example
13437: cfa = *ip++;
13438: ca = *cfa;
13439: goto *ca;
13440: @end example
13441: @cindex instruction pointer
13442: For those unfamiliar with the names: @code{ip} is the Forth instruction
13443: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
13444: execution token and points to the code field of the next word to be
13445: executed; The @code{ca} (code address) fetched from there points to some
13446: executable code, e.g., a primitive or the colon definition handler
13447: @code{docol}.
13448:
13449: @cindex @code{NEXT}, direct threaded
13450: @cindex direct threaded inner interpreter
13451: @cindex inner interpreter, direct threaded
13452: Direct threading is even simpler:
13453: @example
13454: ca = *ip++;
13455: goto *ca;
13456: @end example
13457:
13458: Of course we have packaged the whole thing neatly in macros called
13459: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
13460:
13461: @menu
13462: * Scheduling::
13463: * Direct or Indirect Threaded?::
13464: * DOES>::
13465: @end menu
13466:
13467: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
13468: @subsection Scheduling
13469: @cindex inner interpreter optimization
13470:
13471: There is a little complication: Pipelined and superscalar processors,
13472: i.e., RISC and some modern CISC machines can process independent
13473: instructions while waiting for the results of an instruction. The
13474: compiler usually reorders (schedules) the instructions in a way that
13475: achieves good usage of these delay slots. However, on our first tries
13476: the compiler did not do well on scheduling primitives. E.g., for
13477: @code{+} implemented as
13478: @example
13479: n=sp[0]+sp[1];
13480: sp++;
13481: sp[0]=n;
13482: NEXT;
13483: @end example
13484: the @code{NEXT} comes strictly after the other code, i.e., there is nearly no
13485: scheduling. After a little thought the problem becomes clear: The
13486: compiler cannot know that @code{sp} and @code{ip} point to different
13487: addresses (and the version of @code{gcc} we used would not know it even
13488: if it was possible), so it could not move the load of the cfa above the
13489: store to the TOS. Indeed the pointers could be the same, if code on or
13490: very near the top of stack were executed. In the interest of speed we
13491: chose to forbid this probably unused ``feature'' and helped the compiler
13492: in scheduling: @code{NEXT} is divided into the loading part (@code{NEXT_P1})
13493: and the goto part (@code{NEXT_P2}). @code{+} now looks like:
13494: @example
13495: n=sp[0]+sp[1];
13496: sp++;
13497: NEXT_P1;
13498: sp[0]=n;
13499: NEXT_P2;
13500: @end example
13501: This can be scheduled optimally by the compiler.
13502:
13503: This division can be turned off with the switch @code{-DCISC_NEXT}. This
13504: switch is on by default on machines that do not profit from scheduling
13505: (e.g., the 80386), in order to preserve registers.
13506:
13507: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
13508: @subsection Direct or Indirect Threaded?
13509: @cindex threading, direct or indirect?
13510:
13511: @cindex -DDIRECT_THREADED
13512: Both! After packaging the nasty details in macro definitions we
13513: realized that we could switch between direct and indirect threading by
13514: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
13515: defining a few machine-specific macros for the direct-threading case.
13516: On the Forth level we also offer access words that hide the
13517: differences between the threading methods (@pxref{Threading Words}).
13518:
13519: Indirect threading is implemented completely machine-independently.
13520: Direct threading needs routines for creating jumps to the executable
13521: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
13522: machine-dependent, but they do not amount to many source lines. Therefore,
13523: even porting direct threading to a new machine requires little effort.
13524:
13525: @cindex --enable-indirect-threaded, configuration flag
13526: @cindex --enable-direct-threaded, configuration flag
13527: The default threading method is machine-dependent. You can enforce a
13528: specific threading method when building Gforth with the configuration
13529: flag @code{--enable-direct-threaded} or
13530: @code{--enable-indirect-threaded}. Note that direct threading is not
13531: supported on all machines.
13532:
13533: @node DOES>, , Direct or Indirect Threaded?, Threading
13534: @subsection DOES>
13535: @cindex @code{DOES>} implementation
13536:
13537: @cindex @code{dodoes} routine
13538: @cindex @code{DOES>}-code
13539: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
13540: the chunk of code executed by every word defined by a
13541: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
13542: the Forth code to be executed, i.e. the code after the
13543: @code{DOES>} (the @code{DOES>}-code)? There are two solutions:
13544:
13545: In fig-Forth the code field points directly to the @code{dodoes} and the
13546: @code{DOES>}-code address is stored in the cell after the code address (i.e. at
13547: @code{@i{CFA} cell+}). It may seem that this solution is illegal in
13548: the Forth-79 and all later standards, because in fig-Forth this address
13549: lies in the body (which is illegal in these standards). However, by
13550: making the code field larger for all words this solution becomes legal
13551: again. We use this approach for the indirect threaded version and for
13552: direct threading on some machines. Leaving a cell unused in most words
13553: is a bit wasteful, but on the machines we are targeting this is hardly a
13554: problem. The other reason for having a code field size of two cells is
13555: to avoid having different image files for direct and indirect threaded
13556: systems (direct threaded systems require two-cell code fields on many
13557: machines).
13558:
13559: @cindex @code{DOES>}-handler
13560: The other approach is that the code field points or jumps to the cell
13561: after @code{DOES>}. In this variant there is a jump to @code{dodoes} at
13562: this address (the @code{DOES>}-handler). @code{dodoes} can then get the
13563: @code{DOES>}-code address by computing the code address, i.e., the address of
13564: the jump to @code{dodoes}, and add the length of that jump field. A variant of
13565: this is to have a call to @code{dodoes} after the @code{DOES>}; then the
13566: return address (which can be found in the return register on RISCs) is
13567: the @code{DOES>}-code address. Since the two cells available in the code field
13568: are used up by the jump to the code address in direct threading on many
13569: architectures, we use this approach for direct threading on these
13570: architectures. We did not want to add another cell to the code field.
13571:
13572: @node Primitives, Performance, Threading, Engine
13573: @section Primitives
13574: @cindex primitives, implementation
13575: @cindex virtual machine instructions, implementation
13576:
13577: @menu
13578: * Automatic Generation::
13579: * TOS Optimization::
13580: * Produced code::
13581: @end menu
13582:
13583: @node Automatic Generation, TOS Optimization, Primitives, Primitives
13584: @subsection Automatic Generation
13585: @cindex primitives, automatic generation
13586:
13587: @cindex @file{prims2x.fs}
13588: Since the primitives are implemented in a portable language, there is no
13589: longer any need to minimize the number of primitives. On the contrary,
13590: having many primitives has an advantage: speed. In order to reduce the
13591: number of errors in primitives and to make programming them easier, we
13592: provide a tool, the primitive generator (@file{prims2x.fs}), that
13593: automatically generates most (and sometimes all) of the C code for a
13594: primitive from the stack effect notation. The source for a primitive
13595: has the following form:
13596:
13597: @cindex primitive source format
13598: @format
13599: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
13600: [@code{""}@i{glossary entry}@code{""}]
13601: @i{C code}
13602: [@code{:}
13603: @i{Forth code}]
13604: @end format
13605:
13606: The items in brackets are optional. The category and glossary fields
13607: are there for generating the documentation, the Forth code is there
13608: for manual implementations on machines without GNU C. E.g., the source
13609: for the primitive @code{+} is:
13610: @example
13611: + ( n1 n2 -- n ) core plus
13612: n = n1+n2;
13613: @end example
13614:
13615: This looks like a specification, but in fact @code{n = n1+n2} is C
13616: code. Our primitive generation tool extracts a lot of information from
13617: the stack effect notations@footnote{We use a one-stack notation, even
13618: though we have separate data and floating-point stacks; The separate
13619: notation can be generated easily from the unified notation.}: The number
13620: of items popped from and pushed on the stack, their type, and by what
13621: name they are referred to in the C code. It then generates a C code
13622: prelude and postlude for each primitive. The final C code for @code{+}
13623: looks like this:
13624:
13625: @example
13626: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
13627: /* */ /* documentation */
13628: @{
13629: DEF_CA /* definition of variable ca (indirect threading) */
13630: Cell n1; /* definitions of variables */
13631: Cell n2;
13632: Cell n;
13633: n1 = (Cell) sp[1]; /* input */
13634: n2 = (Cell) TOS;
13635: sp += 1; /* stack adjustment */
13636: NAME("+") /* debugging output (with -DDEBUG) */
13637: @{
13638: n = n1+n2; /* C code taken from the source */
13639: @}
13640: NEXT_P1; /* NEXT part 1 */
13641: TOS = (Cell)n; /* output */
13642: NEXT_P2; /* NEXT part 2 */
13643: @}
13644: @end example
13645:
13646: This looks long and inefficient, but the GNU C compiler optimizes quite
13647: well and produces optimal code for @code{+} on, e.g., the R3000 and the
13648: HP RISC machines: Defining the @code{n}s does not produce any code, and
13649: using them as intermediate storage also adds no cost.
13650:
13651: There are also other optimizations that are not illustrated by this
13652: example: assignments between simple variables are usually for free (copy
13653: propagation). If one of the stack items is not used by the primitive
13654: (e.g. in @code{drop}), the compiler eliminates the load from the stack
13655: (dead code elimination). On the other hand, there are some things that
13656: the compiler does not do, therefore they are performed by
13657: @file{prims2x.fs}: The compiler does not optimize code away that stores
13658: a stack item to the place where it just came from (e.g., @code{over}).
13659:
13660: While programming a primitive is usually easy, there are a few cases
13661: where the programmer has to take the actions of the generator into
13662: account, most notably @code{?dup}, but also words that do not (always)
13663: fall through to @code{NEXT}.
13664:
13665: @node TOS Optimization, Produced code, Automatic Generation, Primitives
13666: @subsection TOS Optimization
13667: @cindex TOS optimization for primitives
13668: @cindex primitives, keeping the TOS in a register
13669:
13670: An important optimization for stack machine emulators, e.g., Forth
13671: engines, is keeping one or more of the top stack items in
13672: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
13673: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
13674: @itemize @bullet
13675: @item
13676: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
13677: due to fewer loads from and stores to the stack.
13678: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
13679: @i{y<n}, due to additional moves between registers.
13680: @end itemize
13681:
13682: @cindex -DUSE_TOS
13683: @cindex -DUSE_NO_TOS
13684: In particular, keeping one item in a register is never a disadvantage,
13685: if there are enough registers. Keeping two items in registers is a
13686: disadvantage for frequent words like @code{?branch}, constants,
13687: variables, literals and @code{i}. Therefore our generator only produces
13688: code that keeps zero or one items in registers. The generated C code
13689: covers both cases; the selection between these alternatives is made at
13690: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
13691: code for @code{+} is just a simple variable name in the one-item case,
13692: otherwise it is a macro that expands into @code{sp[0]}. Note that the
13693: GNU C compiler tries to keep simple variables like @code{TOS} in
13694: registers, and it usually succeeds, if there are enough registers.
13695:
13696: @cindex -DUSE_FTOS
13697: @cindex -DUSE_NO_FTOS
13698: The primitive generator performs the TOS optimization for the
13699: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
13700: operations the benefit of this optimization is even larger:
13701: floating-point operations take quite long on most processors, but can be
13702: performed in parallel with other operations as long as their results are
13703: not used. If the FP-TOS is kept in a register, this works. If
13704: it is kept on the stack, i.e., in memory, the store into memory has to
13705: wait for the result of the floating-point operation, lengthening the
13706: execution time of the primitive considerably.
13707:
13708: The TOS optimization makes the automatic generation of primitives a
13709: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
13710: @code{TOS} is not sufficient. There are some special cases to
13711: consider:
13712: @itemize @bullet
13713: @item In the case of @code{dup ( w -- w w )} the generator must not
13714: eliminate the store to the original location of the item on the stack,
13715: if the TOS optimization is turned on.
13716: @item Primitives with stack effects of the form @code{--}
13717: @i{out1}...@i{outy} must store the TOS to the stack at the start.
13718: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
13719: must load the TOS from the stack at the end. But for the null stack
13720: effect @code{--} no stores or loads should be generated.
13721: @end itemize
13722:
13723: @node Produced code, , TOS Optimization, Primitives
13724: @subsection Produced code
13725: @cindex primitives, assembly code listing
13726:
13727: @cindex @file{engine.s}
13728: To see what assembly code is produced for the primitives on your machine
13729: with your compiler and your flag settings, type @code{make engine.s} and
13730: look at the resulting file @file{engine.s}.
13731:
13732: @node Performance, , Primitives, Engine
13733: @section Performance
13734: @cindex performance of some Forth interpreters
13735: @cindex engine performance
13736: @cindex benchmarking Forth systems
13737: @cindex Gforth performance
13738:
13739: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
13740: impossible to write a significantly faster engine.
13741:
13742: On register-starved machines like the 386 architecture processors
13743: improvements are possible, because @code{gcc} does not utilize the
13744: registers as well as a human, even with explicit register declarations;
13745: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
13746: and hand-tuned it for the 486; this system is 1.19 times faster on the
13747: Sieve benchmark on a 486DX2/66 than Gforth compiled with
13748: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
13749: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
13750: registers fit in real registers (and we can even afford to use the TOS
13751: optimization), resulting in a speedup of 1.14 on the sieve over the
13752: earlier results.
13753:
13754: @cindex Win32Forth performance
13755: @cindex NT Forth performance
13756: @cindex eforth performance
13757: @cindex ThisForth performance
13758: @cindex PFE performance
13759: @cindex TILE performance
13760: The potential advantage of assembly language implementations
13761: is not necessarily realized in complete Forth systems: We compared
13762: Gforth-0.4.9 (direct threaded, compiled with @code{gcc-2.95.1} and
13763: @code{-DFORCE_REG}) with Win32Forth 1.2093, LMI's NT Forth (Beta, May
13764: 1994) and Eforth (with and without peephole (aka pinhole) optimization
13765: of the threaded code); all these systems were written in assembly
13766: language. We also compared Gforth with three systems written in C:
13767: PFE-0.9.14 (compiled with @code{gcc-2.6.3} with the default
13768: configuration for Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS
13769: -DUNROLL_NEXT}), ThisForth Beta (compiled with @code{gcc-2.6.3 -O3
13770: -fomit-frame-pointer}; ThisForth employs peephole optimization of the
13771: threaded code) and TILE (compiled with @code{make opt}). We benchmarked
13772: Gforth, PFE, ThisForth and TILE on a 486DX2/66 under Linux. Kenneth
13773: O'Heskin kindly provided the results for Win32Forth and NT Forth on a
13774: 486DX2/66 with similar memory performance under Windows NT. Marcel
13775: Hendrix ported Eforth to Linux, then extended it to run the benchmarks,
13776: added the peephole optimizer, ran the benchmarks and reported the
13777: results.
13778:
13779: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
13780: matrix multiplication come from the Stanford integer benchmarks and have
13781: been translated into Forth by Martin Fraeman; we used the versions
13782: included in the TILE Forth package, but with bigger data set sizes; and
13783: a recursive Fibonacci number computation for benchmarking calling
13784: performance. The following table shows the time taken for the benchmarks
13785: scaled by the time taken by Gforth (in other words, it shows the speedup
13786: factor that Gforth achieved over the other systems).
13787:
13788: @example
13789: relative Win32- NT eforth This-
13790: time Gforth Forth Forth eforth +opt PFE Forth TILE
13791: sieve 1.00 1.58 1.30 1.58 0.97 1.80 3.63 9.79
13792: bubble 1.00 1.55 1.67 1.75 1.04 1.78 4.59
13793: matmul 1.00 1.67 1.53 1.66 0.84 1.79 4.63
13794: fib 1.00 1.75 1.53 1.40 0.99 1.99 3.43 4.93
13795: @end example
13796:
13797: You may be quite surprised by the good performance of Gforth when
13798: compared with systems written in assembly language. One important reason
13799: for the disappointing performance of these other systems is probably
13800: that they are not written optimally for the 486 (e.g., they use the
13801: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
13802: but costly method for relocating the Forth image: like @code{cforth}, it
13803: computes the actual addresses at run time, resulting in two address
13804: computations per @code{NEXT} (@pxref{Image File Background}).
13805:
13806: Only Eforth with the peephole optimizer performs comparable to
13807: Gforth. The speedups achieved with peephole optimization of threaded
13808: code are quite remarkable. Adding a peephole optimizer to Gforth should
13809: cause similar speedups.
13810:
13811: The speedup of Gforth over PFE, ThisForth and TILE can be easily
13812: explained with the self-imposed restriction of the latter systems to
13813: standard C, which makes efficient threading impossible (however, the
13814: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
13815: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
13816: Moreover, current C compilers have a hard time optimizing other aspects
13817: of the ThisForth and the TILE source.
13818:
13819: The performance of Gforth on 386 architecture processors varies widely
13820: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
13821: allocate any of the virtual machine registers into real machine
13822: registers by itself and would not work correctly with explicit register
13823: declarations, giving a 1.5 times slower engine (on a 486DX2/66 running
13824: the Sieve) than the one measured above.
13825:
13826: Note that there have been several releases of Win32Forth since the
13827: release presented here, so the results presented above may have little
13828: predictive value for the performance of Win32Forth today (results for
13829: the current release on an i486DX2/66 are welcome).
13830:
13831: @cindex @file{Benchres}
13832: In @cite{Translating Forth to Efficient C} by M. Anton Ertl and Martin
13833: Maierhofer (presented at EuroForth '95), an indirect threaded version of
13834: Gforth is compared with Win32Forth, NT Forth, PFE, and ThisForth; that
13835: version of Gforth is slower on a 486 than the direct threaded version
13836: used here. The paper available at
13837: @*@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz};
13838: it also contains numbers for some native code systems. You can find a
13839: newer version of these measurements at
13840: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
13841: find numbers for Gforth on various machines in @file{Benchres}.
13842:
13843: @c ******************************************************************
13844: @node Binding to System Library, Cross Compiler, Engine, Top
13845: @chapter Binding to System Library
13846:
13847: @node Cross Compiler, Bugs, Binding to System Library, Top
13848: @chapter Cross Compiler
13849: @cindex @file{cross.fs}
13850: @cindex cross-compiler
13851: @cindex metacompiler
13852: @cindex target compiler
13853:
13854: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
13855: mostly written in Forth, including crucial parts like the outer
13856: interpreter and compiler, it needs compiled Forth code to get
13857: started. The cross compiler allows to create new images for other
13858: architectures, even running under another Forth system.
13859:
13860: @menu
13861: * Using the Cross Compiler::
13862: * How the Cross Compiler Works::
13863: @end menu
13864:
13865: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
13866: @section Using the Cross Compiler
13867:
13868: The cross compiler uses a language that resembles Forth, but isn't. The
13869: main difference is that you can execute Forth code after definition,
13870: while you usually can't execute the code compiled by cross, because the
13871: code you are compiling is typically for a different computer than the
13872: one you are compiling on.
13873:
13874: The Makefile is already set up to allow you to create kernels for new
13875: architectures with a simple make command. The generic kernels using the
13876: GCC compiled virtual machine are created in the normal build process
13877: with @code{make}. To create a embedded Gforth executable for e.g. the
13878: 8086 processor (running on a DOS machine), type
13879:
13880: @example
13881: make kernl-8086.fi
13882: @end example
13883:
13884: This will use the machine description from the @file{arch/8086}
13885: directory to create a new kernel. A machine file may look like that:
13886:
13887: @example
13888: \ Parameter for target systems 06oct92py
13889:
13890: 4 Constant cell \ cell size in bytes
13891: 2 Constant cell<< \ cell shift to bytes
13892: 5 Constant cell>bit \ cell shift to bits
13893: 8 Constant bits/char \ bits per character
13894: 8 Constant bits/byte \ bits per byte [default: 8]
13895: 8 Constant float \ bytes per float
13896: 8 Constant /maxalign \ maximum alignment in bytes
13897: false Constant bigendian \ byte order
13898: ( true=big, false=little )
13899:
13900: include machpc.fs \ feature list
13901: @end example
13902:
13903: This part is obligatory for the cross compiler itself, the feature list
13904: is used by the kernel to conditionally compile some features in and out,
13905: depending on whether the target supports these features.
13906:
13907: There are some optional features, if you define your own primitives,
13908: have an assembler, or need special, nonstandard preparation to make the
13909: boot process work. @code{asm-include} include an assembler,
13910: @code{prims-include} includes primitives, and @code{>boot} prepares for
13911: booting.
13912:
13913: @example
13914: : asm-include ." Include assembler" cr
13915: s" arch/8086/asm.fs" included ;
13916:
13917: : prims-include ." Include primitives" cr
13918: s" arch/8086/prim.fs" included ;
13919:
13920: : >boot ." Prepare booting" cr
13921: s" ' boot >body into-forth 1+ !" evaluate ;
13922: @end example
13923:
13924: These words are used as sort of macro during the cross compilation in
13925: the file @file{kernel/main.fs}. Instead of using this macros, it would
13926: be possible --- but more complicated --- to write a new kernel project
13927: file, too.
13928:
13929: @file{kernel/main.fs} expects the machine description file name on the
13930: stack; the cross compiler itself (@file{cross.fs}) assumes that either
13931: @code{mach-file} leaves a counted string on the stack, or
13932: @code{machine-file} leaves an address, count pair of the filename on the
13933: stack.
13934:
13935: The feature list is typically controlled using @code{SetValue}, generic
13936: files that are used by several projects can use @code{DefaultValue}
13937: instead. Both functions work like @code{Value}, when the value isn't
13938: defined, but @code{SetValue} works like @code{to} if the value is
13939: defined, and @code{DefaultValue} doesn't set anything, if the value is
13940: defined.
13941:
13942: @example
13943: \ generic mach file for pc gforth 03sep97jaw
13944:
13945: true DefaultValue NIL \ relocating
13946:
13947: >ENVIRON
13948:
13949: true DefaultValue file \ controls the presence of the
13950: \ file access wordset
13951: true DefaultValue OS \ flag to indicate a operating system
13952:
13953: true DefaultValue prims \ true: primitives are c-code
13954:
13955: true DefaultValue floating \ floating point wordset is present
13956:
13957: true DefaultValue glocals \ gforth locals are present
13958: \ will be loaded
13959: true DefaultValue dcomps \ double number comparisons
13960:
13961: true DefaultValue hash \ hashing primitives are loaded/present
13962:
13963: true DefaultValue xconds \ used together with glocals,
13964: \ special conditionals supporting gforths'
13965: \ local variables
13966: true DefaultValue header \ save a header information
13967:
13968: true DefaultValue backtrace \ enables backtrace code
13969:
13970: false DefaultValue ec
13971: false DefaultValue crlf
13972:
13973: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
13974:
13975: &16 KB DefaultValue stack-size
13976: &15 KB &512 + DefaultValue fstack-size
13977: &15 KB DefaultValue rstack-size
13978: &14 KB &512 + DefaultValue lstack-size
13979: @end example
13980:
13981: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
13982: @section How the Cross Compiler Works
13983:
13984: @node Bugs, Origin, Cross Compiler, Top
13985: @appendix Bugs
13986: @cindex bug reporting
13987:
13988: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
13989:
13990: If you find a bug, please send a bug report to
13991: @email{bug-gforth@@gnu.org}. A bug report should include this
13992: information:
13993:
13994: @itemize @bullet
13995: @item
13996: The Gforth version used (it is announced at the start of an
13997: interactive Gforth session).
13998: @item
13999: The machine and operating system (on Unix
14000: systems @code{uname -a} will report this information).
14001: @item
14002: The installation options (send the file @file{config.status}).
14003: @item
14004: A complete list of changes (if any) you (or your installer) have made to the
14005: Gforth sources.
14006: @item
14007: A program (or a sequence of keyboard commands) that reproduces the bug.
14008: @item
14009: A description of what you think constitutes the buggy behaviour.
14010: @end itemize
14011:
14012: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
14013: to Report Bugs, gcc.info, GNU C Manual}.
14014:
14015:
14016: @node Origin, Forth-related information, Bugs, Top
14017: @appendix Authors and Ancestors of Gforth
14018:
14019: @section Authors and Contributors
14020: @cindex authors of Gforth
14021: @cindex contributors to Gforth
14022:
14023: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
14024: Ertl. The third major author was Jens Wilke. Lennart Benschop (who was
14025: one of Gforth's first users, in mid-1993) and Stuart Ramsden inspired us
14026: with their continuous feedback. Lennart Benshop contributed
14027: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
14028: support for calling C libraries. Helpful comments also came from Paul
14029: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
14030: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, and
14031: Robert Epprecht. Since the release of Gforth-0.2.1 there were also
14032: helpful comments from many others; thank you all, sorry for not listing
14033: you here (but digging through my mailbox to extract your names is on my
14034: to-do list). Since the release of Gforth-0.4.0 Neal Crook worked on the
14035: manual.
14036:
14037: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
14038: and autoconf, among others), and to the creators of the Internet: Gforth
14039: was developed across the Internet, and its authors did not meet
14040: physically for the first 4 years of development.
14041:
14042: @section Pedigree
14043: @cindex pedigree of Gforth
14044:
14045: Gforth descends from bigFORTH (1993) and fig-Forth. Gforth and PFE (by
14046: Dirk Zoller) will cross-fertilize each other. Of course, a significant
14047: part of the design of Gforth was prescribed by ANS Forth.
14048:
14049: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
14050: 32 bit native code version of VolksForth for the Atari ST, written
14051: mostly by Dietrich Weineck.
14052:
14053: VolksForth descends from F83. It was written by Klaus Schleisiek, Bernd
14054: Pennemann, Georg Rehfeld and Dietrich Weineck for the C64 (called
14055: UltraForth there) in the mid-80s and ported to the Atari ST in 1986.
14056:
14057: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
14058: Forth-83 standard. !! Pedigree? When?
14059:
14060: A team led by Bill Ragsdale implemented fig-Forth on many processors in
14061: 1979. Robert Selzer and Bill Ragsdale developed the original
14062: implementation of fig-Forth for the 6502 based on microForth.
14063:
14064: The principal architect of microForth was Dean Sanderson. microForth was
14065: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
14066: the 1802, and subsequently implemented on the 8080, the 6800 and the
14067: Z80.
14068:
14069: All earlier Forth systems were custom-made, usually by Charles Moore,
14070: who discovered (as he puts it) Forth during the late 60s. The first full
14071: Forth existed in 1971.
14072:
14073: A part of the information in this section comes from @cite{The Evolution
14074: of Forth} by Elizabeth D. Rather, Donald R. Colburn and Charles
14075: H. Moore, presented at the HOPL-II conference and preprinted in SIGPLAN
14076: Notices 28(3), 1993. You can find more historical and genealogical
14077: information about Forth there.
14078:
14079: @node Forth-related information, Word Index, Origin, Top
14080: @appendix Other Forth-related information
14081: @cindex Forth-related information
14082:
14083: @menu
14084: * Internet resources::
14085: * Books::
14086: * The Forth Interest Group::
14087: * Conferences::
14088: @end menu
14089:
14090:
14091: @node Internet resources, Books, Forth-related information, Forth-related information
14092: @section Internet resources
14093: @cindex internet resources
14094:
14095: @cindex comp.lang.forth
14096: @cindex frequently asked questions
14097: There is an active news group (comp.lang.forth) discussing Forth and
14098: Forth-related issues. A frequently-asked-questions (FAQ) list
14099: is posted to the news group regularly, and archived at these sites:
14100:
14101: @itemize @bullet
14102: @item
14103: @uref{ftp://rtfm.mit.edu/pub/usenet-by-group/comp.lang.forth/}
14104: @item
14105: @uref{ftp://ftp.forth.org/pub/Forth/FAQ/}
14106: @end itemize
14107:
14108: The FAQ list should be considered mandatory reading before posting to
14109: the news group.
14110:
14111: Here are some other web sites holding Forth-related material:
14112:
14113: @itemize @bullet
14114: @item
14115: @uref{http://www.taygeta.com/forth.html} -- Skip Carter's Forth pages.
14116: @item
14117: @uref{http://www.jwdt.com/~paysan/gforth.html} -- the Gforth home page.
14118: @item
14119: @uref{http://www.minerva.com/uathena.htm} -- home of ANS Forth Standard.
14120: @item
14121: @uref{http://dec.bournemouth.ac.uk/forth/index.html} -- the Forth
14122: Research page, including links to the Journal of Forth Application and
14123: Research (JFAR) and a searchable Forth bibliography.
14124: @end itemize
14125:
14126:
14127: @node Books, The Forth Interest Group, Internet resources, Forth-related information
14128: @section Books
14129: @cindex books on Forth
14130:
14131: As the Standard is relatively new, there are not many books out yet. It
14132: is not recommended to learn Forth by using Gforth and a book that is not
14133: written for ANS Forth, as you will not know your mistakes from the
14134: deviations of the book. However, books based on the Forth-83 standard
14135: should be ok, because ANS Forth is primarily an extension of Forth-83.
14136: Refer to the Forth FAQ for details of Forth-related books.
14137:
14138: @cindex standard document for ANS Forth
14139: @cindex ANS Forth document
14140: The definite reference if you want to write ANS Forth programs is, of
14141: course, the ANS Forth document. It is available in printed form from the
14142: National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
14143: Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about
14144: $200. You can also get it from Global Engineering Documents (Tel.: USA
14145: (800) 854-7179; Fax.: (303) 843-9880) for about $300.
14146:
14147: @cite{dpANS6}, the last draft of the standard, which was then submitted
14148: to ANSI for publication is available electronically and for free in some
14149: MS Word format, and it has been converted to HTML
14150: (@uref{http://www.taygeta.com/forth/dpans.html}; this HTML version also
14151: includes the answers to Requests for Interpretation (RFIs). Some
14152: pointers to these versions can be found through
14153: @*@uref{http://www.complang.tuwien.ac.at/projects/forth.html}.
14154:
14155:
14156: @node The Forth Interest Group, Conferences, Books, Forth-related information
14157: @section The Forth Interest Group
14158: @cindex Forth interest group (FIG)
14159:
14160: The Forth Interest Group (FIG) is a world-wide, non-profit,
14161: member-supported organisation. It publishes a regular magazine,
14162: @var{FORTH Dimensions}, and offers other benefits of membership. You can
14163: contact the FIG through their office email address:
14164: @email{office@@forth.org} or by visiting their web site at
14165: @uref{http://www.forth.org/}. This web site also includes links to FIG
14166: chapters in other countries and American cities
14167: (@uref{http://www.forth.org/chapters.html}).
14168:
14169: @node Conferences, , The Forth Interest Group, Forth-related information
14170: @section Conferences
14171: @cindex Conferences
14172:
14173: There are several regular conferences related to Forth. They are all
14174: well-publicised in @var{FORTH Dimensions} and on the comp.lang.forth
14175: news group:
14176:
14177: @itemize @bullet
14178: @item
14179: FORML -- the Forth modification laboratory convenes every year near
14180: Monterey, California.
14181: @item
14182: The Rochester Forth Conference -- an annual conference traditionally
14183: held in Rochester, New York.
14184: @item
14185: EuroForth -- this European conference takes place annually.
14186: @end itemize
14187:
14188:
14189: @node Word Index, Name Index, Forth-related information, Top
14190: @unnumbered Word Index
14191:
14192: This index is a list of Forth words that have ``glossary'' entries
14193: within this manual. Each word is listed with its stack effect and
14194: wordset.
14195:
14196: @printindex fn
14197:
14198: @node Name Index, Concept Index, Word Index, Top
14199: @unnumbered Name Index
14200:
14201: This index is a list of Forth words that have ``glossary'' entries
14202: within this manual.
14203:
14204: @printindex ky
14205:
14206: @node Concept Index, , Name Index, Top
14207: @unnumbered Concept and Word Index
14208:
14209: Not all entries listed in this index are present verbatim in the
14210: text. This index also duplicates, in abbreviated form, all of the words
14211: listed in the Word Index (only the names are listed for the words here).
14212:
14213: @printindex cp
14214:
14215: @contents
14216: @bye
14217:
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