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: @c POSTPONE, COMPILE, [COMPILE], LITERAL should have their own section
18:
19: @comment %**start of header (This is for running Texinfo on a region.)
20: @setfilename gforth.info
21: @settitle Gforth Manual
22: @dircategory GNU programming tools
23: @direntry
24: * Gforth: (gforth). A fast interpreter for the Forth language.
25: @end direntry
26: @c The Texinfo manual also recommends doing this, but for Gforth it may
27: @c not make much sense
28: @c @dircategory Individual utilities
29: @c @direntry
30: @c * Gforth: (gforth)Invoking Gforth. gforth, gforth-fast, gforthmi
31: @c @end direntry
32:
33: @comment @setchapternewpage odd
34: @comment TODO this gets left in by HTML converter
35: @macro progstyle {}
36: Programming style note:
37: @end macro
38:
39: @macro assignment {}
40: @table @i
41: @item Assignment:
42: @end macro
43: @macro endassignment {}
44: @end table
45: @end macro
46:
47: @comment %**end of header (This is for running Texinfo on a region.)
48:
49:
50: @comment ----------------------------------------------------------
51: @comment macros for beautifying glossary entries
52: @comment if these are used, need to strip them out for HTML converter
53: @comment else they get repeated verbatim in HTML output.
54: @comment .. not working yet.
55:
56: @macro GLOSS-START {}
57: @iftex
58: @ninerm
59: @end iftex
60: @end macro
61:
62: @macro GLOSS-END {}
63: @iftex
64: @rm
65: @end iftex
66: @end macro
67:
68: @comment ----------------------------------------------------------
69:
70:
71: @include version.texi
72:
73: @ifnottex
74: This file documents Gforth @value{VERSION}
75:
76: Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
77:
78: Permission is granted to make and distribute verbatim copies of
79: this manual provided the copyright notice and this permission notice
80: are preserved on all copies.
81:
82: @ignore
83: Permission is granted to process this file through TeX and print the
84: results, provided the printed document carries a copying permission
85: notice identical to this one except for the removal of this paragraph
86: (this paragraph not being relevant to the printed manual).
87:
88: @end ignore
89: Permission is granted to copy and distribute modified versions of this
90: manual under the conditions for verbatim copying, provided also that the
91: sections entitled "Distribution" and "General Public License" are
92: included exactly as in the original, and provided that the entire
93: resulting derived work is distributed under the terms of a permission
94: notice identical to this one.
95:
96: Permission is granted to copy and distribute translations of this manual
97: into another language, under the above conditions for modified versions,
98: except that the sections entitled "Distribution" and "General Public
99: License" may be included in a translation approved by the author instead
100: of in the original English.
101: @end ifnottex
102:
103: @finalout
104: @titlepage
105: @sp 10
106: @center @titlefont{Gforth Manual}
107: @sp 2
108: @center for version @value{VERSION}
109: @sp 2
110: @center Neal Crook
111: @center Anton Ertl
112: @center Bernd Paysan
113: @center Jens Wilke
114: @sp 3
115: @center This manual is permanently under construction and was last updated on 15-Mar-2000
116:
117: @comment The following two commands start the copyright page.
118: @page
119: @vskip 0pt plus 1filll
120: Copyright @copyright{} 1995--2000 Free Software Foundation, Inc.
121:
122: @comment !! Published by ... or You can get a copy of this manual ...
123:
124: Permission is granted to make and distribute verbatim copies of
125: this manual provided the copyright notice and this permission notice
126: are preserved on all copies.
127:
128: Permission is granted to copy and distribute modified versions of this
129: manual under the conditions for verbatim copying, provided also that the
130: sections entitled "Distribution" and "General Public License" are
131: included exactly as in the original, and provided that the entire
132: resulting derived work is distributed under the terms of a permission
133: notice identical to this one.
134:
135: Permission is granted to copy and distribute translations of this manual
136: into another language, under the above conditions for modified versions,
137: except that the sections entitled "Distribution" and "General Public
138: License" may be included in a translation approved by the author instead
139: of in the original English.
140: @end titlepage
141:
142: @node Top, License, (dir), (dir)
143: @ifnottex
144: Gforth is a free implementation of ANS Forth available on many
145: personal machines. This manual corresponds to version @value{VERSION}.
146: @end ifnottex
147:
148: @menu
149: * License:: The GPL
150: * Goals:: About the Gforth Project
151: * Gforth Environment:: Starting (and exiting) Gforth
152: * Tutorial:: Hands-on Forth Tutorial
153: * Introduction:: An introduction to ANS Forth
154: * Words:: Forth words available in Gforth
155: * Error messages:: How to interpret them
156: * Tools:: Programming tools
157: * ANS conformance:: Implementation-defined options etc.
158: * Standard vs Extensions:: Should I use extensions?
159: * Model:: The abstract machine of Gforth
160: * Integrating Gforth:: Forth as scripting language for applications
161: * Emacs and Gforth:: The Gforth Mode
162: * Image Files:: @code{.fi} files contain compiled code
163: * Engine:: The inner interpreter and the primitives
164: * Binding to System Library::
165: * Cross Compiler:: The Cross Compiler
166: * Bugs:: How to report them
167: * Origin:: Authors and ancestors of Gforth
168: * Forth-related information:: Books and places to look on the WWW
169: * Word Index:: An item for each Forth word
170: * Name Index:: Forth words, only names listed
171: * Concept Index:: A menu covering many topics
172:
173: @detailmenu --- The Detailed Node Listing ---
174:
175: Gforth Environment
176:
177: * Invoking Gforth:: Getting in
178: * Leaving Gforth:: Getting out
179: * Command-line editing::
180: * Environment variables:: that affect how Gforth starts up
181: * Gforth Files:: What gets installed and where
182: * Startup speed:: When 35ms is not fast enough ...
183:
184: Forth Tutorial
185:
186: * Starting Gforth Tutorial::
187: * Syntax Tutorial::
188: * Crash Course Tutorial::
189: * Stack Tutorial::
190: * Arithmetics Tutorial::
191: * Stack Manipulation Tutorial::
192: * Using files for Forth code Tutorial::
193: * Comments Tutorial::
194: * Colon Definitions Tutorial::
195: * Decompilation Tutorial::
196: * Stack-Effect Comments Tutorial::
197: * Types Tutorial::
198: * Factoring Tutorial::
199: * Designing the stack effect Tutorial::
200: * Local Variables Tutorial::
201: * Conditional execution Tutorial::
202: * Flags and Comparisons Tutorial::
203: * General Loops Tutorial::
204: * Counted loops Tutorial::
205: * Recursion Tutorial::
206: * Leaving definitions or loops Tutorial::
207: * Return Stack Tutorial::
208: * Memory Tutorial::
209: * Characters and Strings Tutorial::
210: * Alignment Tutorial::
211: * Interpretation and Compilation Semantics and Immediacy Tutorial::
212: * Execution Tokens Tutorial::
213: * Exceptions Tutorial::
214: * Defining Words Tutorial::
215: * Arrays and Records Tutorial::
216: * POSTPONE Tutorial::
217: * Literal Tutorial::
218: * Advanced macros Tutorial::
219: * Compilation Tokens Tutorial::
220: * Wordlists and Search Order Tutorial::
221:
222: An Introduction to ANS Forth
223:
224: * Introducing the Text Interpreter::
225: * Stacks and Postfix notation::
226: * Your first definition::
227: * How does that work?::
228: * Forth is written in Forth::
229: * Review - elements of a Forth system::
230: * Where to go next::
231: * Exercises::
232:
233: Forth Words
234:
235: * Notation::
236: * Case insensitivity::
237: * Comments::
238: * Boolean Flags::
239: * Arithmetic::
240: * Stack Manipulation::
241: * Memory::
242: * Control Structures::
243: * Defining Words::
244: * Interpretation and Compilation Semantics::
245: * Tokens for Words::
246: * The Text Interpreter::
247: * Word Lists::
248: * Environmental Queries::
249: * Files::
250: * Blocks::
251: * Other I/O::
252: * Programming Tools::
253: * Assembler and Code Words::
254: * Threading Words::
255: * Locals::
256: * Structures::
257: * Object-oriented Forth::
258: * Passing Commands to the OS::
259: * Keeping track of Time::
260: * Miscellaneous Words::
261:
262: Arithmetic
263:
264: * Single precision::
265: * Bitwise operations::
266: * Double precision:: Double-cell integer arithmetic
267: * Numeric comparison::
268: * Mixed precision:: Operations with single and double-cell integers
269: * Floating Point::
270:
271: Stack Manipulation
272:
273: * Data stack::
274: * Floating point stack::
275: * Return stack::
276: * Locals stack::
277: * Stack pointer manipulation::
278:
279: Memory
280:
281: * Memory model::
282: * Dictionary allocation::
283: * Heap Allocation::
284: * Memory Access::
285: * Address arithmetic::
286: * Memory Blocks::
287:
288: Control Structures
289:
290: * Selection:: IF ... ELSE ... ENDIF
291: * Simple Loops:: BEGIN ...
292: * Counted Loops:: DO
293: * Arbitrary control structures::
294: * Calls and returns::
295: * Exception Handling::
296:
297: Defining Words
298:
299: * CREATE::
300: * Variables:: Variables and user variables
301: * Constants::
302: * Values:: Initialised variables
303: * Colon Definitions::
304: * Anonymous Definitions:: Definitions without names
305: * User-defined Defining Words::
306: * Deferred words:: Allow forward references
307: * Aliases::
308: * Supplying names::
309:
310: User-defined Defining Words
311:
312: * CREATE..DOES> applications::
313: * CREATE..DOES> details::
314: * Advanced does> usage example::
315:
316: Interpretation and Compilation Semantics
317:
318: * Combined words::
319:
320: The Text Interpreter
321:
322: * Input Sources::
323: * Number Conversion::
324: * Interpret/Compile states::
325: * Literals::
326: * Interpreter Directives::
327:
328: Word Lists
329:
330: * Why use word lists?::
331: * Word list examples::
332:
333: Files
334:
335: * Forth source files::
336: * General files::
337: * Search Paths::
338:
339: Search Paths
340:
341: * Forth Search Paths::
342: * General Search Paths::
343:
344: Other I/O
345:
346: * Simple numeric output:: Predefined formats
347: * Formatted numeric output:: Formatted (pictured) output
348: * String Formats:: How Forth stores strings in memory
349: * Displaying characters and strings:: Other stuff
350: * Input:: Input
351:
352: Programming Tools
353:
354: * Debugging:: Simple and quick.
355: * Assertions:: Making your programs self-checking.
356: * Singlestep Debugger:: Executing your program word by word.
357:
358: Assembler and Code Words
359:
360: * Code and ;code::
361: * Common Assembler:: Assembler Syntax
362: * Common Disassembler::
363: * 386 Assembler:: Deviations and special cases
364: * Alpha Assembler:: Deviations and special cases
365: * MIPS assembler:: Deviations and special cases
366: * Other assemblers:: How to write them
367:
368: Locals
369:
370: * Gforth locals::
371: * ANS Forth locals::
372:
373: Gforth locals
374:
375: * Where are locals visible by name?::
376: * How long do locals live?::
377: * Programming Style::
378: * Implementation::
379:
380: Structures
381:
382: * Why explicit structure support?::
383: * Structure Usage::
384: * Structure Naming Convention::
385: * Structure Implementation::
386: * Structure Glossary::
387:
388: Object-oriented Forth
389:
390: * Why object-oriented programming?::
391: * Object-Oriented Terminology::
392: * Objects::
393: * OOF::
394: * Mini-OOF::
395: * Comparison with other object models::
396:
397: The @file{objects.fs} model
398:
399: * Properties of the Objects model::
400: * Basic Objects Usage::
401: * The Objects base class::
402: * Creating objects::
403: * Object-Oriented Programming Style::
404: * Class Binding::
405: * Method conveniences::
406: * Classes and Scoping::
407: * Dividing classes::
408: * Object Interfaces::
409: * Objects Implementation::
410: * Objects Glossary::
411:
412: The @file{oof.fs} model
413:
414: * Properties of the OOF model::
415: * Basic OOF Usage::
416: * The OOF base class::
417: * Class Declaration::
418: * Class Implementation::
419:
420: The @file{mini-oof.fs} model
421:
422: * Basic Mini-OOF Usage::
423: * Mini-OOF Example::
424: * Mini-OOF Implementation::
425: * Comparison with other object models::
426:
427: Tools
428:
429: * ANS Report:: Report the words used, sorted by wordset.
430:
431: ANS conformance
432:
433: * The Core Words::
434: * The optional Block word set::
435: * The optional Double Number word set::
436: * The optional Exception word set::
437: * The optional Facility word set::
438: * The optional File-Access word set::
439: * The optional Floating-Point word set::
440: * The optional Locals word set::
441: * The optional Memory-Allocation word set::
442: * The optional Programming-Tools word set::
443: * The optional Search-Order word set::
444:
445: The Core Words
446:
447: * core-idef:: Implementation Defined Options
448: * core-ambcond:: Ambiguous Conditions
449: * core-other:: Other System Documentation
450:
451: The optional Block word set
452:
453: * block-idef:: Implementation Defined Options
454: * block-ambcond:: Ambiguous Conditions
455: * block-other:: Other System Documentation
456:
457: The optional Double Number word set
458:
459: * double-ambcond:: Ambiguous Conditions
460:
461: The optional Exception word set
462:
463: * exception-idef:: Implementation Defined Options
464:
465: The optional Facility word set
466:
467: * facility-idef:: Implementation Defined Options
468: * facility-ambcond:: Ambiguous Conditions
469:
470: The optional File-Access word set
471:
472: * file-idef:: Implementation Defined Options
473: * file-ambcond:: Ambiguous Conditions
474:
475: The optional Floating-Point word set
476:
477: * floating-idef:: Implementation Defined Options
478: * floating-ambcond:: Ambiguous Conditions
479:
480: The optional Locals word set
481:
482: * locals-idef:: Implementation Defined Options
483: * locals-ambcond:: Ambiguous Conditions
484:
485: The optional Memory-Allocation word set
486:
487: * memory-idef:: Implementation Defined Options
488:
489: The optional Programming-Tools word set
490:
491: * programming-idef:: Implementation Defined Options
492: * programming-ambcond:: Ambiguous Conditions
493:
494: The optional Search-Order word set
495:
496: * search-idef:: Implementation Defined Options
497: * search-ambcond:: Ambiguous Conditions
498:
499: Image Files
500:
501: * Image Licensing Issues:: Distribution terms for images.
502: * Image File Background:: Why have image files?
503: * Non-Relocatable Image Files:: don't always work.
504: * Data-Relocatable Image Files:: are better.
505: * Fully Relocatable Image Files:: better yet.
506: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
507: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
508: * Modifying the Startup Sequence:: and turnkey applications.
509:
510: Fully Relocatable Image Files
511:
512: * gforthmi:: The normal way
513: * cross.fs:: The hard way
514:
515: Engine
516:
517: * Portability::
518: * Threading::
519: * Primitives::
520: * Performance::
521:
522: Threading
523:
524: * Scheduling::
525: * Direct or Indirect Threaded?::
526: * DOES>::
527:
528: Primitives
529:
530: * Automatic Generation::
531: * TOS Optimization::
532: * Produced code::
533:
534: Cross Compiler
535:
536: * Using the Cross Compiler::
537: * How the Cross Compiler Works::
538:
539: Other Forth-related information
540:
541: * Internet resources::
542: * Books::
543: * The Forth Interest Group::
544: * Conferences::
545:
546: @end detailmenu
547: @end menu
548:
549: @node License, Goals, Top, Top
550: @unnumbered GNU GENERAL PUBLIC LICENSE
551: @center Version 2, June 1991
552:
553: @display
554: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
555: 675 Mass Ave, Cambridge, MA 02139, USA
556:
557: Everyone is permitted to copy and distribute verbatim copies
558: of this license document, but changing it is not allowed.
559: @end display
560:
561: @unnumberedsec Preamble
562:
563: The licenses for most software are designed to take away your
564: freedom to share and change it. By contrast, the GNU General Public
565: License is intended to guarantee your freedom to share and change free
566: software---to make sure the software is free for all its users. This
567: General Public License applies to most of the Free Software
568: Foundation's software and to any other program whose authors commit to
569: using it. (Some other Free Software Foundation software is covered by
570: the GNU Library General Public License instead.) You can apply it to
571: your programs, too.
572:
573: When we speak of free software, we are referring to freedom, not
574: price. Our General Public Licenses are designed to make sure that you
575: have the freedom to distribute copies of free software (and charge for
576: this service if you wish), that you receive source code or can get it
577: if you want it, that you can change the software or use pieces of it
578: in new free programs; and that you know you can do these things.
579:
580: To protect your rights, we need to make restrictions that forbid
581: anyone to deny you these rights or to ask you to surrender the rights.
582: These restrictions translate to certain responsibilities for you if you
583: distribute copies of the software, or if you modify it.
584:
585: For example, if you distribute copies of such a program, whether
586: gratis or for a fee, you must give the recipients all the rights that
587: you have. You must make sure that they, too, receive or can get the
588: source code. And you must show them these terms so they know their
589: rights.
590:
591: We protect your rights with two steps: (1) copyright the software, and
592: (2) offer you this license which gives you legal permission to copy,
593: distribute and/or modify the software.
594:
595: Also, for each author's protection and ours, we want to make certain
596: that everyone understands that there is no warranty for this free
597: software. If the software is modified by someone else and passed on, we
598: want its recipients to know that what they have is not the original, so
599: that any problems introduced by others will not reflect on the original
600: authors' reputations.
601:
602: Finally, any free program is threatened constantly by software
603: patents. We wish to avoid the danger that redistributors of a free
604: program will individually obtain patent licenses, in effect making the
605: program proprietary. To prevent this, we have made it clear that any
606: patent must be licensed for everyone's free use or not licensed at all.
607:
608: The precise terms and conditions for copying, distribution and
609: modification follow.
610:
611: @iftex
612: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
613: @end iftex
614: @ifnottex
615: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
616: @end ifnottex
617:
618: @enumerate 0
619: @item
620: This License applies to any program or other work which contains
621: a notice placed by the copyright holder saying it may be distributed
622: under the terms of this General Public License. The ``Program'', below,
623: refers to any such program or work, and a ``work based on the Program''
624: means either the Program or any derivative work under copyright law:
625: that is to say, a work containing the Program or a portion of it,
626: either verbatim or with modifications and/or translated into another
627: language. (Hereinafter, translation is included without limitation in
628: the term ``modification''.) Each licensee is addressed as ``you''.
629:
630: Activities other than copying, distribution and modification are not
631: covered by this License; they are outside its scope. The act of
632: running the Program is not restricted, and the output from the Program
633: is covered only if its contents constitute a work based on the
634: Program (independent of having been made by running the Program).
635: Whether that is true depends on what the Program does.
636:
637: @item
638: You may copy and distribute verbatim copies of the Program's
639: source code as you receive it, in any medium, provided that you
640: conspicuously and appropriately publish on each copy an appropriate
641: copyright notice and disclaimer of warranty; keep intact all the
642: notices that refer to this License and to the absence of any warranty;
643: and give any other recipients of the Program a copy of this License
644: along with the Program.
645:
646: You may charge a fee for the physical act of transferring a copy, and
647: you may at your option offer warranty protection in exchange for a fee.
648:
649: @item
650: You may modify your copy or copies of the Program or any portion
651: of it, thus forming a work based on the Program, and copy and
652: distribute such modifications or work under the terms of Section 1
653: above, provided that you also meet all of these conditions:
654:
655: @enumerate a
656: @item
657: You must cause the modified files to carry prominent notices
658: stating that you changed the files and the date of any change.
659:
660: @item
661: You must cause any work that you distribute or publish, that in
662: whole or in part contains or is derived from the Program or any
663: part thereof, to be licensed as a whole at no charge to all third
664: parties under the terms of this License.
665:
666: @item
667: If the modified program normally reads commands interactively
668: when run, you must cause it, when started running for such
669: interactive use in the most ordinary way, to print or display an
670: announcement including an appropriate copyright notice and a
671: notice that there is no warranty (or else, saying that you provide
672: a warranty) and that users may redistribute the program under
673: these conditions, and telling the user how to view a copy of this
674: License. (Exception: if the Program itself is interactive but
675: does not normally print such an announcement, your work based on
676: the Program is not required to print an announcement.)
677: @end enumerate
678:
679: These requirements apply to the modified work as a whole. If
680: identifiable sections of that work are not derived from the Program,
681: and can be reasonably considered independent and separate works in
682: themselves, then this License, and its terms, do not apply to those
683: sections when you distribute them as separate works. But when you
684: distribute the same sections as part of a whole which is a work based
685: on the Program, the distribution of the whole must be on the terms of
686: this License, whose permissions for other licensees extend to the
687: entire whole, and thus to each and every part regardless of who wrote it.
688:
689: Thus, it is not the intent of this section to claim rights or contest
690: your rights to work written entirely by you; rather, the intent is to
691: exercise the right to control the distribution of derivative or
692: collective works based on the Program.
693:
694: In addition, mere aggregation of another work not based on the Program
695: with the Program (or with a work based on the Program) on a volume of
696: a storage or distribution medium does not bring the other work under
697: the scope of this License.
698:
699: @item
700: You may copy and distribute the Program (or a work based on it,
701: under Section 2) in object code or executable form under the terms of
702: Sections 1 and 2 above provided that you also do one of the following:
703:
704: @enumerate a
705: @item
706: Accompany it with the complete corresponding machine-readable
707: source code, which must be distributed under the terms of Sections
708: 1 and 2 above on a medium customarily used for software interchange; or,
709:
710: @item
711: Accompany it with a written offer, valid for at least three
712: years, to give any third party, for a charge no more than your
713: cost of physically performing source distribution, a complete
714: machine-readable copy of the corresponding source code, to be
715: distributed under the terms of Sections 1 and 2 above on a medium
716: customarily used for software interchange; or,
717:
718: @item
719: Accompany it with the information you received as to the offer
720: to distribute corresponding source code. (This alternative is
721: allowed only for noncommercial distribution and only if you
722: received the program in object code or executable form with such
723: an offer, in accord with Subsection b above.)
724: @end enumerate
725:
726: The source code for a work means the preferred form of the work for
727: making modifications to it. For an executable work, complete source
728: code means all the source code for all modules it contains, plus any
729: associated interface definition files, plus the scripts used to
730: control compilation and installation of the executable. However, as a
731: special exception, the source code distributed need not include
732: anything that is normally distributed (in either source or binary
733: form) with the major components (compiler, kernel, and so on) of the
734: operating system on which the executable runs, unless that component
735: itself accompanies the executable.
736:
737: If distribution of executable or object code is made by offering
738: access to copy from a designated place, then offering equivalent
739: access to copy the source code from the same place counts as
740: distribution of the source code, even though third parties are not
741: compelled to copy the source along with the object code.
742:
743: @item
744: You may not copy, modify, sublicense, or distribute the Program
745: except as expressly provided under this License. Any attempt
746: otherwise to copy, modify, sublicense or distribute the Program is
747: void, and will automatically terminate your rights under this License.
748: However, parties who have received copies, or rights, from you under
749: this License will not have their licenses terminated so long as such
750: parties remain in full compliance.
751:
752: @item
753: You are not required to accept this License, since you have not
754: signed it. However, nothing else grants you permission to modify or
755: distribute the Program or its derivative works. These actions are
756: prohibited by law if you do not accept this License. Therefore, by
757: modifying or distributing the Program (or any work based on the
758: Program), you indicate your acceptance of this License to do so, and
759: all its terms and conditions for copying, distributing or modifying
760: the Program or works based on it.
761:
762: @item
763: Each time you redistribute the Program (or any work based on the
764: Program), the recipient automatically receives a license from the
765: original licensor to copy, distribute or modify the Program subject to
766: these terms and conditions. You may not impose any further
767: restrictions on the recipients' exercise of the rights granted herein.
768: You are not responsible for enforcing compliance by third parties to
769: this License.
770:
771: @item
772: If, as a consequence of a court judgment or allegation of patent
773: infringement or for any other reason (not limited to patent issues),
774: conditions are imposed on you (whether by court order, agreement or
775: otherwise) that contradict the conditions of this License, they do not
776: excuse you from the conditions of this License. If you cannot
777: distribute so as to satisfy simultaneously your obligations under this
778: License and any other pertinent obligations, then as a consequence you
779: may not distribute the Program at all. For example, if a patent
780: license would not permit royalty-free redistribution of the Program by
781: all those who receive copies directly or indirectly through you, then
782: the only way you could satisfy both it and this License would be to
783: refrain entirely from distribution of the Program.
784:
785: If any portion of this section is held invalid or unenforceable under
786: any particular circumstance, the balance of the section is intended to
787: apply and the section as a whole is intended to apply in other
788: circumstances.
789:
790: It is not the purpose of this section to induce you to infringe any
791: patents or other property right claims or to contest validity of any
792: such claims; this section has the sole purpose of protecting the
793: integrity of the free software distribution system, which is
794: implemented by public license practices. Many people have made
795: generous contributions to the wide range of software distributed
796: through that system in reliance on consistent application of that
797: system; it is up to the author/donor to decide if he or she is willing
798: to distribute software through any other system and a licensee cannot
799: impose that choice.
800:
801: This section is intended to make thoroughly clear what is believed to
802: be a consequence of the rest of this License.
803:
804: @item
805: If the distribution and/or use of the Program is restricted in
806: certain countries either by patents or by copyrighted interfaces, the
807: original copyright holder who places the Program under this License
808: may add an explicit geographical distribution limitation excluding
809: those countries, so that distribution is permitted only in or among
810: countries not thus excluded. In such case, this License incorporates
811: the limitation as if written in the body of this License.
812:
813: @item
814: The Free Software Foundation may publish revised and/or new versions
815: of the General Public License from time to time. Such new versions will
816: be similar in spirit to the present version, but may differ in detail to
817: address new problems or concerns.
818:
819: Each version is given a distinguishing version number. If the Program
820: specifies a version number of this License which applies to it and ``any
821: later version'', you have the option of following the terms and conditions
822: either of that version or of any later version published by the Free
823: Software Foundation. If the Program does not specify a version number of
824: this License, you may choose any version ever published by the Free Software
825: Foundation.
826:
827: @item
828: If you wish to incorporate parts of the Program into other free
829: programs whose distribution conditions are different, write to the author
830: to ask for permission. For software which is copyrighted by the Free
831: Software Foundation, write to the Free Software Foundation; we sometimes
832: make exceptions for this. Our decision will be guided by the two goals
833: of preserving the free status of all derivatives of our free software and
834: of promoting the sharing and reuse of software generally.
835:
836: @iftex
837: @heading NO WARRANTY
838: @end iftex
839: @ifnottex
840: @center NO WARRANTY
841: @end ifnottex
842:
843: @item
844: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
845: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
846: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
847: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
848: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
849: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
850: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
851: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
852: REPAIR OR CORRECTION.
853:
854: @item
855: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
856: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
857: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
858: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
859: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
860: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
861: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
862: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
863: POSSIBILITY OF SUCH DAMAGES.
864: @end enumerate
865:
866: @iftex
867: @heading END OF TERMS AND CONDITIONS
868: @end iftex
869: @ifnottex
870: @center END OF TERMS AND CONDITIONS
871: @end ifnottex
872:
873: @page
874: @unnumberedsec How to Apply These Terms to Your New Programs
875:
876: If you develop a new program, and you want it to be of the greatest
877: possible use to the public, the best way to achieve this is to make it
878: free software which everyone can redistribute and change under these terms.
879:
880: To do so, attach the following notices to the program. It is safest
881: to attach them to the start of each source file to most effectively
882: convey the exclusion of warranty; and each file should have at least
883: the ``copyright'' line and a pointer to where the full notice is found.
884:
885: @smallexample
886: @var{one line to give the program's name and a brief idea of what it does.}
887: Copyright (C) 19@var{yy} @var{name of author}
888:
889: This program is free software; you can redistribute it and/or modify
890: it under the terms of the GNU General Public License as published by
891: the Free Software Foundation; either version 2 of the License, or
892: (at your option) any later version.
893:
894: This program is distributed in the hope that it will be useful,
895: but WITHOUT ANY WARRANTY; without even the implied warranty of
896: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
897: GNU General Public License for more details.
898:
899: You should have received a copy of the GNU General Public License
900: along with this program; if not, write to the Free Software
901: Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
902: @end smallexample
903:
904: Also add information on how to contact you by electronic and paper mail.
905:
906: If the program is interactive, make it output a short notice like this
907: when it starts in an interactive mode:
908:
909: @smallexample
910: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
911: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
912: type `show w'.
913: This is free software, and you are welcome to redistribute it
914: under certain conditions; type `show c' for details.
915: @end smallexample
916:
917: The hypothetical commands @samp{show w} and @samp{show c} should show
918: the appropriate parts of the General Public License. Of course, the
919: commands you use may be called something other than @samp{show w} and
920: @samp{show c}; they could even be mouse-clicks or menu items---whatever
921: suits your program.
922:
923: You should also get your employer (if you work as a programmer) or your
924: school, if any, to sign a ``copyright disclaimer'' for the program, if
925: necessary. Here is a sample; alter the names:
926:
927: @smallexample
928: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
929: `Gnomovision' (which makes passes at compilers) written by James Hacker.
930:
931: @var{signature of Ty Coon}, 1 April 1989
932: Ty Coon, President of Vice
933: @end smallexample
934:
935: This General Public License does not permit incorporating your program into
936: proprietary programs. If your program is a subroutine library, you may
937: consider it more useful to permit linking proprietary applications with the
938: library. If this is what you want to do, use the GNU Library General
939: Public License instead of this License.
940:
941: @iftex
942: @unnumbered Preface
943: @cindex Preface
944: This manual documents Gforth. Some introductory material is provided for
945: readers who are unfamiliar with Forth or who are migrating to Gforth
946: from other Forth compilers. However, this manual is primarily a
947: reference manual.
948: @end iftex
949:
950: @comment TODO much more blurb here.
951:
952: @c ******************************************************************
953: @node Goals, Gforth Environment, License, Top
954: @comment node-name, next, previous, up
955: @chapter Goals of Gforth
956: @cindex goals of the Gforth project
957: The goal of the Gforth Project is to develop a standard model for
958: ANS Forth. This can be split into several subgoals:
959:
960: @itemize @bullet
961: @item
962: Gforth should conform to the ANS Forth Standard.
963: @item
964: It should be a model, i.e. it should define all the
965: implementation-dependent things.
966: @item
967: It should become standard, i.e. widely accepted and used. This goal
968: is the most difficult one.
969: @end itemize
970:
971: To achieve these goals Gforth should be
972: @itemize @bullet
973: @item
974: Similar to previous models (fig-Forth, F83)
975: @item
976: Powerful. It should provide for all the things that are considered
977: necessary today and even some that are not yet considered necessary.
978: @item
979: Efficient. It should not get the reputation of being exceptionally
980: slow.
981: @item
982: Free.
983: @item
984: Available on many machines/easy to port.
985: @end itemize
986:
987: Have we achieved these goals? Gforth conforms to the ANS Forth
988: standard. It may be considered a model, but we have not yet documented
989: which parts of the model are stable and which parts we are likely to
990: change. It certainly has not yet become a de facto standard, but it
991: appears to be quite popular. It has some similarities to and some
992: differences from previous models. It has some powerful features, but not
993: yet everything that we envisioned. We certainly have achieved our
994: execution speed goals (@pxref{Performance})@footnote{However, in 1998
995: the bar was raised when the major commercial Forth vendors switched to
996: native code compilers.}. It is free and available on many machines.
997:
998: @c ******************************************************************
999: @node Gforth Environment, Tutorial, Goals, Top
1000: @chapter Gforth Environment
1001: @cindex Gforth environment
1002:
1003: Note: ultimately, the Gforth man page will be auto-generated from the
1004: material in this chapter.
1005:
1006: @menu
1007: * Invoking Gforth:: Getting in
1008: * Leaving Gforth:: Getting out
1009: * Command-line editing::
1010: * Environment variables:: that affect how Gforth starts up
1011: * Gforth Files:: What gets installed and where
1012: * Startup speed:: When 35ms is not fast enough ...
1013: @end menu
1014:
1015: For related information about the creation of images see @ref{Image Files}.
1016:
1017: @comment ----------------------------------------------
1018: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1019: @section Invoking Gforth
1020: @cindex invoking Gforth
1021: @cindex running Gforth
1022: @cindex command-line options
1023: @cindex options on the command line
1024: @cindex flags on the command line
1025:
1026: Gforth is made up of two parts; an executable ``engine'' (named
1027: @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
1028: will usually just say @code{gforth} -- this automatically loads the
1029: default image file @file{gforth.fi}. In many other cases the default
1030: Gforth image will be invoked like this:
1031: @example
1032: gforth [file | -e forth-code] ...
1033: @end example
1034: @noindent
1035: This interprets the contents of the files and the Forth code in the order they
1036: are given.
1037:
1038: In addition to the @file{gforth} engine, there is also an engine called
1039: @file{gforth-fast}, which is faster, but gives less informative error
1040: messages (@pxref{Error messages}).
1041:
1042: In general, the command line looks like this:
1043:
1044: @example
1045: gforth[-fast] [engine options] [image options]
1046: @end example
1047:
1048: The engine options must come before the rest of the command
1049: line. They are:
1050:
1051: @table @code
1052: @cindex -i, command-line option
1053: @cindex --image-file, command-line option
1054: @item --image-file @i{file}
1055: @itemx -i @i{file}
1056: Loads the Forth image @i{file} instead of the default
1057: @file{gforth.fi} (@pxref{Image Files}).
1058:
1059: @cindex --appl-image, command-line option
1060: @item --appl-image @i{file}
1061: Loads the image @i{file} and leaves all further command-line arguments
1062: to the image (instead of processing them as engine options). This is
1063: useful for building executable application images on Unix, built with
1064: @code{gforthmi --application ...}.
1065:
1066: @cindex --path, command-line option
1067: @cindex -p, command-line option
1068: @item --path @i{path}
1069: @itemx -p @i{path}
1070: Uses @i{path} for searching the image file and Forth source code files
1071: instead of the default in the environment variable @code{GFORTHPATH} or
1072: the path specified at installation time (e.g.,
1073: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1074: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1075:
1076: @cindex --dictionary-size, command-line option
1077: @cindex -m, command-line option
1078: @cindex @i{size} parameters for command-line options
1079: @cindex size of the dictionary and the stacks
1080: @item --dictionary-size @i{size}
1081: @itemx -m @i{size}
1082: Allocate @i{size} space for the Forth dictionary space instead of
1083: using the default specified in the image (typically 256K). The
1084: @i{size} specification for this and subsequent options consists of
1085: an integer and a unit (e.g.,
1086: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1087: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1088: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1089: @code{e} is used.
1090:
1091: @cindex --data-stack-size, command-line option
1092: @cindex -d, command-line option
1093: @item --data-stack-size @i{size}
1094: @itemx -d @i{size}
1095: Allocate @i{size} space for the data stack instead of using the
1096: default specified in the image (typically 16K).
1097:
1098: @cindex --return-stack-size, command-line option
1099: @cindex -r, command-line option
1100: @item --return-stack-size @i{size}
1101: @itemx -r @i{size}
1102: Allocate @i{size} space for the return stack instead of using the
1103: default specified in the image (typically 15K).
1104:
1105: @cindex --fp-stack-size, command-line option
1106: @cindex -f, command-line option
1107: @item --fp-stack-size @i{size}
1108: @itemx -f @i{size}
1109: Allocate @i{size} space for the floating point stack instead of
1110: using the default specified in the image (typically 15.5K). In this case
1111: the unit specifier @code{e} refers to floating point numbers.
1112:
1113: @cindex --locals-stack-size, command-line option
1114: @cindex -l, command-line option
1115: @item --locals-stack-size @i{size}
1116: @itemx -l @i{size}
1117: Allocate @i{size} space for the locals stack instead of using the
1118: default specified in the image (typically 14.5K).
1119:
1120: @cindex -h, command-line option
1121: @cindex --help, command-line option
1122: @item --help
1123: @itemx -h
1124: Print a message about the command-line options
1125:
1126: @cindex -v, command-line option
1127: @cindex --version, command-line option
1128: @item --version
1129: @itemx -v
1130: Print version and exit
1131:
1132: @cindex --debug, command-line option
1133: @item --debug
1134: Print some information useful for debugging on startup.
1135:
1136: @cindex --offset-image, command-line option
1137: @item --offset-image
1138: Start the dictionary at a slightly different position than would be used
1139: otherwise (useful for creating data-relocatable images,
1140: @pxref{Data-Relocatable Image Files}).
1141:
1142: @cindex --no-offset-im, command-line option
1143: @item --no-offset-im
1144: Start the dictionary at the normal position.
1145:
1146: @cindex --clear-dictionary, command-line option
1147: @item --clear-dictionary
1148: Initialize all bytes in the dictionary to 0 before loading the image
1149: (@pxref{Data-Relocatable Image Files}).
1150:
1151: @cindex --die-on-signal, command-line-option
1152: @item --die-on-signal
1153: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1154: or the segmentation violation SIGSEGV) by translating it into a Forth
1155: @code{THROW}. With this option, Gforth exits if it receives such a
1156: signal. This option is useful when the engine and/or the image might be
1157: severely broken (such that it causes another signal before recovering
1158: from the first); this option avoids endless loops in such cases.
1159: @end table
1160:
1161: @cindex loading files at startup
1162: @cindex executing code on startup
1163: @cindex batch processing with Gforth
1164: As explained above, the image-specific command-line arguments for the
1165: default image @file{gforth.fi} consist of a sequence of filenames and
1166: @code{-e @var{forth-code}} options that are interpreted in the sequence
1167: in which they are given. The @code{-e @var{forth-code}} or
1168: @code{--evaluate @var{forth-code}} option evaluates the Forth
1169: code. This option takes only one argument; if you want to evaluate more
1170: Forth words, you have to quote them or use @code{-e} several times. To exit
1171: after processing the command line (instead of entering interactive mode)
1172: append @code{-e bye} to the command line.
1173:
1174: @cindex versions, invoking other versions of Gforth
1175: If you have several versions of Gforth installed, @code{gforth} will
1176: invoke the version that was installed last. @code{gforth-@i{version}}
1177: invokes a specific version. If your environment contains the variable
1178: @code{GFORTHPATH}, you may want to override it by using the
1179: @code{--path} option.
1180:
1181: Not yet implemented:
1182: On startup the system first executes the system initialization file
1183: (unless the option @code{--no-init-file} is given; note that the system
1184: resulting from using this option may not be ANS Forth conformant). Then
1185: the user initialization file @file{.gforth.fs} is executed, unless the
1186: option @code{--no-rc} is given; this file is searched for in @file{.},
1187: then in @file{~}, then in the normal path (see above).
1188:
1189:
1190:
1191: @comment ----------------------------------------------
1192: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1193: @section Leaving Gforth
1194: @cindex Gforth - leaving
1195: @cindex leaving Gforth
1196:
1197: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1198: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1199: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1200: data are discarded. For ways of saving the state of the system before
1201: leaving Gforth see @ref{Image Files}.
1202:
1203: doc-bye
1204:
1205:
1206: @comment ----------------------------------------------
1207: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1208: @section Command-line editing
1209: @cindex command-line editing
1210:
1211: Gforth maintains a history file that records every line that you type to
1212: the text interpreter. This file is preserved between sessions, and is
1213: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1214: repeatedly you can recall successively older commands from this (or
1215: previous) session(s). The full list of command-line editing facilities is:
1216:
1217: @itemize @bullet
1218: @item
1219: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1220: commands from the history buffer.
1221: @item
1222: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1223: from the history buffer.
1224: @item
1225: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1226: @item
1227: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1228: @item
1229: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1230: closing up the line.
1231: @item
1232: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1233: @item
1234: @kbd{Ctrl-a} to move the cursor to the start of the line.
1235: @item
1236: @kbd{Ctrl-e} to move the cursor to the end of the line.
1237: @item
1238: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1239: line.
1240: @item
1241: @key{TAB} to step through all possible full-word completions of the word
1242: currently being typed.
1243: @item
1244: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
1245: using @code{bye}).
1246: @item
1247: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
1248: character under the cursor.
1249: @end itemize
1250:
1251: When editing, displayable characters are inserted to the left of the
1252: cursor position; the line is always in ``insert'' (as opposed to
1253: ``overstrike'') mode.
1254:
1255: @cindex history file
1256: @cindex @file{.gforth-history}
1257: On Unix systems, the history file is @file{~/.gforth-history} by
1258: default@footnote{i.e. it is stored in the user's home directory.}. You
1259: can find out the name and location of your history file using:
1260:
1261: @example
1262: history-file type \ Unix-class systems
1263:
1264: history-file type \ Other systems
1265: history-dir type
1266: @end example
1267:
1268: If you enter long definitions by hand, you can use a text editor to
1269: paste them out of the history file into a Forth source file for reuse at
1270: a later time.
1271:
1272: Gforth never trims the size of the history file, so you should do this
1273: periodically, if necessary.
1274:
1275: @comment this is all defined in history.fs
1276: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1277: @comment chosen?
1278:
1279:
1280: @comment ----------------------------------------------
1281: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1282: @section Environment variables
1283: @cindex environment variables
1284:
1285: Gforth uses these environment variables:
1286:
1287: @itemize @bullet
1288: @item
1289: @cindex @code{GFORTHHIST} -- environment variable
1290: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1291: open/create the history file, @file{.gforth-history}. Default:
1292: @code{$HOME}.
1293:
1294: @item
1295: @cindex @code{GFORTHPATH} -- environment variable
1296: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1297: for Forth source-code files.
1298:
1299: @item
1300: @cindex @code{GFORTH} -- environment variable
1301: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1302:
1303: @item
1304: @cindex @code{GFORTHD} -- environment variable
1305: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1306:
1307: @item
1308: @cindex @code{TMP}, @code{TEMP} - environment variable
1309: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1310: location for the history file.
1311: @end itemize
1312:
1313: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1314: @comment mentioning these.
1315:
1316: All the Gforth environment variables default to sensible values if they
1317: are not set.
1318:
1319:
1320: @comment ----------------------------------------------
1321: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1322: @section Gforth files
1323: @cindex Gforth files
1324:
1325: When you install Gforth on a Unix system, it installs files in these
1326: locations by default:
1327:
1328: @itemize @bullet
1329: @item
1330: @file{/usr/local/bin/gforth}
1331: @item
1332: @file{/usr/local/bin/gforthmi}
1333: @item
1334: @file{/usr/local/man/man1/gforth.1} - man page.
1335: @item
1336: @file{/usr/local/info} - the Info version of this manual.
1337: @item
1338: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1339: @item
1340: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1341: @item
1342: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1343: @item
1344: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1345: @end itemize
1346:
1347: You can select different places for installation by using
1348: @code{configure} options (listed with @code{configure --help}).
1349:
1350: @comment ----------------------------------------------
1351: @node Startup speed, , Gforth Files, Gforth Environment
1352: @section Startup speed
1353: @cindex Startup speed
1354: @cindex speed, startup
1355:
1356: If Gforth is used for CGI scripts or in shell scripts, its startup
1357: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1358: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1359: system time.
1360:
1361: If startup speed is a problem, you may consider the following ways to
1362: improve it; or you may consider ways to reduce the number of startups
1363: (for example, by using Fast-CGI).
1364:
1365: The first step to improve startup speed is to statically link Gforth, by
1366: building it with @code{XLDFLAGS=-static}. This requires more memory for
1367: the code and will therefore slow down the first invocation, but
1368: subsequent invocations avoid the dynamic linking overhead. Another
1369: disadvantage is that Gforth won't profit from library upgrades. As a
1370: result, @code{gforth-static -e bye} takes about 17.1ms user and
1371: 8.2ms system time.
1372:
1373: The next step to improve startup speed is to use a non-relocatable image
1374: (@pxref{Non-Relocatable Image Files}). You can create this image with
1375: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1376: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1377: and a part of the copy-on-write overhead. The disadvantage is that the
1378: non-relocatable image does not work if the OS gives Gforth a different
1379: address for the dictionary, for whatever reason; so you better provide a
1380: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1381: bye} takes about 15.3ms user and 7.5ms system time.
1382:
1383: The final step is to disable dictionary hashing in Gforth. Gforth
1384: builds the hash table on startup, which takes much of the startup
1385: overhead. You can do this by commenting out the @code{include hash.fs}
1386: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1387: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1388: The disadvantages are that functionality like @code{table} and
1389: @code{ekey} is missing and that text interpretation (e.g., compiling)
1390: now takes much longer. So, you should only use this method if there is
1391: no significant text interpretation to perform (the script should be
1392: compiled into the image, amongst other things). @code{gforth-static -i
1393: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1394:
1395: @c ******************************************************************
1396: @node Tutorial, Introduction, Gforth Environment, Top
1397: @chapter Forth Tutorial
1398: @cindex Tutorial
1399: @cindex Forth Tutorial
1400:
1401: This tutorial can be used with any ANS-compliant Forth; any
1402: Gforth-specific features are marked as such and you can skip them if you
1403: work with another Forth. This tutorial does not explain all features of
1404: Forth, just enough to get you started and give you some ideas about the
1405: facilities available in Forth. Read the rest of the manual and the
1406: standard when you are through this.
1407:
1408: The intended way to use this tutorial is that you work through it while
1409: sitting in front of the console, take a look at the examples and predict
1410: what they will do, then try them out; if the outcome is not as expected,
1411: find out why (e.g., by trying out variations of the example), so you
1412: understand what's going on. There are also some assignments that you
1413: should solve.
1414:
1415: This tutorial assumes that you have programmed before and know what,
1416: e.g., a loop is.
1417:
1418: @c !! explain compat library
1419:
1420: @menu
1421: * Starting Gforth Tutorial::
1422: * Syntax Tutorial::
1423: * Crash Course Tutorial::
1424: * Stack Tutorial::
1425: * Arithmetics Tutorial::
1426: * Stack Manipulation Tutorial::
1427: * Using files for Forth code Tutorial::
1428: * Comments Tutorial::
1429: * Colon Definitions Tutorial::
1430: * Decompilation Tutorial::
1431: * Stack-Effect Comments Tutorial::
1432: * Types Tutorial::
1433: * Factoring Tutorial::
1434: * Designing the stack effect Tutorial::
1435: * Local Variables Tutorial::
1436: * Conditional execution Tutorial::
1437: * Flags and Comparisons Tutorial::
1438: * General Loops Tutorial::
1439: * Counted loops Tutorial::
1440: * Recursion Tutorial::
1441: * Leaving definitions or loops Tutorial::
1442: * Return Stack Tutorial::
1443: * Memory Tutorial::
1444: * Characters and Strings Tutorial::
1445: * Alignment Tutorial::
1446: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1447: * Execution Tokens Tutorial::
1448: * Exceptions Tutorial::
1449: * Defining Words Tutorial::
1450: * Arrays and Records Tutorial::
1451: * POSTPONE Tutorial::
1452: * Literal Tutorial::
1453: * Advanced macros Tutorial::
1454: * Compilation Tokens Tutorial::
1455: * Wordlists and Search Order Tutorial::
1456: @end menu
1457:
1458: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1459: @section Starting Gforth
1460: @cindex starting Gforth tutorial
1461: You can start Gforth by typing its name:
1462:
1463: @example
1464: gforth
1465: @end example
1466:
1467: That puts you into interactive mode; you can leave Gforth by typing
1468: @code{bye}. While in Gforth, you can edit the command line and access
1469: the command line history with cursor keys, similar to bash.
1470:
1471:
1472: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1473: @section Syntax
1474: @cindex syntax tutorial
1475:
1476: A @dfn{word} is a sequence of arbitrary characters (expcept white
1477: space). Words are separated by white space. E.g., each of the
1478: following lines contains exactly one word:
1479:
1480: @example
1481: word
1482: !@@#$%^&*()
1483: 1234567890
1484: 5!a
1485: @end example
1486:
1487: A frequent beginner's error is to leave away necessary white space,
1488: resulting in an error like @samp{Undefined word}; so if you see such an
1489: error, check if you have put spaces wherever necessary.
1490:
1491: @example
1492: ." hello, world" \ correct
1493: ."hello, world" \ gives an "Undefined word" error
1494: @end example
1495:
1496: Gforth and most other Forth systems ignore differences in case (they are
1497: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1498: your system is case-sensitive, you may have to type all the examples
1499: given here in upper case.
1500:
1501:
1502: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1503: @section Crash Course
1504:
1505: Type
1506:
1507: @example
1508: 0 0 !
1509: here execute
1510: ' catch >body 20 erase abort
1511: ' (quit) >body 20 erase
1512: @end example
1513:
1514: The last two examples are guaranteed to destroy parts of Gforth (and
1515: most other systems), so you better leave Gforth afterwards (if it has
1516: not finished by itself). On some systems you may have to kill gforth
1517: from outside (e.g., in Unix with @code{kill}).
1518:
1519: Now that you know how to produce crashes (and that there's not much to
1520: them), let's learn how to produce meaningful programs.
1521:
1522:
1523: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1524: @section Stack
1525: @cindex stack tutorial
1526:
1527: The most obvious feature of Forth is the stack. When you type in a
1528: number, it is pushed on the stack. You can display the content of the
1529: stack with @code{.s}.
1530:
1531: @example
1532: 1 2 .s
1533: 3 .s
1534: @end example
1535:
1536: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1537: appear in @code{.s} output as they appeared in the input.
1538:
1539: You can print the top of stack element with @code{.}.
1540:
1541: @example
1542: 1 2 3 . . .
1543: @end example
1544:
1545: In general, words consume their stack arguments (@code{.s} is an
1546: exception).
1547:
1548: @assignment
1549: What does the stack contain after @code{5 6 7 .}?
1550: @endassignment
1551:
1552:
1553: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1554: @section Arithmetics
1555: @cindex arithmetics tutorial
1556:
1557: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1558: operate on the top two stack items:
1559:
1560: @example
1561: 2 2 + .
1562: 2 1 - .
1563: 7 3 mod .
1564: @end example
1565:
1566: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1567: as in the corresponding infix expression (this is generally the case in
1568: Forth).
1569:
1570: Parentheses are superfluous (and not available), because the order of
1571: the words unambiguously determines the order of evaluation and the
1572: operands:
1573:
1574: @example
1575: 3 4 + 5 * .
1576: 3 4 5 * + .
1577: @end example
1578:
1579: @assignment
1580: What are the infix expressions corresponding to the Forth code above?
1581: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1582: known as Postfix or RPN (Reverse Polish Notation).}.
1583: @endassignment
1584:
1585: To change the sign, use @code{negate}:
1586:
1587: @example
1588: 2 negate .
1589: @end example
1590:
1591: @assignment
1592: Convert -(-3)*4-5 to Forth.
1593: @endassignment
1594:
1595: @code{/mod} performs both @code{/} and @code{mod}.
1596:
1597: @example
1598: 7 3 /mod . .
1599: @end example
1600:
1601: Reference: @ref{Arithmetic}.
1602:
1603:
1604: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1605: @section Stack Manipulation
1606: @cindex stack manipulation tutorial
1607:
1608: Stack manipulation words rearrange the data on the stack.
1609:
1610: @example
1611: 1 .s drop .s
1612: 1 .s dup .s drop drop .s
1613: 1 2 .s over .s drop drop drop
1614: 1 2 .s swap .s drop drop
1615: 1 2 3 .s rot .s drop drop drop
1616: @end example
1617:
1618: These are the most important stack manipulation words. There are also
1619: variants that manipulate twice as many stack items:
1620:
1621: @example
1622: 1 2 3 4 .s 2swap .s 2drop 2drop
1623: @end example
1624:
1625: Two more stack manipulation words are:
1626:
1627: @example
1628: 1 2 .s nip .s drop
1629: 1 2 .s tuck .s 2drop drop
1630: @end example
1631:
1632: @assignment
1633: Replace @code{nip} and @code{tuck} with combinations of other stack
1634: manipulation words.
1635:
1636: @example
1637: Given: How do you get:
1638: 1 2 3 3 2 1
1639: 1 2 3 1 2 3 2
1640: 1 2 3 1 2 3 3
1641: 1 2 3 1 3 3
1642: 1 2 3 2 1 3
1643: 1 2 3 4 4 3 2 1
1644: 1 2 3 1 2 3 1 2 3
1645: 1 2 3 4 1 2 3 4 1 2
1646: 1 2 3
1647: 1 2 3 1 2 3 4
1648: 1 2 3 1 3
1649: @end example
1650: @endassignment
1651:
1652: @example
1653: 5 dup * .
1654: @end example
1655:
1656: @assignment
1657: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1658: Write a piece of Forth code that expects two numbers on the stack
1659: (@var{a} and @var{b}, with @var{b} on top) and computes
1660: @code{(a-b)(a+1)}.
1661: @endassignment
1662:
1663: Reference: @ref{Stack Manipulation}.
1664:
1665:
1666: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1667: @section Using files for Forth code
1668: @cindex loading Forth code, tutorial
1669: @cindex files containing Forth code, tutorial
1670:
1671: While working at the Forth command line is convenient for one-line
1672: examples and short one-off code, you probably want to store your source
1673: code in files for convenient editing and persistence. You can use your
1674: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1675: Gforth}) to create @var{file} and use
1676:
1677: @example
1678: s" @var{file}" included
1679: @end example
1680:
1681: to load it into your Forth system. The file name extension I use for
1682: Forth files is @samp{.fs}.
1683:
1684: You can easily start Gforth with some files loaded like this:
1685:
1686: @example
1687: gforth @var{file1} @var{file2}
1688: @end example
1689:
1690: If an error occurs during loading these files, Gforth terminates,
1691: whereas an error during @code{INCLUDED} within Gforth usually gives you
1692: a Gforth command line. Starting the Forth system every time gives you a
1693: clean start every time, without interference from the results of earlier
1694: tries.
1695:
1696: I often put all the tests in a file, then load the code and run the
1697: tests with
1698:
1699: @example
1700: gforth @var{code} @var{tests} -e bye
1701: @end example
1702:
1703: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1704: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1705: restart this command without ado.
1706:
1707: The advantage of this approach is that the tests can be repeated easily
1708: every time the program ist changed, making it easy to catch bugs
1709: introduced by the change.
1710:
1711: Reference: @ref{Forth source files}.
1712:
1713:
1714: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1715: @section Comments
1716: @cindex comments tutorial
1717:
1718: @example
1719: \ That's a comment; it ends at the end of the line
1720: ( Another comment; it ends here: ) .s
1721: @end example
1722:
1723: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1724: separated with white space from the following text.
1725:
1726: @example
1727: \This gives an "Undefined word" error
1728: @end example
1729:
1730: The first @code{)} ends a comment started with @code{(}, so you cannot
1731: nest @code{(}-comments; and you cannot comment out text containing a
1732: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1733: avoid @code{)} in word names.}.
1734:
1735: I use @code{\}-comments for descriptive text and for commenting out code
1736: of one or more line; I use @code{(}-comments for describing the stack
1737: effect, the stack contents, or for commenting out sub-line pieces of
1738: code.
1739:
1740: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1741: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1742: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1743: with @kbd{M-q}.
1744:
1745: Reference: @ref{Comments}.
1746:
1747:
1748: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1749: @section Colon Definitions
1750: @cindex colon definitions, tutorial
1751: @cindex definitions, tutorial
1752: @cindex procedures, tutorial
1753: @cindex functions, tutorial
1754:
1755: are similar to procedures and functions in other programming languages.
1756:
1757: @example
1758: : squared ( n -- n^2 )
1759: dup * ;
1760: 5 squared .
1761: 7 squared .
1762: @end example
1763:
1764: @code{:} starts the colon definition; its name is @code{squared}. The
1765: following comment describes its stack effect. The words @code{dup *}
1766: are not executed, but compiled into the definition. @code{;} ends the
1767: colon definition.
1768:
1769: The newly-defined word can be used like any other word, including using
1770: it in other definitions:
1771:
1772: @example
1773: : cubed ( n -- n^3 )
1774: dup squared * ;
1775: -5 cubed .
1776: : fourth-power ( n -- n^4 )
1777: squared squared ;
1778: 3 fourth-power .
1779: @end example
1780:
1781: @assignment
1782: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1783: @code{/mod} in terms of other Forth words, and check if they work (hint:
1784: test your tests on the originals first). Don't let the
1785: @samp{redefined}-Messages spook you, they are just warnings.
1786: @endassignment
1787:
1788: Reference: @ref{Colon Definitions}.
1789:
1790:
1791: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1792: @section Decompilation
1793: @cindex decompilation tutorial
1794: @cindex see tutorial
1795:
1796: You can decompile colon definitions with @code{see}:
1797:
1798: @example
1799: see squared
1800: see cubed
1801: @end example
1802:
1803: In Gforth @code{see} shows you a reconstruction of the source code from
1804: the executable code. Informations that were present in the source, but
1805: not in the executable code, are lost (e.g., comments).
1806:
1807: You can also decompile the predefined words:
1808:
1809: @example
1810: see .
1811: see +
1812: @end example
1813:
1814:
1815: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1816: @section Stack-Effect Comments
1817: @cindex stack-effect comments, tutorial
1818: @cindex --, tutorial
1819: By convention the comment after the name of a definition describes the
1820: stack effect: The part in from of the @samp{--} describes the state of
1821: the stack before the execution of the definition, i.e., the parameters
1822: that are passed into the colon definition; the part behind the @samp{--}
1823: is the state of the stack after the execution of the definition, i.e.,
1824: the results of the definition. The stack comment only shows the top
1825: stack items that the definition accesses and/or changes.
1826:
1827: You should put a correct stack effect on every definition, even if it is
1828: just @code{( -- )}. You should also add some descriptive comment to
1829: more complicated words (I usually do this in the lines following
1830: @code{:}). If you don't do this, your code becomes unreadable (because
1831: you have to work through every definition before you can undertsand
1832: any).
1833:
1834: @assignment
1835: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1836: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1837: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1838: are done, you can compare your stack effects to those in this manual
1839: (@pxref{Word Index}).
1840: @endassignment
1841:
1842: Sometimes programmers put comments at various places in colon
1843: definitions that describe the contents of the stack at that place (stack
1844: comments); i.e., they are like the first part of a stack-effect
1845: comment. E.g.,
1846:
1847: @example
1848: : cubed ( n -- n^3 )
1849: dup squared ( n n^2 ) * ;
1850: @end example
1851:
1852: In this case the stack comment is pretty superfluous, because the word
1853: is simple enough. If you think it would be a good idea to add such a
1854: comment to increase readability, you should also consider factoring the
1855: word into several simpler words (@pxref{Factoring Tutorial,,
1856: Factoring}), which typically eliminates the need for the stack comment;
1857: however, if you decide not to refactor it, then having such a comment is
1858: better than not having it.
1859:
1860: The names of the stack items in stack-effect and stack comments in the
1861: standard, in this manual, and in many programs specify the type through
1862: a type prefix, similar to Fortran and Hungarian notation. The most
1863: frequent prefixes are:
1864:
1865: @table @code
1866: @item n
1867: signed integer
1868: @item u
1869: unsigned integer
1870: @item c
1871: character
1872: @item f
1873: Boolean flags, i.e. @code{false} or @code{true}.
1874: @item a-addr,a-
1875: Cell-aligned address
1876: @item c-addr,c-
1877: Char-aligned address (note that a Char may have two bytes in Windows NT)
1878: @item xt
1879: Execution token, same size as Cell
1880: @item w,x
1881: Cell, can contain an integer or an address. It usually takes 32, 64 or
1882: 16 bits (depending on your platform and Forth system). A cell is more
1883: commonly known as machine word, but the term @emph{word} already means
1884: something different in Forth.
1885: @item d
1886: signed double-cell integer
1887: @item ud
1888: unsigned double-cell integer
1889: @item r
1890: Float (on the FP stack)
1891: @end table
1892:
1893: You can find a more complete list in @ref{Notation}.
1894:
1895: @assignment
1896: Write stack-effect comments for all definitions you have written up to
1897: now.
1898: @endassignment
1899:
1900:
1901: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1902: @section Types
1903: @cindex types tutorial
1904:
1905: In Forth the names of the operations are not overloaded; so similar
1906: operations on different types need different names; e.g., @code{+} adds
1907: integers, and you have to use @code{f+} to add floating-point numbers.
1908: The following prefixes are often used for related operations on
1909: different types:
1910:
1911: @table @code
1912: @item (none)
1913: signed integer
1914: @item u
1915: unsigned integer
1916: @item c
1917: character
1918: @item d
1919: signed double-cell integer
1920: @item ud, du
1921: unsigned double-cell integer
1922: @item 2
1923: two cells (not-necessarily double-cell numbers)
1924: @item m, um
1925: mixed single-cell and double-cell operations
1926: @item f
1927: floating-point (note that in stack comments @samp{f} represents flags,
1928: and @samp{r} represents FP numbers).
1929: @end table
1930:
1931: If there are no differences between the signed and the unsigned variant
1932: (e.g., for @code{+}), there is only the prefix-less variant.
1933:
1934: Forth does not perform type checking, neither at compile time, nor at
1935: run time. If you use the wrong oeration, the data are interpreted
1936: incorrectly:
1937:
1938: @example
1939: -1 u.
1940: @end example
1941:
1942: If you have only experience with type-checked languages until now, and
1943: have heard how important type-checking is, don't panic! In my
1944: experience (and that of other Forthers), type errors in Forth code are
1945: usually easy to find (once you get used to it), the increased vigilance
1946: of the programmer tends to catch some harder errors in addition to most
1947: type errors, and you never have to work around the type system, so in
1948: most situations the lack of type-checking seems to be a win (projects to
1949: add type checking to Forth have not caught on).
1950:
1951:
1952: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1953: @section Factoring
1954: @cindex factoring tutorial
1955:
1956: If you try to write longer definitions, you will soon find it hard to
1957: keep track of the stack contents. Therefore, good Forth programmers
1958: tend to write only short definitions (e.g., three lines). The art of
1959: finding meaningful short definitions is known as factoring (as in
1960: factoring polynomials).
1961:
1962: Well-factored programs offer additional advantages: smaller, more
1963: general words, are easier to test and debug and can be reused more and
1964: better than larger, specialized words.
1965:
1966: So, if you run into difficulties with stack management, when writing
1967: code, try to define meaningful factors for the word, and define the word
1968: in terms of those. Even if a factor contains only two words, it is
1969: often helpful.
1970:
1971: Good factoring is not easy, and it takes some practice to get the knack
1972: for it; but even experienced Forth programmers often don't find the
1973: right solution right away, but only when rewriting the program. So, if
1974: you don't come up with a good solution immediately, keep trying, don't
1975: despair.
1976:
1977: @c example !!
1978:
1979:
1980: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
1981: @section Designing the stack effect
1982: @cindex Stack effect design, tutorial
1983: @cindex design of stack effects, tutorial
1984:
1985: In other languages you can use an arbitrary order of parameters for a
1986: function; and since there is only one result, you don't have to deal with
1987: the order of results, either.
1988:
1989: In Forth (and other stack-based languages, e.g., Postscript) the
1990: parameter and result order of a definition is important and should be
1991: designed well. The general guideline is to design the stack effect such
1992: that the word is simple to use in most cases, even if that complicates
1993: the implementation of the word. Some concrete rules are:
1994:
1995: @itemize @bullet
1996:
1997: @item
1998: Words consume all of their parameters (e.g., @code{.}).
1999:
2000: @item
2001: If there is a convention on the order of parameters (e.g., from
2002: mathematics or another programming language), stick with it (e.g.,
2003: @code{-}).
2004:
2005: @item
2006: If one parameter usually requires only a short computation (e.g., it is
2007: a constant), pass it on the top of the stack. Conversely, parameters
2008: that usually require a long sequence of code to compute should be passed
2009: as the bottom (i.e., first) parameter. This makes the code easier to
2010: read, because reader does not need to keep track of the bottom item
2011: through a long sequence of code (or, alternatively, through stack
2012: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
2013: address on top of the stack because it is usually simpler to compute
2014: than the stored value (often the address is just a variable).
2015:
2016: @item
2017: Similarly, results that are usually consumed quickly should be returned
2018: on the top of stack, whereas a result that is often used in long
2019: computations should be passed as bottom result. E.g., the file words
2020: like @code{open-file} return the error code on the top of stack, because
2021: it is usually consumed quickly by @code{throw}; moreover, the error code
2022: has to be checked before doing anything with the other results.
2023:
2024: @end itemize
2025:
2026: These rules are just general guidelines, don't lose sight of the overall
2027: goal to make the words easy to use. E.g., if the convention rule
2028: conflicts with the computation-length rule, you might decide in favour
2029: of the convention if the word will be used rarely, and in favour of the
2030: computation-length rule if the word will be used frequently (because
2031: with frequent use the cost of breaking the computation-length rule would
2032: be quite high, and frequent use makes it easier to remember an
2033: unconventional order).
2034:
2035: @c example !! structure package
2036:
2037:
2038: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2039: @section Local Variables
2040: @cindex local variables, tutorial
2041:
2042: You can define local variables (@emph{locals}) in a colon definition:
2043:
2044: @example
2045: : swap @{ a b -- b a @}
2046: b a ;
2047: 1 2 swap .s 2drop
2048: @end example
2049:
2050: (If your Forth system does not support this syntax, include
2051: @file{compat/anslocals.fs} first).
2052:
2053: In this example @code{@{ a b -- b a @}} is the locals definition; it
2054: takes two cells from the stack, puts the top of stack in @code{b} and
2055: the next stack element in @code{a}. @code{--} starts a comment ending
2056: with @code{@}}. After the locals definition, using the name of the
2057: local will push its value on the stack. You can leave the comment
2058: part (@code{-- b a}) away:
2059:
2060: @example
2061: : swap ( x1 x2 -- x2 x1 )
2062: @{ a b @} b a ;
2063: @end example
2064:
2065: In Gforth you can have several locals definitions, anywhere in a colon
2066: definition; in contrast, in a standard program you can have only one
2067: locals definition per colon definition, and that locals definition must
2068: be outside any controll structure.
2069:
2070: With locals you can write slightly longer definitions without running
2071: into stack trouble. However, I recommend trying to write colon
2072: definitions without locals for exercise purposes to help you gain the
2073: essential factoring skills.
2074:
2075: @assignment
2076: Rewrite your definitions until now with locals
2077: @endassignment
2078:
2079: Reference: @ref{Locals}.
2080:
2081:
2082: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2083: @section Conditional execution
2084: @cindex conditionals, tutorial
2085: @cindex if, tutorial
2086:
2087: In Forth you can use control structures only inside colon definitions.
2088: An @code{if}-structure looks like this:
2089:
2090: @example
2091: : abs ( n1 -- +n2 )
2092: dup 0 < if
2093: negate
2094: endif ;
2095: 5 abs .
2096: -5 abs .
2097: @end example
2098:
2099: @code{if} takes a flag from the stack. If the flag is non-zero (true),
2100: the following code is performed, otherwise execution continues after the
2101: @code{endif} (or @code{else}). @code{<} compares the top two stack
2102: elements and prioduces a flag:
2103:
2104: @example
2105: 1 2 < .
2106: 2 1 < .
2107: 1 1 < .
2108: @end example
2109:
2110: Actually the standard name for @code{endif} is @code{then}. This
2111: tutorial presents the examples using @code{endif}, because this is often
2112: less confusing for people familiar with other programming languages
2113: where @code{then} has a different meaning. If your system does not have
2114: @code{endif}, define it with
2115:
2116: @example
2117: : endif postpone then ; immediate
2118: @end example
2119:
2120: You can optionally use an @code{else}-part:
2121:
2122: @example
2123: : min ( n1 n2 -- n )
2124: 2dup < if
2125: drop
2126: else
2127: nip
2128: endif ;
2129: 2 3 min .
2130: 3 2 min .
2131: @end example
2132:
2133: @assignment
2134: Write @code{min} without @code{else}-part (hint: what's the definition
2135: of @code{nip}?).
2136: @endassignment
2137:
2138: Reference: @ref{Selection}.
2139:
2140:
2141: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2142: @section Flags and Comparisons
2143: @cindex flags tutorial
2144: @cindex comparison tutorial
2145:
2146: In a false-flag all bits are clear (0 when interpreted as integer). In
2147: a canonical true-flag all bits are set (-1 as a twos-complement signed
2148: integer); in many contexts (e.g., @code{if}) any non-zero value is
2149: treated as true flag.
2150:
2151: @example
2152: false .
2153: true .
2154: true hex u. decimal
2155: @end example
2156:
2157: Comparison words produce canonical flags:
2158:
2159: @example
2160: 1 1 = .
2161: 1 0= .
2162: 0 1 < .
2163: 0 0 < .
2164: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2165: -1 1 < .
2166: @end example
2167:
2168: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
2169: (or none) and the comparisons @code{= <> < > <= >=}. Only a part of
2170: these combinations are standard (for details see the standard,
2171: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
2172:
2173: You can use @code{and or xor invert} can be used as operations on
2174: canonical flags. Actually they are bitwise operations:
2175:
2176: @example
2177: 1 2 and .
2178: 1 2 or .
2179: 1 3 xor .
2180: 1 invert .
2181: @end example
2182:
2183: You can convert a zero/non-zero flag into a canonical flag with
2184: @code{0<>} (and complement it on the way with @code{0=}).
2185:
2186: @example
2187: 1 0= .
2188: 1 0<> .
2189: @end example
2190:
2191: You can use the all-bits-set feature of canonical flags and the bitwise
2192: operation of the Boolean operations to avoid @code{if}s:
2193:
2194: @example
2195: : foo ( n1 -- n2 )
2196: 0= if
2197: 14
2198: else
2199: 0
2200: endif ;
2201: 0 foo .
2202: 1 foo .
2203:
2204: : foo ( n1 -- n2 )
2205: 0= 14 and ;
2206: 0 foo .
2207: 1 foo .
2208: @end example
2209:
2210: @assignment
2211: Write @code{min} without @code{if}.
2212: @endassignment
2213:
2214: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
2215: @ref{Bitwise operations}.
2216:
2217:
2218: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2219: @section General Loops
2220: @cindex loops, indefinite, tutorial
2221:
2222: The endless loop is the most simple one:
2223:
2224: @example
2225: : endless ( -- )
2226: 0 begin
2227: dup . 1+
2228: again ;
2229: endless
2230: @end example
2231:
2232: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2233: does nothing at run-time, @code{again} jumps back to @code{begin}.
2234:
2235: A loop with one exit at any place looks like this:
2236:
2237: @example
2238: : log2 ( +n1 -- n2 )
2239: \ logarithmus dualis of n1>0, rounded down to the next integer
2240: assert( dup 0> )
2241: 2/ 0 begin
2242: over 0> while
2243: 1+ swap 2/ swap
2244: repeat
2245: nip ;
2246: 7 log2 .
2247: 8 log2 .
2248: @end example
2249:
2250: At run-time @code{while} consumes a flag; if it is 0, execution
2251: continues behind the @code{repeat}; if the flag is non-zero, execution
2252: continues behind the @code{while}. @code{Repeat} jumps back to
2253: @code{begin}, just like @code{again}.
2254:
2255: In Forth there are many combinations/abbreviations, like @code{1+}.
2256: However, @code{2/} is not one of them; it shifts it's argument right by
2257: one bit (arithmetic shift right):
2258:
2259: @example
2260: -5 2 / .
2261: -5 2/ .
2262: @end example
2263:
2264: @code{assert(} is no standard word, but you can get it on systems other
2265: then Gforth by including @file{compat/assert.fs}. You can see what it
2266: does by trying
2267:
2268: @example
2269: 0 log2 .
2270: @end example
2271:
2272: Here's a loop with an exit at the end:
2273:
2274: @example
2275: : log2 ( +n1 -- n2 )
2276: \ logarithmus dualis of n1>0, rounded down to the next integer
2277: assert( dup 0 > )
2278: -1 begin
2279: 1+ swap 2/ swap
2280: over 0 <=
2281: until
2282: nip ;
2283: @end example
2284:
2285: @code{Until} consumes a flag; if it is non-zero, execution continues at
2286: the @code{begin}, otherwise after the @code{until}.
2287:
2288: @assignment
2289: Write a definition for computing the greatest common divisor.
2290: @endassignment
2291:
2292: Reference: @ref{Simple Loops}.
2293:
2294:
2295: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2296: @section Counted loops
2297: @cindex loops, counted, tutorial
2298:
2299: @example
2300: : ^ ( n1 u -- n )
2301: \ n = the uth power of u1
2302: 1 swap 0 u+do
2303: over *
2304: loop
2305: nip ;
2306: 3 2 ^ .
2307: 4 3 ^ .
2308: @end example
2309:
2310: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2311: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2312: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2313: times (or not at all, if @code{u3-u4<0}).
2314:
2315: You can see the stack effect design rules at work in the stack effect of
2316: the loop start words: Since the start value of the loop is more
2317: frequently constant than the end value, the start value is passed on
2318: the top-of-stack.
2319:
2320: You can access the counter of a counted loop with @code{i}:
2321:
2322: @example
2323: : fac ( u -- u! )
2324: 1 swap 1+ 1 u+do
2325: i *
2326: loop ;
2327: 5 fac .
2328: 7 fac .
2329: @end example
2330:
2331: There is also @code{+do}, which expects signed numbers (important for
2332: deciding whether to enter the loop).
2333:
2334: @assignment
2335: Write a definition for computing the nth Fibonacci number.
2336: @endassignment
2337:
2338: You can also use increments other than 1:
2339:
2340: @example
2341: : up2 ( n1 n2 -- )
2342: +do
2343: i .
2344: 2 +loop ;
2345: 10 0 up2
2346:
2347: : down2 ( n1 n2 -- )
2348: -do
2349: i .
2350: 2 -loop ;
2351: 0 10 down2
2352: @end example
2353:
2354: Reference: @ref{Counted Loops}.
2355:
2356:
2357: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2358: @section Recursion
2359: @cindex recursion tutorial
2360:
2361: Usually the name of a definition is not visible in the definition; but
2362: earlier definitions are usually visible:
2363:
2364: @example
2365: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2366: : / ( n1 n2 -- n )
2367: dup 0= if
2368: -10 throw \ report division by zero
2369: endif
2370: / \ old version
2371: ;
2372: 1 0 /
2373: @end example
2374:
2375: For recursive definitions you can use @code{recursive} (non-standard) or
2376: @code{recurse}:
2377:
2378: @example
2379: : fac1 ( n -- n! ) recursive
2380: dup 0> if
2381: dup 1- fac1 *
2382: else
2383: drop 1
2384: endif ;
2385: 7 fac1 .
2386:
2387: : fac2 ( n -- n! )
2388: dup 0> if
2389: dup 1- recurse *
2390: else
2391: drop 1
2392: endif ;
2393: 8 fac2 .
2394: @end example
2395:
2396: @assignment
2397: Write a recursive definition for computing the nth Fibonacci number.
2398: @endassignment
2399:
2400: Reference (including indirect recursion): @xref{Calls and returns}.
2401:
2402:
2403: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2404: @section Leaving definitions or loops
2405: @cindex leaving definitions, tutorial
2406: @cindex leaving loops, tutorial
2407:
2408: @code{EXIT} exits the current definition right away. For every counted
2409: loop that is left in this way, an @code{UNLOOP} has to be performed
2410: before the @code{EXIT}:
2411:
2412: @c !! real examples
2413: @example
2414: : ...
2415: ... u+do
2416: ... if
2417: ... unloop exit
2418: endif
2419: ...
2420: loop
2421: ... ;
2422: @end example
2423:
2424: @code{LEAVE} leaves the innermost counted loop right away:
2425:
2426: @example
2427: : ...
2428: ... u+do
2429: ... if
2430: ... leave
2431: endif
2432: ...
2433: loop
2434: ... ;
2435: @end example
2436:
2437: @c !! example
2438:
2439: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
2440:
2441:
2442: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2443: @section Return Stack
2444: @cindex return stack tutorial
2445:
2446: In addition to the data stack Forth also has a second stack, the return
2447: stack; most Forth systems store the return addresses of procedure calls
2448: there (thus its name). Programmers can also use this stack:
2449:
2450: @example
2451: : foo ( n1 n2 -- )
2452: .s
2453: >r .s
2454: r@@ .
2455: >r .s
2456: r@@ .
2457: r> .
2458: r@@ .
2459: r> . ;
2460: 1 2 foo
2461: @end example
2462:
2463: @code{>r} takes an element from the data stack and pushes it onto the
2464: return stack; conversely, @code{r>} moves an elementm from the return to
2465: the data stack; @code{r@@} pushes a copy of the top of the return stack
2466: on the return stack.
2467:
2468: Forth programmers usually use the return stack for storing data
2469: temporarily, if using the data stack alone would be too complex, and
2470: factoring and locals are not an option:
2471:
2472: @example
2473: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2474: rot >r rot r> ;
2475: @end example
2476:
2477: The return address of the definition and the loop control parameters of
2478: counted loops usually reside on the return stack, so you have to take
2479: all items, that you have pushed on the return stack in a colon
2480: definition or counted loop, from the return stack before the definition
2481: or loop ends. You cannot access items that you pushed on the return
2482: stack outside some definition or loop within the definition of loop.
2483:
2484: If you miscount the return stack items, this usually ends in a crash:
2485:
2486: @example
2487: : crash ( n -- )
2488: >r ;
2489: 5 crash
2490: @end example
2491:
2492: You cannot mix using locals and using the return stack (according to the
2493: standard; Gforth has no problem). However, they solve the same
2494: problems, so this shouldn't be an issue.
2495:
2496: @assignment
2497: Can you rewrite any of the definitions you wrote until now in a better
2498: way using the return stack?
2499: @endassignment
2500:
2501: Reference: @ref{Return stack}.
2502:
2503:
2504: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2505: @section Memory
2506: @cindex memory access/allocation tutorial
2507:
2508: You can create a global variable @code{v} with
2509:
2510: @example
2511: variable v ( -- addr )
2512: @end example
2513:
2514: @code{v} pushes the address of a cell in memory on the stack. This cell
2515: was reserved by @code{variable}. You can use @code{!} (store) to store
2516: values into this cell and @code{@@} (fetch) to load the value from the
2517: stack into memory:
2518:
2519: @example
2520: v .
2521: 5 v ! .s
2522: v @@ .
2523: @end example
2524:
2525: You can see a raw dump of memory with @code{dump}:
2526:
2527: @example
2528: v 1 cells .s dump
2529: @end example
2530:
2531: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2532: generally, address units (aus)) that @code{n1 cells} occupy. You can
2533: also reserve more memory:
2534:
2535: @example
2536: create v2 20 cells allot
2537: v2 20 cells dump
2538: @end example
2539:
2540: creates a word @code{v2} and reserves 20 uninitialized cells; the
2541: address pushed by @code{v2} points to the start of these 20 cells. You
2542: can use address arithmetic to access these cells:
2543:
2544: @example
2545: 3 v2 5 cells + !
2546: v2 20 cells dump
2547: @end example
2548:
2549: You can reserve and initialize memory with @code{,}:
2550:
2551: @example
2552: create v3
2553: 5 , 4 , 3 , 2 , 1 ,
2554: v3 @@ .
2555: v3 cell+ @@ .
2556: v3 2 cells + @@ .
2557: v3 5 cells dump
2558: @end example
2559:
2560: @assignment
2561: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2562: @code{u} cells, with the first of these cells at @code{addr}, the next
2563: one at @code{addr cell+} etc.
2564: @endassignment
2565:
2566: You can also reserve memory without creating a new word:
2567:
2568: @example
2569: here 10 cells allot .
2570: here .
2571: @end example
2572:
2573: @code{Here} pushes the start address of the memory area. You should
2574: store it somewhere, or you will have a hard time finding the memory area
2575: again.
2576:
2577: @code{Allot} manages dictionary memory. The dictionary memory contains
2578: the system's data structures for words etc. on Gforth and most other
2579: Forth systems. It is managed like a stack: You can free the memory that
2580: you have just @code{allot}ed with
2581:
2582: @example
2583: -10 cells allot
2584: here .
2585: @end example
2586:
2587: Note that you cannot do this if you have created a new word in the
2588: meantime (because then your @code{allot}ed memory is no longer on the
2589: top of the dictionary ``stack'').
2590:
2591: Alternatively, you can use @code{allocate} and @code{free} which allow
2592: freeing memory in any order:
2593:
2594: @example
2595: 10 cells allocate throw .s
2596: 20 cells allocate throw .s
2597: swap
2598: free throw
2599: free throw
2600: @end example
2601:
2602: The @code{throw}s deal with errors (e.g., out of memory).
2603:
2604: And there is also a
2605: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2606: garbage collector}, which eliminates the need to @code{free} memory
2607: explicitly.
2608:
2609: Reference: @ref{Memory}.
2610:
2611:
2612: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2613: @section Characters and Strings
2614: @cindex strings tutorial
2615: @cindex characters tutorial
2616:
2617: On the stack characters take up a cell, like numbers. In memory they
2618: have their own size (one 8-bit byte on most systems), and therefore
2619: require their own words for memory access:
2620:
2621: @example
2622: create v4
2623: 104 c, 97 c, 108 c, 108 c, 111 c,
2624: v4 4 chars + c@@ .
2625: v4 5 chars dump
2626: @end example
2627:
2628: The preferred representation of strings on the stack is @code{addr
2629: u-count}, where @code{addr} is the address of the first character and
2630: @code{u-count} is the number of characters in the string.
2631:
2632: @example
2633: v4 5 type
2634: @end example
2635:
2636: You get a string constant with
2637:
2638: @example
2639: s" hello, world" .s
2640: type
2641: @end example
2642:
2643: Make sure you have a space between @code{s"} and the string; @code{s"}
2644: is a normal Forth word and must be delimited with white space (try what
2645: happens when you remove the space).
2646:
2647: However, this interpretive use of @code{s"} is quite restricted: the
2648: string exists only until the next call of @code{s"} (some Forth systems
2649: keep more than one of these strings, but usually they still have a
2650: limited lifetime).
2651:
2652: @example
2653: s" hello," s" world" .s
2654: type
2655: type
2656: @end example
2657:
2658: You can also use @code{s"} in a definition, and the resulting
2659: strings then live forever (well, for as long as the definition):
2660:
2661: @example
2662: : foo s" hello," s" world" ;
2663: foo .s
2664: type
2665: type
2666: @end example
2667:
2668: @assignment
2669: @code{Emit ( c -- )} types @code{c} as character (not a number).
2670: Implement @code{type ( addr u -- )}.
2671: @endassignment
2672:
2673: Reference: @ref{Memory Blocks}.
2674:
2675:
2676: @node Alignment Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Characters and Strings Tutorial, Tutorial
2677: @section Alignment
2678: @cindex alignment tutorial
2679: @cindex memory alignment tutorial
2680:
2681: On many processors cells have to be aligned in memory, if you want to
2682: access them with @code{@@} and @code{!} (and even if the processor does
2683: not require alignment, access to aligned cells is faster).
2684:
2685: @code{Create} aligns @code{here} (i.e., the place where the next
2686: allocation will occur, and that the @code{create}d word points to).
2687: Likewise, the memory produced by @code{allocate} starts at an aligned
2688: address. Adding a number of @code{cells} to an aligned address produces
2689: another aligned address.
2690:
2691: However, address arithmetic involving @code{char+} and @code{chars} can
2692: create an address that is not cell-aligned. @code{Aligned ( addr --
2693: a-addr )} produces the next aligned address:
2694:
2695: @example
2696: v3 char+ aligned .s @@ .
2697: v3 char+ .s @@ .
2698: @end example
2699:
2700: Similarly, @code{align} advances @code{here} to the next aligned
2701: address:
2702:
2703: @example
2704: create v5 97 c,
2705: here .
2706: align here .
2707: 1000 ,
2708: @end example
2709:
2710: Note that you should use aligned addresses even if your processor does
2711: not require them, if you want your program to be portable.
2712:
2713: Reference: @ref{Address arithmetic}.
2714:
2715:
2716: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Alignment Tutorial, Tutorial
2717: @section Interpretation and Compilation Semantics and Immediacy
2718: @cindex semantics tutorial
2719: @cindex interpretation semantics tutorial
2720: @cindex compilation semantics tutorial
2721: @cindex immediate, tutorial
2722:
2723: When a word is compiled, it behaves differently from being interpreted.
2724: E.g., consider @code{+}:
2725:
2726: @example
2727: 1 2 + .
2728: : foo + ;
2729: @end example
2730:
2731: These two behaviours are known as compilation and interpretation
2732: semantics. For normal words (e.g., @code{+}), the compilation semantics
2733: is to append the interpretation semantics to the currently defined word
2734: (@code{foo} in the example above). I.e., when @code{foo} is executed
2735: later, the interpretation semantics of @code{+} (i.e., adding two
2736: numbers) will be performed.
2737:
2738: However, there are words with non-default compilation semantics, e.g.,
2739: the control-flow words like @code{if}. You can use @code{immediate} to
2740: change the compilation semantics of the last defined word to be equal to
2741: the interpretation semantics:
2742:
2743: @example
2744: : [FOO] ( -- )
2745: 5 . ; immediate
2746:
2747: [FOO]
2748: : bar ( -- )
2749: [FOO] ;
2750: bar
2751: see bar
2752: @end example
2753:
2754: Two conventions to mark words with non-default compilation semnatics are
2755: names with brackets (more frequently used) and to write them all in
2756: upper case (less frequently used).
2757:
2758: In Gforth (and many other systems) you can also remove the
2759: interpretation semantics with @code{compile-only} (the compilation
2760: semantics is derived from the original interpretation semantics):
2761:
2762: @example
2763: : flip ( -- )
2764: 6 . ; compile-only \ but not immediate
2765: flip
2766:
2767: : flop ( -- )
2768: flip ;
2769: flop
2770: @end example
2771:
2772: In this example the interpretation semantics of @code{flop} is equal to
2773: the original interpretation semantics of @code{flip}.
2774:
2775: The text interpreter has two states: in interpret state, it performs the
2776: interpretation semantics of words it encounters; in compile state, it
2777: performs the compilation semantics of these words.
2778:
2779: Among other things, @code{:} switches into compile state, and @code{;}
2780: switches back to interpret state. They contain the factors @code{]}
2781: (switch to compile state) and @code{[} (switch to interpret state), that
2782: do nothing but switch the state.
2783:
2784: @example
2785: : xxx ( -- )
2786: [ 5 . ]
2787: ;
2788:
2789: xxx
2790: see xxx
2791: @end example
2792:
2793: These brackets are also the source of the naming convention mentioned
2794: above.
2795:
2796: Reference: @ref{Interpretation and Compilation Semantics}.
2797:
2798:
2799: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2800: @section Execution Tokens
2801: @cindex execution tokens tutorial
2802: @cindex XT tutorial
2803:
2804: @code{' word} gives you the execution token (XT) of a word. The XT is a
2805: cell representing the interpretation semantics of a word. You can
2806: execute this semantics with @code{execute}:
2807:
2808: @example
2809: ' + .s
2810: 1 2 rot execute .
2811: @end example
2812:
2813: The XT is similar to a function pointer in C. However, parameter
2814: passing through the stack makes it a little more flexible:
2815:
2816: @example
2817: : map-array ( ... addr u xt -- ... )
2818: \ executes xt ( ... x -- ... ) for every element of the array starting
2819: \ at addr and containing u elements
2820: @{ xt @}
2821: cells over + swap ?do
2822: i @@ xt execute
2823: 1 cells +loop ;
2824:
2825: create a 3 , 4 , 2 , -1 , 4 ,
2826: a 5 ' . map-array .s
2827: 0 a 5 ' + map-array .
2828: s" max-n" environment? drop .s
2829: a 5 ' min map-array .
2830: @end example
2831:
2832: You can use map-array with the XTs of words that consume one element
2833: more than they produce. In theory you can also use it with other XTs,
2834: but the stack effect then depends on the size of the array, which is
2835: hard to understand.
2836:
2837: Since XTs are cell-sized, you can store them in memory and manipulate
2838: them on the stack like other cells. You can also compile the XT into a
2839: word with @code{compile,}:
2840:
2841: @example
2842: : foo1 ( n1 n2 -- n )
2843: [ ' + compile, ] ;
2844: see foo
2845: @end example
2846:
2847: This is non-standard, because @code{compile,} has no compilation
2848: semantics in the standard, but it works in good Forth systems. For the
2849: broken ones, use
2850:
2851: @example
2852: : [compile,] compile, ; immediate
2853:
2854: : foo1 ( n1 n2 -- n )
2855: [ ' + ] [compile,] ;
2856: see foo
2857: @end example
2858:
2859: @code{'} is a word with default compilation semantics; it parses the
2860: next word when its interpretation semantics are executed, not during
2861: compilation:
2862:
2863: @example
2864: : foo ( -- xt )
2865: ' ;
2866: see foo
2867: : bar ( ... "word" -- ... )
2868: ' execute ;
2869: see bar
2870: 1 2 bar + .
2871: @end example
2872:
2873: You often want to parse a word during compilation and compile its XT so
2874: it will be pushed on the stack at run-time. @code{[']} does this:
2875:
2876: @example
2877: : xt-+ ( -- xt )
2878: ['] + ;
2879: see xt-+
2880: 1 2 xt-+ execute .
2881: @end example
2882:
2883: Many programmers tend to see @code{'} and the word it parses as one
2884: unit, and expect it to behave like @code{[']} when compiled, and are
2885: confused by the actual behaviour. If you are, just remember that the
2886: Forth system just takes @code{'} as one unit and has no idea that it is
2887: a parsing word (attempts to convenience programmers in this issue have
2888: usually resulted in even worse pitfalls, see
2889: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
2890: @code{State}-smartness---Why it is evil and How to Exorcise it}).
2891:
2892: Note that the state of the interpreter does not come into play when
2893: creating and executing XTs. I.e., even when you execute @code{'} in
2894: compile state, it still gives you the interpretation semantics. And
2895: whatever that state is, @code{execute} performs the semantics
2896: represented by the XT (i.e., for XTs produced with @code{'} the
2897: interpretation semantics).
2898:
2899: Reference: @ref{Tokens for Words}.
2900:
2901:
2902: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
2903: @section Exceptions
2904: @cindex exceptions tutorial
2905:
2906: @code{throw ( n -- )} causes an exception unless n is zero.
2907:
2908: @example
2909: 100 throw .s
2910: 0 throw .s
2911: @end example
2912:
2913: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
2914: it catches exceptions and pushes the number of the exception on the
2915: stack (or 0, if the xt executed without exception). If there was an
2916: exception, the stacks have the same depth as when entering @code{catch}:
2917:
2918: @example
2919: .s
2920: 3 0 ' / catch .s
2921: 3 2 ' / catch .s
2922: @end example
2923:
2924: @assignment
2925: Try the same with @code{execute} instead of @code{catch}.
2926: @endassignment
2927:
2928: @code{Throw} always jumps to the dynamically next enclosing
2929: @code{catch}, even if it has to leave several call levels to achieve
2930: this:
2931:
2932: @example
2933: : foo 100 throw ;
2934: : foo1 foo ." after foo" ;
2935: : bar ['] foo1 catch ;
2936: bar .
2937: @end example
2938:
2939: It is often important to restore a value upon leaving a definition, even
2940: if the definition is left through an exception. You can ensure this
2941: like this:
2942:
2943: @example
2944: : ...
2945: save-x
2946: ['] word-changing-x catch ( ... n )
2947: restore-x
2948: ( ... n ) throw ;
2949: @end example
2950:
2951: Gforth provides an alternative syntax in addition to @code{catch}:
2952: @code{try ... recover ... endtry}. If the code between @code{try} and
2953: @code{recover} has an exception, the stack depths are restored, the
2954: exception number is pushed on the stack, and the code between
2955: @code{recover} and @code{endtry} is performed. E.g., the definition for
2956: @code{catch} is
2957:
2958: @example
2959: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
2960: try
2961: execute 0
2962: recover
2963: nip
2964: endtry ;
2965: @end example
2966:
2967: The equivalent to the restoration code above is
2968:
2969: @example
2970: : ...
2971: save-x
2972: try
2973: word-changing-x
2974: end-try
2975: restore-x
2976: throw ;
2977: @end example
2978:
2979: As you can see, the @code{recover} part is optional.
2980:
2981: Reference: @ref{Exception Handling}.
2982:
2983:
2984: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
2985: @section Defining Words
2986: @cindex defining words tutorial
2987: @cindex does> tutorial
2988: @cindex create...does> tutorial
2989:
2990: @c before semantics?
2991:
2992: @code{:}, @code{create}, and @code{variable} are definition words: They
2993: define other words. @code{Constant} is another definition word:
2994:
2995: @example
2996: 5 constant foo
2997: foo .
2998: @end example
2999:
3000: You can also use the prefixes @code{2} (double-cell) and @code{f}
3001: (floating point) with @code{variable} and @code{constant}.
3002:
3003: You can also define your own defining words. E.g.:
3004:
3005: @example
3006: : variable ( "name" -- )
3007: create 0 , ;
3008: @end example
3009:
3010: You can also define defining words that create words that do something
3011: other than just producing their address:
3012:
3013: @example
3014: : constant ( n "name" -- )
3015: create ,
3016: does> ( -- n )
3017: ( addr ) @@ ;
3018:
3019: 5 constant foo
3020: foo .
3021: @end example
3022:
3023: The definition of @code{constant} above ends at the @code{does>}; i.e.,
3024: @code{does>} replaces @code{;}, but it also does something else: It
3025: changes the last defined word such that it pushes the address of the
3026: body of the word and then performs the code after the @code{does>}
3027: whenever it is called.
3028:
3029: In the example above, @code{constant} uses @code{,} to store 5 into the
3030: body of @code{foo}. When @code{foo} executes, it pushes the address of
3031: the body onto the stack, then (in the code after the @code{does>})
3032: fetches the 5 from there.
3033:
3034: The stack comment near the @code{does>} reflects the stack effect of the
3035: defined word, not the stack effect of the code after the @code{does>}
3036: (the difference is that the code expects the address of the body that
3037: the stack comment does not show).
3038:
3039: You can use these definition words to do factoring in cases that involve
3040: (other) definition words. E.g., a field offset is always added to an
3041: address. Instead of defining
3042:
3043: @example
3044: 2 cells constant offset-field1
3045: @end example
3046:
3047: and using this like
3048:
3049: @example
3050: ( addr ) offset-field1 +
3051: @end example
3052:
3053: you can define a definition word
3054:
3055: @example
3056: : simple-field ( n "name" -- )
3057: create ,
3058: does> ( n1 -- n1+n )
3059: ( addr ) @@ + ;
3060: @end example
3061:
3062: Definition and use of field offsets now look like this:
3063:
3064: @example
3065: 2 cells simple-field field1
3066: create mystruct 4 cells allot
3067: mystruct .s field1 .s drop
3068: @end example
3069:
3070: If you want to do something with the word without performing the code
3071: after the @code{does>}, you can access the body of a @code{create}d word
3072: with @code{>body ( xt -- addr )}:
3073:
3074: @example
3075: : value ( n "name" -- )
3076: create ,
3077: does> ( -- n1 )
3078: @@ ;
3079: : to ( n "name" -- )
3080: ' >body ! ;
3081:
3082: 5 value foo
3083: foo .
3084: 7 to foo
3085: foo .
3086: @end example
3087:
3088: @assignment
3089: Define @code{defer ( "name" -- )}, which creates a word that stores an
3090: XT (at the start the XT of @code{abort}), and upon execution
3091: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3092: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3093: recursion is one application of @code{defer}.
3094: @endassignment
3095:
3096: Reference: @ref{User-defined Defining Words}.
3097:
3098:
3099: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3100: @section Arrays and Records
3101: @cindex arrays tutorial
3102: @cindex records tutorial
3103: @cindex structs tutorial
3104:
3105: Forth has no standard words for defining data structures such as arrays
3106: and records (structs in C terminology), but you can build them yourself
3107: based on address arithmetic. You can also define words for defining
3108: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
3109:
3110: One of the first projects a Forth newcomer sets out upon when learning
3111: about defining words is an array defining word (possibly for
3112: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3113: learn something from it. However, don't be disappointed when you later
3114: learn that you have little use for these words (inappropriate use would
3115: be even worse). I have not yet found a set of useful array words yet;
3116: the needs are just too diverse, and named, global arrays (the result of
3117: naive use of defining words) are often not flexible enough (e.g.,
3118: consider how to pass them as parameters). Another such project is a set
3119: of words to help dealing with strings.
3120:
3121: On the other hand, there is a useful set of record words, and it has
3122: been defined in @file{compat/struct.fs}; these words are predefined in
3123: Gforth. They are explained in depth elsewhere in this manual (see
3124: @pxref{Structures}). The @code{simple-field} example above is
3125: simplified variant of fields in this package.
3126:
3127:
3128: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3129: @section @code{POSTPONE}
3130: @cindex postpone tutorial
3131:
3132: You can compile the compilation semantics (instead of compiling the
3133: interpretation semantics) of a word with @code{POSTPONE}:
3134:
3135: @example
3136: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3137: POSTPONE + ; immediate
3138: : foo ( n1 n2 -- n )
3139: MY-+ ;
3140: 1 2 foo .
3141: see foo
3142: @end example
3143:
3144: During the definition of @code{foo} the text interpreter performs the
3145: compilation semantics of @code{MY-+}, which performs the compilation
3146: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3147:
3148: This example also displays separate stack comments for the compilation
3149: semantics and for the stack effect of the compiled code. For words with
3150: default compilation semantics these stack effects are usually not
3151: displayed; the stack effect of the compilation semantics is always
3152: @code{( -- )} for these words, the stack effect for the compiled code is
3153: the stack effect of the interpretation semantics.
3154:
3155: Note that the state of the interpreter does not come into play when
3156: performing the compilation semantics in this way. You can also perform
3157: it interpretively, e.g.:
3158:
3159: @example
3160: : foo2 ( n1 n2 -- n )
3161: [ MY-+ ] ;
3162: 1 2 foo .
3163: see foo
3164: @end example
3165:
3166: However, there are some broken Forth systems where this does not always
3167: work, and therefore this practice was been declared non-standard in
3168: 1999.
3169: @c !! repair.fs
3170:
3171: Here is another example for using @code{POSTPONE}:
3172:
3173: @example
3174: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3175: POSTPONE negate POSTPONE + ; immediate compile-only
3176: : bar ( n1 n2 -- n )
3177: MY-- ;
3178: 2 1 bar .
3179: see bar
3180: @end example
3181:
3182: You can define @code{ENDIF} in this way:
3183:
3184: @example
3185: : ENDIF ( Compilation: orig -- )
3186: POSTPONE then ; immediate
3187: @end example
3188:
3189: @assignment
3190: Write @code{MY-2DUP} that has compilation semantics equivalent to
3191: @code{2dup}, but compiles @code{over over}.
3192: @endassignment
3193:
3194: @c !! @xref{Macros} for reference
3195:
3196:
3197: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3198: @section @code{Literal}
3199: @cindex literal tutorial
3200:
3201: You cannot @code{POSTPONE} numbers:
3202:
3203: @example
3204: : [FOO] POSTPONE 500 ; immediate
3205: @end example
3206:
3207: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
3208:
3209: @example
3210: : [FOO] ( compilation: --; run-time: -- n )
3211: 500 POSTPONE literal ; immediate
3212:
3213: : flip [FOO] ;
3214: flip .
3215: see flip
3216: @end example
3217:
3218: @code{LITERAL} consumes a number at compile-time (when it's compilation
3219: semantics are executed) and pushes it at run-time (when the code it
3220: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3221: number computed at compile time into the current word:
3222:
3223: @example
3224: : bar ( -- n )
3225: [ 2 2 + ] literal ;
3226: see bar
3227: @end example
3228:
3229: @assignment
3230: Write @code{]L} which allows writing the example above as @code{: bar (
3231: -- n ) [ 2 2 + ]L ;}
3232: @endassignment
3233:
3234: @c !! @xref{Macros} for reference
3235:
3236:
3237: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3238: @section Advanced macros
3239: @cindex macros, advanced tutorial
3240: @cindex run-time code generation, tutorial
3241:
3242: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
3243: Execution Tokens}. It frequently performs @code{execute}, a relatively
3244: expensive operation in some Forth implementations. You can use
3245: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3246: and produce a word that contains the word to be performed directly:
3247:
3248: @c use ]] ... [[
3249: @example
3250: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3251: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3252: \ array beginning at addr and containing u elements
3253: @{ xt @}
3254: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
3255: POSTPONE i POSTPONE @@ xt compile,
3256: 1 cells POSTPONE literal POSTPONE +loop ;
3257:
3258: : sum-array ( addr u -- n )
3259: 0 rot rot [ ' + compile-map-array ] ;
3260: see sum-array
3261: a 5 sum-array .
3262: @end example
3263:
3264: You can use the full power of Forth for generating the code; here's an
3265: example where the code is generated in a loop:
3266:
3267: @example
3268: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3269: \ n2=n1+(addr1)*n, addr2=addr1+cell
3270: POSTPONE tuck POSTPONE @@
3271: POSTPONE literal POSTPONE * POSTPONE +
3272: POSTPONE swap POSTPONE cell+ ;
3273:
3274: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
3275: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
3276: 0 postpone literal postpone swap
3277: [ ' compile-vmul-step compile-map-array ]
3278: postpone drop ;
3279: see compile-vmul
3280:
3281: : a-vmul ( addr -- n )
3282: \ n=a*v, where v is a vector that's as long as a and starts at addr
3283: [ a 5 compile-vmul ] ;
3284: see a-vmul
3285: a a-vmul .
3286: @end example
3287:
3288: This example uses @code{compile-map-array} to show off, but you could
3289: also use @code{map-array} instead (try it now!).
3290:
3291: You can use this technique for efficient multiplication of large
3292: matrices. In matrix multiplication, you multiply every line of one
3293: matrix with every column of the other matrix. You can generate the code
3294: for one line once, and use it for every column. The only downside of
3295: this technique is that it is cumbersome to recover the memory consumed
3296: by the generated code when you are done (and in more complicated cases
3297: it is not possible portably).
3298:
3299: @c !! @xref{Macros} for reference
3300:
3301:
3302: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3303: @section Compilation Tokens
3304: @cindex compilation tokens, tutorial
3305: @cindex CT, tutorial
3306:
3307: This section is Gforth-specific. You can skip it.
3308:
3309: @code{' word compile,} compiles the interpretation semantics. For words
3310: with default compilation semantics this is the same as performing the
3311: compilation semantics. To represent the compilation semantics of other
3312: words (e.g., words like @code{if} that have no interpretation
3313: semantics), Gforth has the concept of a compilation token (CT,
3314: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3315: You can perform the compilation semantics represented by a CT with
3316: @code{execute}:
3317:
3318: @example
3319: : foo2 ( n1 n2 -- n )
3320: [ comp' + execute ] ;
3321: see foo
3322: @end example
3323:
3324: You can compile the compilation semantics represented by a CT with
3325: @code{postpone,}:
3326:
3327: @example
3328: : foo3 ( -- )
3329: [ comp' + postpone, ] ;
3330: see foo3
3331: @end example
3332:
3333: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
3334: @code{comp'} is particularly useful for words that have no
3335: interpretation semantics:
3336:
3337: @example
3338: ' if
3339: comp' if .s 2drop
3340: @end example
3341:
3342: Reference: @ref{Tokens for Words}.
3343:
3344:
3345: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3346: @section Wordlists and Search Order
3347: @cindex wordlists tutorial
3348: @cindex search order, tutorial
3349:
3350: The dictionary is not just a memory area that allows you to allocate
3351: memory with @code{allot}, it also contains the Forth words, arranged in
3352: several wordlists. When searching for a word in a wordlist,
3353: conceptually you start searching at the youngest and proceed towards
3354: older words (in reality most systems nowadays use hash-tables); i.e., if
3355: you define a word with the same name as an older word, the new word
3356: shadows the older word.
3357:
3358: Which wordlists are searched in which order is determined by the search
3359: order. You can display the search order with @code{order}. It displays
3360: first the search order, starting with the wordlist searched first, then
3361: it displays the wordlist that will contain newly defined words.
3362:
3363: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
3364:
3365: @example
3366: wordlist constant mywords
3367: @end example
3368:
3369: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3370: defined words (the @emph{current} wordlist):
3371:
3372: @example
3373: mywords set-current
3374: order
3375: @end example
3376:
3377: Gforth does not display a name for the wordlist in @code{mywords}
3378: because this wordlist was created anonymously with @code{wordlist}.
3379:
3380: You can get the current wordlist with @code{get-current ( -- wid)}. If
3381: you want to put something into a specific wordlist without overall
3382: effect on the current wordlist, this typically looks like this:
3383:
3384: @example
3385: get-current mywords set-current ( wid )
3386: create someword
3387: ( wid ) set-current
3388: @end example
3389:
3390: You can write the search order with @code{set-order ( wid1 .. widn n --
3391: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3392: searched wordlist is topmost.
3393:
3394: @example
3395: get-order mywords swap 1+ set-order
3396: order
3397: @end example
3398:
3399: Yes, the order of wordlists in the output of @code{order} is reversed
3400: from stack comments and the output of @code{.s} and thus unintuitive.
3401:
3402: @assignment
3403: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3404: wordlist to the search order. Define @code{previous ( -- )}, which
3405: removes the first searched wordlist from the search order. Experiment
3406: with boundary conditions (you will see some crashes or situations that
3407: are hard or impossible to leave).
3408: @endassignment
3409:
3410: The search order is a powerful foundation for providing features similar
3411: to Modula-2 modules and C++ namespaces. However, trying to modularize
3412: programs in this way has disadvantages for debugging and reuse/factoring
3413: that overcome the advantages in my experience (I don't do huge projects,
3414: though). These disadvantages are not so clear in other
3415: languages/programming environments, because these langauges are not so
3416: strong in debugging and reuse.
3417:
3418: @c !! example
3419:
3420: Reference: @ref{Word Lists}.
3421:
3422: @c ******************************************************************
3423: @node Introduction, Words, Tutorial, Top
3424: @comment node-name, next, previous, up
3425: @chapter An Introduction to ANS Forth
3426: @cindex Forth - an introduction
3427:
3428: The primary purpose of this manual is to document Gforth. However, since
3429: Forth is not a widely-known language and there is a lack of up-to-date
3430: teaching material, it seems worthwhile to provide some introductory
3431: material. For other sources of Forth-related
3432: information, see @ref{Forth-related information}.
3433:
3434: The examples in this section should work on any ANS Forth; the
3435: output shown was produced using Gforth. Each example attempts to
3436: reproduce the exact output that Gforth produces. If you try out the
3437: examples (and you should), what you should type is shown @kbd{like this}
3438: and Gforth's response is shown @code{like this}. The single exception is
3439: that, where the example shows @key{RET} it means that you should
3440: press the ``carriage return'' key. Unfortunately, some output formats for
3441: this manual cannot show the difference between @kbd{this} and
3442: @code{this} which will make trying out the examples harder (but not
3443: impossible).
3444:
3445: Forth is an unusual language. It provides an interactive development
3446: environment which includes both an interpreter and compiler. Forth
3447: programming style encourages you to break a problem down into many
3448: @cindex factoring
3449: small fragments (@dfn{factoring}), and then to develop and test each
3450: fragment interactively. Forth advocates assert that breaking the
3451: edit-compile-test cycle used by conventional programming languages can
3452: lead to great productivity improvements.
3453:
3454: @menu
3455: * Introducing the Text Interpreter::
3456: * Stacks and Postfix notation::
3457: * Your first definition::
3458: * How does that work?::
3459: * Forth is written in Forth::
3460: * Review - elements of a Forth system::
3461: * Where to go next::
3462: * Exercises::
3463: @end menu
3464:
3465: @comment ----------------------------------------------
3466: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3467: @section Introducing the Text Interpreter
3468: @cindex text interpreter
3469: @cindex outer interpreter
3470:
3471: @c IMO this is too detailed and the pace is too slow for
3472: @c an introduction. If you know German, take a look at
3473: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3474: @c to see how I do it - anton
3475:
3476: @c nac-> Where I have accepted your comments 100% and modified the text
3477: @c accordingly, I have deleted your comments. Elsewhere I have added a
3478: @c response like this to attempt to rationalise what I have done. Of
3479: @c course, this is a very clumsy mechanism for something that would be
3480: @c done far more efficiently over a beer. Please delete any dialogue
3481: @c you consider closed.
3482:
3483: When you invoke the Forth image, you will see a startup banner printed
3484: and nothing else (if you have Gforth installed on your system, try
3485: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
3486: its command line interpreter, which is called the @dfn{Text Interpreter}
3487: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
3488: about the text interpreter as you read through this chapter, for more
3489: detail @pxref{The Text Interpreter}).
3490:
3491: Although it's not obvious, Forth is actually waiting for your
3492: input. Type a number and press the @key{RET} key:
3493:
3494: @example
3495: @kbd{45@key{RET}} ok
3496: @end example
3497:
3498: Rather than give you a prompt to invite you to input something, the text
3499: interpreter prints a status message @i{after} it has processed a line
3500: of input. The status message in this case (``@code{ ok}'' followed by
3501: carriage-return) indicates that the text interpreter was able to process
3502: all of your input successfully. Now type something illegal:
3503:
3504: @example
3505: @kbd{qwer341@key{RET}}
3506: :1: Undefined word
3507: qwer341
3508: ^^^^^^^
3509: $400D2BA8 Bounce
3510: $400DBDA8 no.extensions
3511: @end example
3512:
3513: The exact text, other than the ``Undefined word'' may differ slightly on
3514: your system, but the effect is the same; when the text interpreter
3515: detects an error, it discards any remaining text on a line, resets
3516: certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
3517: messages}.
3518:
3519: The text interpreter waits for you to press carriage-return, and then
3520: processes your input line. Starting at the beginning of the line, it
3521: breaks the line into groups of characters separated by spaces. For each
3522: group of characters in turn, it makes two attempts to do something:
3523:
3524: @itemize @bullet
3525: @item
3526: @cindex name dictionary
3527: It tries to treat it as a command. It does this by searching a @dfn{name
3528: dictionary}. If the group of characters matches an entry in the name
3529: dictionary, the name dictionary provides the text interpreter with
3530: information that allows the text interpreter perform some actions. In
3531: Forth jargon, we say that the group
3532: @cindex word
3533: @cindex definition
3534: @cindex execution token
3535: @cindex xt
3536: of characters names a @dfn{word}, that the dictionary search returns an
3537: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3538: word, and that the text interpreter executes the xt. Often, the terms
3539: @dfn{word} and @dfn{definition} are used interchangeably.
3540: @item
3541: If the text interpreter fails to find a match in the name dictionary, it
3542: tries to treat the group of characters as a number in the current number
3543: base (when you start up Forth, the current number base is base 10). If
3544: the group of characters legitimately represents a number, the text
3545: interpreter pushes the number onto a stack (we'll learn more about that
3546: in the next section).
3547: @end itemize
3548:
3549: If the text interpreter is unable to do either of these things with any
3550: group of characters, it discards the group of characters and the rest of
3551: the line, then prints an error message. If the text interpreter reaches
3552: the end of the line without error, it prints the status message ``@code{ ok}''
3553: followed by carriage-return.
3554:
3555: This is the simplest command we can give to the text interpreter:
3556:
3557: @example
3558: @key{RET} ok
3559: @end example
3560:
3561: The text interpreter did everything we asked it to do (nothing) without
3562: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3563: command:
3564:
3565: @example
3566: @kbd{12 dup fred dup@key{RET}}
3567: :1: Undefined word
3568: 12 dup fred dup
3569: ^^^^
3570: $400D2BA8 Bounce
3571: $400DBDA8 no.extensions
3572: @end example
3573:
3574: When you press the carriage-return key, the text interpreter starts to
3575: work its way along the line:
3576:
3577: @itemize @bullet
3578: @item
3579: When it gets to the space after the @code{2}, it takes the group of
3580: characters @code{12} and looks them up in the name
3581: dictionary@footnote{We can't tell if it found them or not, but assume
3582: for now that it did not}. There is no match for this group of characters
3583: in the name dictionary, so it tries to treat them as a number. It is
3584: able to do this successfully, so it puts the number, 12, ``on the stack''
3585: (whatever that means).
3586: @item
3587: The text interpreter resumes scanning the line and gets the next group
3588: of characters, @code{dup}. It looks it up in the name dictionary and
3589: (you'll have to take my word for this) finds it, and executes the word
3590: @code{dup} (whatever that means).
3591: @item
3592: Once again, the text interpreter resumes scanning the line and gets the
3593: group of characters @code{fred}. It looks them up in the name
3594: dictionary, but can't find them. It tries to treat them as a number, but
3595: they don't represent any legal number.
3596: @end itemize
3597:
3598: At this point, the text interpreter gives up and prints an error
3599: message. The error message shows exactly how far the text interpreter
3600: got in processing the line. In particular, it shows that the text
3601: interpreter made no attempt to do anything with the final character
3602: group, @code{dup}, even though we have good reason to believe that the
3603: text interpreter would have no problem looking that word up and
3604: executing it a second time.
3605:
3606:
3607: @comment ----------------------------------------------
3608: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3609: @section Stacks, postfix notation and parameter passing
3610: @cindex text interpreter
3611: @cindex outer interpreter
3612:
3613: In procedural programming languages (like C and Pascal), the
3614: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3615: functions or procedures are called with @dfn{explicit parameters}. For
3616: example, in C we might write:
3617:
3618: @example
3619: total = total + new_volume(length,height,depth);
3620: @end example
3621:
3622: @noindent
3623: where new_volume is a function-call to another piece of code, and total,
3624: length, height and depth are all variables. length, height and depth are
3625: parameters to the function-call.
3626:
3627: In Forth, the equivalent of the function or procedure is the
3628: @dfn{definition} and parameters are implicitly passed between
3629: definitions using a shared stack that is visible to the
3630: programmer. Although Forth does support variables, the existence of the
3631: stack means that they are used far less often than in most other
3632: programming languages. When the text interpreter encounters a number, it
3633: will place (@dfn{push}) it on the stack. There are several stacks (the
3634: actual number is implementation-dependent ...) and the particular stack
3635: used for any operation is implied unambiguously by the operation being
3636: performed. The stack used for all integer operations is called the @dfn{data
3637: stack} and, since this is the stack used most commonly, references to
3638: ``the data stack'' are often abbreviated to ``the stack''.
3639:
3640: The stacks have a last-in, first-out (LIFO) organisation. If you type:
3641:
3642: @example
3643: @kbd{1 2 3@key{RET}} ok
3644: @end example
3645:
3646: Then this instructs the text interpreter to placed three numbers on the
3647: (data) stack. An analogy for the behaviour of the stack is to take a
3648: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3649: the table. The 3 was the last card onto the pile (``last-in'') and if
3650: you take a card off the pile then, unless you're prepared to fiddle a
3651: bit, the card that you take off will be the 3 (``first-out''). The
3652: number that will be first-out of the stack is called the @dfn{top of
3653: stack}, which
3654: @cindex TOS definition
3655: is often abbreviated to @dfn{TOS}.
3656:
3657: To understand how parameters are passed in Forth, consider the
3658: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3659: be surprised to learn that this definition performs addition. More
3660: precisely, it adds two number together and produces a result. Where does
3661: it get the two numbers from? It takes the top two numbers off the
3662: stack. Where does it place the result? On the stack. You can act-out the
3663: behaviour of @code{+} with your playing cards like this:
3664:
3665: @itemize @bullet
3666: @item
3667: Pick up two cards from the stack on the table
3668: @item
3669: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3670: numbers''
3671: @item
3672: Decide that the answer is 5
3673: @item
3674: Shuffle the two cards back into the pack and find a 5
3675: @item
3676: Put a 5 on the remaining ace that's on the table.
3677: @end itemize
3678:
3679: If you don't have a pack of cards handy but you do have Forth running,
3680: you can use the definition @code{.s} to show the current state of the stack,
3681: without affecting the stack. Type:
3682:
3683: @example
3684: @kbd{clearstack 1 2 3@key{RET}} ok
3685: @kbd{.s@key{RET}} <3> 1 2 3 ok
3686: @end example
3687:
3688: The text interpreter looks up the word @code{clearstack} and executes
3689: it; it tidies up the stack and removes any entries that may have been
3690: left on it by earlier examples. The text interpreter pushes each of the
3691: three numbers in turn onto the stack. Finally, the text interpreter
3692: looks up the word @code{.s} and executes it. The effect of executing
3693: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3694: followed by a list of all the items on the stack; the item on the far
3695: right-hand side is the TOS.
3696:
3697: You can now type:
3698:
3699: @example
3700: @kbd{+ .s@key{RET}} <2> 1 5 ok
3701: @end example
3702:
3703: @noindent
3704: which is correct; there are now 2 items on the stack and the result of
3705: the addition is 5.
3706:
3707: If you're playing with cards, try doing a second addition: pick up the
3708: two cards, work out that their sum is 6, shuffle them into the pack,
3709: look for a 6 and place that on the table. You now have just one item on
3710: the stack. What happens if you try to do a third addition? Pick up the
3711: first card, pick up the second card -- ah! There is no second card. This
3712: is called a @dfn{stack underflow} and consitutes an error. If you try to
3713: do the same thing with Forth it will report an error (probably a Stack
3714: Underflow or an Invalid Memory Address error).
3715:
3716: The opposite situation to a stack underflow is a @dfn{stack overflow},
3717: which simply accepts that there is a finite amount of storage space
3718: reserved for the stack. To stretch the playing card analogy, if you had
3719: enough packs of cards and you piled the cards up on the table, you would
3720: eventually be unable to add another card; you'd hit the ceiling. Gforth
3721: allows you to set the maximum size of the stacks. In general, the only
3722: time that you will get a stack overflow is because a definition has a
3723: bug in it and is generating data on the stack uncontrollably.
3724:
3725: There's one final use for the playing card analogy. If you model your
3726: stack using a pack of playing cards, the maximum number of items on
3727: your stack will be 52 (I assume you didn't use the Joker). The maximum
3728: @i{value} of any item on the stack is 13 (the King). In fact, the only
3729: possible numbers are positive integer numbers 1 through 13; you can't
3730: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3731: think about some of the cards, you can accommodate different
3732: numbers. For example, you could think of the Jack as representing 0,
3733: the Queen as representing -1 and the King as representing -2. Your
3734: @i{range} remains unchanged (you can still only represent a total of 13
3735: numbers) but the numbers that you can represent are -2 through 10.
3736:
3737: In that analogy, the limit was the amount of information that a single
3738: stack entry could hold, and Forth has a similar limit. In Forth, the
3739: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3740: implementation dependent and affects the maximum value that a stack
3741: entry can hold. A Standard Forth provides a cell size of at least
3742: 16-bits, and most desktop systems use a cell size of 32-bits.
3743:
3744: Forth does not do any type checking for you, so you are free to
3745: manipulate and combine stack items in any way you wish. A convenient way
3746: of treating stack items is as 2's complement signed integers, and that
3747: is what Standard words like @code{+} do. Therefore you can type:
3748:
3749: @example
3750: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
3751: @end example
3752:
3753: If you use numbers and definitions like @code{+} in order to turn Forth
3754: into a great big pocket calculator, you will realise that it's rather
3755: different from a normal calculator. Rather than typing 2 + 3 = you had
3756: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3757: result). The terminology used to describe this difference is to say that
3758: your calculator uses @dfn{Infix Notation} (parameters and operators are
3759: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3760: operators are separate), also called @dfn{Reverse Polish Notation}.
3761:
3762: Whilst postfix notation might look confusing to begin with, it has
3763: several important advantages:
3764:
3765: @itemize @bullet
3766: @item
3767: it is unambiguous
3768: @item
3769: it is more concise
3770: @item
3771: it fits naturally with a stack-based system
3772: @end itemize
3773:
3774: To examine these claims in more detail, consider these sums:
3775:
3776: @example
3777: 6 + 5 * 4 =
3778: 4 * 5 + 6 =
3779: @end example
3780:
3781: If you're just learning maths or your maths is very rusty, you will
3782: probably come up with the answer 44 for the first and 26 for the
3783: second. If you are a bit of a whizz at maths you will remember the
3784: @i{convention} that multiplication takes precendence over addition, and
3785: you'd come up with the answer 26 both times. To explain the answer 26
3786: to someone who got the answer 44, you'd probably rewrite the first sum
3787: like this:
3788:
3789: @example
3790: 6 + (5 * 4) =
3791: @end example
3792:
3793: If what you really wanted was to perform the addition before the
3794: multiplication, you would have to use parentheses to force it.
3795:
3796: If you did the first two sums on a pocket calculator you would probably
3797: get the right answers, unless you were very cautious and entered them using
3798: these keystroke sequences:
3799:
3800: 6 + 5 = * 4 =
3801: 4 * 5 = + 6 =
3802:
3803: Postfix notation is unambiguous because the order that the operators
3804: are applied is always explicit; that also means that parentheses are
3805: never required. The operators are @i{active} (the act of quoting the
3806: operator makes the operation occur) which removes the need for ``=''.
3807:
3808: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3809: equivalent ways:
3810:
3811: @example
3812: 6 5 4 * + or:
3813: 5 4 * 6 +
3814: @end example
3815:
3816: An important thing that you should notice about this notation is that
3817: the @i{order} of the numbers does not change; if you want to subtract
3818: 2 from 10 you type @code{10 2 -}.
3819:
3820: The reason that Forth uses postfix notation is very simple to explain: it
3821: makes the implementation extremely simple, and it follows naturally from
3822: using the stack as a mechanism for passing parameters. Another way of
3823: thinking about this is to realise that all Forth definitions are
3824: @i{active}; they execute as they are encountered by the text
3825: interpreter. The result of this is that the syntax of Forth is trivially
3826: simple.
3827:
3828:
3829:
3830: @comment ----------------------------------------------
3831: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3832: @section Your first Forth definition
3833: @cindex first definition
3834:
3835: Until now, the examples we've seen have been trivial; we've just been
3836: using Forth as a bigger-than-pocket calculator. Also, each calculation
3837: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3838: again@footnote{That's not quite true. If you press the up-arrow key on
3839: your keyboard you should be able to scroll back to any earlier command,
3840: edit it and re-enter it.} In this section we'll see how to add new
3841: words to Forth's vocabulary.
3842:
3843: The easiest way to create a new word is to use a @dfn{colon
3844: definition}. We'll define a few and try them out before worrying too
3845: much about how they work. Try typing in these examples; be careful to
3846: copy the spaces accurately:
3847:
3848: @example
3849: : add-two 2 + . ;
3850: : greet ." Hello and welcome" ;
3851: : demo 5 add-two ;
3852: @end example
3853:
3854: @noindent
3855: Now try them out:
3856:
3857: @example
3858: @kbd{greet@key{RET}} Hello and welcome ok
3859: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
3860: @kbd{4 add-two@key{RET}} 6 ok
3861: @kbd{demo@key{RET}} 7 ok
3862: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
3863: @end example
3864:
3865: The first new thing that we've introduced here is the pair of words
3866: @code{:} and @code{;}. These are used to start and terminate a new
3867: definition, respectively. The first word after the @code{:} is the name
3868: for the new definition.
3869:
3870: As you can see from the examples, a definition is built up of words that
3871: have already been defined; Forth makes no distinction between
3872: definitions that existed when you started the system up, and those that
3873: you define yourself.
3874:
3875: The examples also introduce the words @code{.} (dot), @code{."}
3876: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
3877: the stack and displays it. It's like @code{.s} except that it only
3878: displays the top item of the stack and it is destructive; after it has
3879: executed, the number is no longer on the stack. There is always one
3880: space printed after the number, and no spaces before it. Dot-quote
3881: defines a string (a sequence of characters) that will be printed when
3882: the word is executed. The string can contain any printable characters
3883: except @code{"}. A @code{"} has a special function; it is not a Forth
3884: word but it acts as a delimiter (the way that delimiters work is
3885: described in the next section). Finally, @code{dup} duplicates the value
3886: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
3887:
3888: We already know that the text interpreter searches through the
3889: dictionary to locate names. If you've followed the examples earlier, you
3890: will already have a definition called @code{add-two}. Lets try modifying
3891: it by typing in a new definition:
3892:
3893: @example
3894: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
3895: @end example
3896:
3897: Forth recognised that we were defining a word that already exists, and
3898: printed a message to warn us of that fact. Let's try out the new
3899: definition:
3900:
3901: @example
3902: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
3903: @end example
3904:
3905: @noindent
3906: All that we've actually done here, though, is to create a new
3907: definition, with a particular name. The fact that there was already a
3908: definition with the same name did not make any difference to the way
3909: that the new definition was created (except that Forth printed a warning
3910: message). The old definition of add-two still exists (try @code{demo}
3911: again to see that this is true). Any new definition will use the new
3912: definition of @code{add-two}, but old definitions continue to use the
3913: version that already existed at the time that they were @code{compiled}.
3914:
3915: Before you go on to the next section, try defining and redefining some
3916: words of your own.
3917:
3918: @comment ----------------------------------------------
3919: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
3920: @section How does that work?
3921: @cindex parsing words
3922:
3923: @c That's pretty deep (IMO way too deep) for an introduction. - anton
3924:
3925: @c Is it a good idea to talk about the interpretation semantics of a
3926: @c number? We don't have an xt to go along with it. - anton
3927:
3928: @c Now that I have eliminated execution semantics, I wonder if it would not
3929: @c be better to keep them (or add run-time semantics), to make it easier to
3930: @c explain what compilation semantics usually does. - anton
3931:
3932: @c nac-> I removed the term ``default compilation sematics'' from the
3933: @c introductory chapter. Removing ``execution semantics'' was making
3934: @c everything simpler to explain, then I think the use of this term made
3935: @c everything more complex again. I replaced it with ``default
3936: @c semantics'' (which is used elsewhere in the manual) by which I mean
3937: @c ``a definition that has neither the immediate nor the compile-only
3938: @c flag set''. I reworded big chunks of the ``how does that work''
3939: @c section (and, unusually for me, I think I even made it shorter!). See
3940: @c what you think -- I know I have not addressed your primary concern
3941: @c that it is too heavy-going for an introduction. From what I understood
3942: @c of your course notes it looks as though they might be a good framework.
3943: @c Things that I've tried to capture here are some things that came as a
3944: @c great revelation here when I first understood them. Also, I like the
3945: @c fact that a very simple code example shows up almost all of the issues
3946: @c that you need to understand to see how Forth works. That's unique and
3947: @c worthwhile to emphasise.
3948:
3949: Now we're going to take another look at the definition of @code{add-two}
3950: from the previous section. From our knowledge of the way that the text
3951: interpreter works, we would have expected this result when we tried to
3952: define @code{add-two}:
3953:
3954: @example
3955: @kbd{: add-two 2 + . ;@key{RET}}
3956: ^^^^^^^
3957: Error: Undefined word
3958: @end example
3959:
3960: The reason that this didn't happen is bound up in the way that @code{:}
3961: works. The word @code{:} does two special things. The first special
3962: thing that it does prevents the text interpreter from ever seeing the
3963: characters @code{add-two}. The text interpreter uses a variable called
3964: @cindex modifying >IN
3965: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
3966: input line. When it encounters the word @code{:} it behaves in exactly
3967: the same way as it does for any other word; it looks it up in the name
3968: dictionary, finds its xt and executes it. When @code{:} executes, it
3969: looks at the input buffer, finds the word @code{add-two} and advances the
3970: value of @code{>IN} to point past it. It then does some other stuff
3971: associated with creating the new definition (including creating an entry
3972: for @code{add-two} in the name dictionary). When the execution of @code{:}
3973: completes, control returns to the text interpreter, which is oblivious
3974: to the fact that it has been tricked into ignoring part of the input
3975: line.
3976:
3977: @cindex parsing words
3978: Words like @code{:} -- words that advance the value of @code{>IN} and so
3979: prevent the text interpreter from acting on the whole of the input line
3980: -- are called @dfn{parsing words}.
3981:
3982: @cindex @code{state} - effect on the text interpreter
3983: @cindex text interpreter - effect of state
3984: The second special thing that @code{:} does is change the value of a
3985: variable called @code{state}, which affects the way that the text
3986: interpreter behaves. When Gforth starts up, @code{state} has the value
3987: 0, and the text interpreter is said to be @dfn{interpreting}. During a
3988: colon definition (started with @code{:}), @code{state} is set to -1 and
3989: the text interpreter is said to be @dfn{compiling}.
3990:
3991: In this example, the text interpreter is compiling when it processes the
3992: string ``@code{2 + . ;}''. It still breaks the string down into
3993: character sequences in the same way. However, instead of pushing the
3994: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
3995: into the definition of @code{add-two} that will make the number @code{2} get
3996: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
3997: the behaviours of @code{+} and @code{.} are also compiled into the
3998: definition.
3999:
4000: One category of words don't get compiled. These so-called @dfn{immediate
4001: words} get executed (performed @i{now}) regardless of whether the text
4002: interpreter is interpreting or compiling. The word @code{;} is an
4003: immediate word. Rather than being compiled into the definition, it
4004: executes. Its effect is to terminate the current definition, which
4005: includes changing the value of @code{state} back to 0.
4006:
4007: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
4008: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
4009: definition.
4010:
4011: In Forth, every word or number can be described in terms of two
4012: properties:
4013:
4014: @itemize @bullet
4015: @item
4016: @cindex interpretation semantics
4017: Its @dfn{interpretation semantics} describe how it will behave when the
4018: text interpreter encounters it in @dfn{interpret} state. The
4019: interpretation semantics of a word are represented by an @dfn{execution
4020: token}.
4021: @item
4022: @cindex compilation semantics
4023: Its @dfn{compilation semantics} describe how it will behave when the
4024: text interpreter encounters it in @dfn{compile} state. The compilation
4025: semantics of a word are represented in an implementation-dependent way;
4026: Gforth uses a @dfn{compilation token}.
4027: @end itemize
4028:
4029: @noindent
4030: Numbers are always treated in a fixed way:
4031:
4032: @itemize @bullet
4033: @item
4034: When the number is @dfn{interpreted}, its behaviour is to push the
4035: number onto the stack.
4036: @item
4037: When the number is @dfn{compiled}, a piece of code is appended to the
4038: current definition that pushes the number when it runs. (In other words,
4039: the compilation semantics of a number are to postpone its interpretation
4040: semantics until the run-time of the definition that it is being compiled
4041: into.)
4042: @end itemize
4043:
4044: Words don't behave in such a regular way, but most have @i{default
4045: semantics} which means that they behave like this:
4046:
4047: @itemize @bullet
4048: @item
4049: The @dfn{interpretation semantics} of the word are to do something useful.
4050: @item
4051: The @dfn{compilation semantics} of the word are to append its
4052: @dfn{interpretation semantics} to the current definition (so that its
4053: run-time behaviour is to do something useful).
4054: @end itemize
4055:
4056: @cindex immediate words
4057: The actual behaviour of any particular word can be controlled by using
4058: the words @code{immediate} and @code{compile-only} when the word is
4059: defined. These words set flags in the name dictionary entry of the most
4060: recently defined word, and these flags are retrieved by the text
4061: interpreter when it finds the word in the name dictionary.
4062:
4063: A word that is marked as @dfn{immediate} has compilation semantics that
4064: are identical to its interpretation semantics. In other words, it
4065: behaves like this:
4066:
4067: @itemize @bullet
4068: @item
4069: The @dfn{interpretation semantics} of the word are to do something useful.
4070: @item
4071: The @dfn{compilation semantics} of the word are to do something useful
4072: (and actually the same thing); i.e., it is executed during compilation.
4073: @end itemize
4074:
4075: Marking a word as @dfn{compile-only} prohibits the text interpreter from
4076: performing the interpretation semantics of the word directly; an attempt
4077: to do so will generate an error. It is never necessary to use
4078: @code{compile-only} (and it is not even part of ANS Forth, though it is
4079: provided by many implementations) but it is good etiquette to apply it
4080: to a word that will not behave correctly (and might have unexpected
4081: side-effects) in interpret state. For example, it is only legal to use
4082: the conditional word @code{IF} within a definition. If you forget this
4083: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4084: @code{compile-only} allows the text interpreter to generate a helpful
4085: error message rather than subjecting you to the consequences of your
4086: folly.
4087:
4088: This example shows the difference between an immediate and a
4089: non-immediate word:
4090:
4091: @example
4092: : show-state state @@ . ;
4093: : show-state-now show-state ; immediate
4094: : word1 show-state ;
4095: : word2 show-state-now ;
4096: @end example
4097:
4098: The word @code{immediate} after the definition of @code{show-state-now}
4099: makes that word an immediate word. These definitions introduce a new
4100: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4101: variable, and leaves it on the stack. Therefore, the behaviour of
4102: @code{show-state} is to print a number that represents the current value
4103: of @code{state}.
4104:
4105: When you execute @code{word1}, it prints the number 0, indicating that
4106: the system is interpreting. When the text interpreter compiled the
4107: definition of @code{word1}, it encountered @code{show-state} whose
4108: compilation semantics are to append its interpretation semantics to the
4109: current definition. When you execute @code{word1}, it performs the
4110: interpretation semantics of @code{show-state}. At the time that @code{word1}
4111: (and therefore @code{show-state}) are executed, the system is
4112: interpreting.
4113:
4114: When you pressed @key{RET} after entering the definition of @code{word2},
4115: you should have seen the number -1 printed, followed by ``@code{
4116: ok}''. When the text interpreter compiled the definition of
4117: @code{word2}, it encountered @code{show-state-now}, an immediate word,
4118: whose compilation semantics are therefore to perform its interpretation
4119: semantics. It is executed straight away (even before the text
4120: interpreter has moved on to process another group of characters; the
4121: @code{;} in this example). The effect of executing it are to display the
4122: value of @code{state} @i{at the time that the definition of}
4123: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4124: system is compiling at this time. If you execute @code{word2} it does
4125: nothing at all.
4126:
4127: @cindex @code{."}, how it works
4128: Before leaving the subject of immediate words, consider the behaviour of
4129: @code{."} in the definition of @code{greet}, in the previous
4130: section. This word is both a parsing word and an immediate word. Notice
4131: that there is a space between @code{."} and the start of the text
4132: @code{Hello and welcome}, but that there is no space between the last
4133: letter of @code{welcome} and the @code{"} character. The reason for this
4134: is that @code{."} is a Forth word; it must have a space after it so that
4135: the text interpreter can identify it. The @code{"} is not a Forth word;
4136: it is a @dfn{delimiter}. The examples earlier show that, when the string
4137: is displayed, there is neither a space before the @code{H} nor after the
4138: @code{e}. Since @code{."} is an immediate word, it executes at the time
4139: that @code{greet} is defined. When it executes, its behaviour is to
4140: search forward in the input line looking for the delimiter. When it
4141: finds the delimiter, it updates @code{>IN} to point past the
4142: delimiter. It also compiles some magic code into the definition of
4143: @code{greet}; the xt of a run-time routine that prints a text string. It
4144: compiles the string @code{Hello and welcome} into memory so that it is
4145: available to be printed later. When the text interpreter gains control,
4146: the next word it finds in the input stream is @code{;} and so it
4147: terminates the definition of @code{greet}.
4148:
4149:
4150: @comment ----------------------------------------------
4151: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4152: @section Forth is written in Forth
4153: @cindex structure of Forth programs
4154:
4155: When you start up a Forth compiler, a large number of definitions
4156: already exist. In Forth, you develop a new application using bottom-up
4157: programming techniques to create new definitions that are defined in
4158: terms of existing definitions. As you create each definition you can
4159: test and debug it interactively.
4160:
4161: If you have tried out the examples in this section, you will probably
4162: have typed them in by hand; when you leave Gforth, your definitions will
4163: be lost. You can avoid this by using a text editor to enter Forth source
4164: code into a file, and then loading code from the file using
4165: @code{include} (@pxref{Forth source files}). A Forth source file is
4166: processed by the text interpreter, just as though you had typed it in by
4167: hand@footnote{Actually, there are some subtle differences -- see
4168: @ref{The Text Interpreter}.}.
4169:
4170: Gforth also supports the traditional Forth alternative to using text
4171: files for program entry (@pxref{Blocks}).
4172:
4173: In common with many, if not most, Forth compilers, most of Gforth is
4174: actually written in Forth. All of the @file{.fs} files in the
4175: installation directory@footnote{For example,
4176: @file{/usr/local/share/gforth...}} are Forth source files, which you can
4177: study to see examples of Forth programming.
4178:
4179: Gforth maintains a history file that records every line that you type to
4180: the text interpreter. This file is preserved between sessions, and is
4181: used to provide a command-line recall facility. If you enter long
4182: definitions by hand, you can use a text editor to paste them out of the
4183: history file into a Forth source file for reuse at a later time
4184: (for more information @pxref{Command-line editing}).
4185:
4186:
4187: @comment ----------------------------------------------
4188: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4189: @section Review - elements of a Forth system
4190: @cindex elements of a Forth system
4191:
4192: To summarise this chapter:
4193:
4194: @itemize @bullet
4195: @item
4196: Forth programs use @dfn{factoring} to break a problem down into small
4197: fragments called @dfn{words} or @dfn{definitions}.
4198: @item
4199: Forth program development is an interactive process.
4200: @item
4201: The main command loop that accepts input, and controls both
4202: interpretation and compilation, is called the @dfn{text interpreter}
4203: (also known as the @dfn{outer interpreter}).
4204: @item
4205: Forth has a very simple syntax, consisting of words and numbers
4206: separated by spaces or carriage-return characters. Any additional syntax
4207: is imposed by @dfn{parsing words}.
4208: @item
4209: Forth uses a stack to pass parameters between words. As a result, it
4210: uses postfix notation.
4211: @item
4212: To use a word that has previously been defined, the text interpreter
4213: searches for the word in the @dfn{name dictionary}.
4214: @item
4215: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
4216: @item
4217: The text interpreter uses the value of @code{state} to select between
4218: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4219: semantics} of a word that it encounters.
4220: @item
4221: The relationship between the @dfn{interpretation semantics} and
4222: @dfn{compilation semantics} for a word
4223: depend upon the way in which the word was defined (for example, whether
4224: it is an @dfn{immediate} word).
4225: @item
4226: Forth definitions can be implemented in Forth (called @dfn{high-level
4227: definitions}) or in some other way (usually a lower-level language and
4228: as a result often called @dfn{low-level definitions}, @dfn{code
4229: definitions} or @dfn{primitives}).
4230: @item
4231: Many Forth systems are implemented mainly in Forth.
4232: @end itemize
4233:
4234:
4235: @comment ----------------------------------------------
4236: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
4237: @section Where To Go Next
4238: @cindex where to go next
4239:
4240: Amazing as it may seem, if you have read (and understood) this far, you
4241: know almost all the fundamentals about the inner workings of a Forth
4242: system. You certainly know enough to be able to read and understand the
4243: rest of this manual and the ANS Forth document, to learn more about the
4244: facilities that Forth in general and Gforth in particular provide. Even
4245: scarier, you know almost enough to implement your own Forth system.
4246: However, that's not a good idea just yet... better to try writing some
4247: programs in Gforth.
4248:
4249: Forth has such a rich vocabulary that it can be hard to know where to
4250: start in learning it. This section suggests a few sets of words that are
4251: enough to write small but useful programs. Use the word index in this
4252: document to learn more about each word, then try it out and try to write
4253: small definitions using it. Start by experimenting with these words:
4254:
4255: @itemize @bullet
4256: @item
4257: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4258: @item
4259: Comparison: @code{MIN MAX =}
4260: @item
4261: Logic: @code{AND OR XOR NOT}
4262: @item
4263: Stack manipulation: @code{DUP DROP SWAP OVER}
4264: @item
4265: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
4266: @item
4267: Input/Output: @code{. ." EMIT CR KEY}
4268: @item
4269: Defining words: @code{: ; CREATE}
4270: @item
4271: Memory allocation words: @code{ALLOT ,}
4272: @item
4273: Tools: @code{SEE WORDS .S MARKER}
4274: @end itemize
4275:
4276: When you have mastered those, go on to:
4277:
4278: @itemize @bullet
4279: @item
4280: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
4281: @item
4282: Memory access: @code{@@ !}
4283: @end itemize
4284:
4285: When you have mastered these, there's nothing for it but to read through
4286: the whole of this manual and find out what you've missed.
4287:
4288: @comment ----------------------------------------------
4289: @node Exercises, , Where to go next, Introduction
4290: @section Exercises
4291: @cindex exercises
4292:
4293: TODO: provide a set of programming excercises linked into the stuff done
4294: already and into other sections of the manual. Provide solutions to all
4295: the exercises in a .fs file in the distribution.
4296:
4297: @c Get some inspiration from Starting Forth and Kelly&Spies.
4298:
4299: @c excercises:
4300: @c 1. take inches and convert to feet and inches.
4301: @c 2. take temperature and convert from fahrenheight to celcius;
4302: @c may need to care about symmetric vs floored??
4303: @c 3. take input line and do character substitution
4304: @c to encipher or decipher
4305: @c 4. as above but work on a file for in and out
4306: @c 5. take input line and convert to pig-latin
4307: @c
4308: @c thing of sets of things to exercise then come up with
4309: @c problems that need those things.
4310:
4311:
4312: @c ******************************************************************
4313: @node Words, Error messages, Introduction, Top
4314: @chapter Forth Words
4315: @cindex words
4316:
4317: @menu
4318: * Notation::
4319: * Case insensitivity::
4320: * Comments::
4321: * Boolean Flags::
4322: * Arithmetic::
4323: * Stack Manipulation::
4324: * Memory::
4325: * Control Structures::
4326: * Defining Words::
4327: * Interpretation and Compilation Semantics::
4328: * Tokens for Words::
4329: * The Text Interpreter::
4330: * Word Lists::
4331: * Environmental Queries::
4332: * Files::
4333: * Blocks::
4334: * Other I/O::
4335: * Programming Tools::
4336: * Assembler and Code Words::
4337: * Threading Words::
4338: * Locals::
4339: * Structures::
4340: * Object-oriented Forth::
4341: * Passing Commands to the OS::
4342: * Keeping track of Time::
4343: * Miscellaneous Words::
4344: @end menu
4345:
4346: @node Notation, Case insensitivity, Words, Words
4347: @section Notation
4348: @cindex notation of glossary entries
4349: @cindex format of glossary entries
4350: @cindex glossary notation format
4351: @cindex word glossary entry format
4352:
4353: The Forth words are described in this section in the glossary notation
4354: that has become a de-facto standard for Forth texts, i.e.,
4355:
4356: @format
4357: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
4358: @end format
4359: @i{Description}
4360:
4361: @table @var
4362: @item word
4363: The name of the word.
4364:
4365: @item Stack effect
4366: @cindex stack effect
4367: The stack effect is written in the notation @code{@i{before} --
4368: @i{after}}, where @i{before} and @i{after} describe the top of
4369: stack entries before and after the execution of the word. The rest of
4370: the stack is not touched by the word. The top of stack is rightmost,
4371: i.e., a stack sequence is written as it is typed in. Note that Gforth
4372: uses a separate floating point stack, but a unified stack
4373: notation. Also, return stack effects are not shown in @i{stack
4374: effect}, but in @i{Description}. The name of a stack item describes
4375: the type and/or the function of the item. See below for a discussion of
4376: the types.
4377:
4378: All words have two stack effects: A compile-time stack effect and a
4379: run-time stack effect. The compile-time stack-effect of most words is
4380: @i{ -- }. If the compile-time stack-effect of a word deviates from
4381: this standard behaviour, or the word does other unusual things at
4382: compile time, both stack effects are shown; otherwise only the run-time
4383: stack effect is shown.
4384:
4385: @cindex pronounciation of words
4386: @item pronunciation
4387: How the word is pronounced.
4388:
4389: @cindex wordset
4390: @item wordset
4391: The ANS Forth standard is divided into several word sets. A standard
4392: system need not support all of them. Therefore, in theory, the fewer
4393: word sets your program uses the more portable it will be. However, we
4394: suspect that most ANS Forth systems on personal machines will feature
4395: all word sets. Words that are not defined in ANS Forth have
4396: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
4397: describes words that will work in future releases of Gforth;
4398: @code{gforth-internal} words are more volatile. Environmental query
4399: strings are also displayed like words; you can recognize them by the
4400: @code{environment} in the word set field.
4401:
4402: @item Description
4403: A description of the behaviour of the word.
4404: @end table
4405:
4406: @cindex types of stack items
4407: @cindex stack item types
4408: The type of a stack item is specified by the character(s) the name
4409: starts with:
4410:
4411: @table @code
4412: @item f
4413: @cindex @code{f}, stack item type
4414: Boolean flags, i.e. @code{false} or @code{true}.
4415: @item c
4416: @cindex @code{c}, stack item type
4417: Char
4418: @item w
4419: @cindex @code{w}, stack item type
4420: Cell, can contain an integer or an address
4421: @item n
4422: @cindex @code{n}, stack item type
4423: signed integer
4424: @item u
4425: @cindex @code{u}, stack item type
4426: unsigned integer
4427: @item d
4428: @cindex @code{d}, stack item type
4429: double sized signed integer
4430: @item ud
4431: @cindex @code{ud}, stack item type
4432: double sized unsigned integer
4433: @item r
4434: @cindex @code{r}, stack item type
4435: Float (on the FP stack)
4436: @item a-
4437: @cindex @code{a_}, stack item type
4438: Cell-aligned address
4439: @item c-
4440: @cindex @code{c_}, stack item type
4441: Char-aligned address (note that a Char may have two bytes in Windows NT)
4442: @item f-
4443: @cindex @code{f_}, stack item type
4444: Float-aligned address
4445: @item df-
4446: @cindex @code{df_}, stack item type
4447: Address aligned for IEEE double precision float
4448: @item sf-
4449: @cindex @code{sf_}, stack item type
4450: Address aligned for IEEE single precision float
4451: @item xt
4452: @cindex @code{xt}, stack item type
4453: Execution token, same size as Cell
4454: @item wid
4455: @cindex @code{wid}, stack item type
4456: Word list ID, same size as Cell
4457: @item f83name
4458: @cindex @code{f83name}, stack item type
4459: Pointer to a name structure
4460: @item "
4461: @cindex @code{"}, stack item type
4462: string in the input stream (not on the stack). The terminating character
4463: is a blank by default. If it is not a blank, it is shown in @code{<>}
4464: quotes.
4465: @end table
4466:
4467: @comment ----------------------------------------------
4468: @node Case insensitivity, Comments, Notation, Words
4469: @section Case insensitivity
4470: @cindex case sensitivity
4471: @cindex upper and lower case
4472:
4473: Gforth is case-insensitive; you can enter definitions and invoke
4474: Standard words using upper, lower or mixed case (however,
4475: @pxref{core-idef, Implementation-defined options, Implementation-defined
4476: options}).
4477:
4478: ANS Forth only @i{requires} implementations to recognise Standard words
4479: when they are typed entirely in upper case. Therefore, a Standard
4480: program must use upper case for all Standard words. You can use whatever
4481: case you like for words that you define, but in a Standard program you
4482: have to use the words in the same case that you defined them.
4483:
4484: Gforth supports case sensitivity through @code{table}s (case-sensitive
4485: wordlists, @pxref{Word Lists}).
4486:
4487: Two people have asked how to convert Gforth to be case-sensitive; while
4488: we think this is a bad idea, you can change all wordlists into tables
4489: like this:
4490:
4491: @example
4492: ' table-find forth-wordlist wordlist-map @ !
4493: @end example
4494:
4495: Note that you now have to type the predefined words in the same case
4496: that we defined them, which are varying. You may want to convert them
4497: to your favourite case before doing this operation (I won't explain how,
4498: because if you are even contemplating doing this, you'd better have
4499: enough knowledge of Forth systems to know this already).
4500:
4501: @node Comments, Boolean Flags, Case insensitivity, Words
4502: @section Comments
4503: @cindex comments
4504:
4505: Forth supports two styles of comment; the traditional @i{in-line} comment,
4506: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
4507:
4508:
4509: doc-(
4510: doc-\
4511: doc-\G
4512:
4513:
4514: @node Boolean Flags, Arithmetic, Comments, Words
4515: @section Boolean Flags
4516: @cindex Boolean flags
4517:
4518: A Boolean flag is cell-sized. A cell with all bits clear represents the
4519: flag @code{false} and a flag with all bits set represents the flag
4520: @code{true}. Words that check a flag (for example, @code{IF}) will treat
4521: a cell that has @i{any} bit set as @code{true}.
4522:
4523:
4524: doc-true
4525: doc-false
4526: doc-on
4527: doc-off
4528:
4529:
4530: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
4531: @section Arithmetic
4532: @cindex arithmetic words
4533:
4534: @cindex division with potentially negative operands
4535: Forth arithmetic is not checked, i.e., you will not hear about integer
4536: overflow on addition or multiplication, you may hear about division by
4537: zero if you are lucky. The operator is written after the operands, but
4538: the operands are still in the original order. I.e., the infix @code{2-1}
4539: corresponds to @code{2 1 -}. Forth offers a variety of division
4540: operators. If you perform division with potentially negative operands,
4541: you do not want to use @code{/} or @code{/mod} with its undefined
4542: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4543: former, @pxref{Mixed precision}).
4544: @comment TODO discuss the different division forms and the std approach
4545:
4546: @menu
4547: * Single precision::
4548: * Bitwise operations::
4549: * Double precision:: Double-cell integer arithmetic
4550: * Numeric comparison::
4551: * Mixed precision:: Operations with single and double-cell integers
4552: * Floating Point::
4553: @end menu
4554:
4555: @node Single precision, Bitwise operations, Arithmetic, Arithmetic
4556: @subsection Single precision
4557: @cindex single precision arithmetic words
4558:
4559: By default, numbers in Forth are single-precision integers that are 1
4560: cell in size. They can be signed or unsigned, depending upon how you
4561: treat them. For the rules used by the text interpreter for recognising
4562: single-precision integers see @ref{Number Conversion}.
4563:
4564:
4565: doc-+
4566: doc-1+
4567: doc--
4568: doc-1-
4569: doc-*
4570: doc-/
4571: doc-mod
4572: doc-/mod
4573: doc-negate
4574: doc-abs
4575: doc-min
4576: doc-max
4577: doc-d>s
4578: doc-floored
4579:
4580:
4581: @node Bitwise operations, Double precision, Single precision, Arithmetic
4582: @subsection Bitwise operations
4583: @cindex bitwise operation words
4584:
4585:
4586: doc-and
4587: doc-or
4588: doc-xor
4589: doc-invert
4590: doc-lshift
4591: doc-rshift
4592: doc-2*
4593: doc-d2*
4594: doc-2/
4595: doc-d2/
4596:
4597:
4598: @node Double precision, Numeric comparison, Bitwise operations, Arithmetic
4599: @subsection Double precision
4600: @cindex double precision arithmetic words
4601:
4602: For the rules used by the text interpreter for
4603: recognising double-precision integers, see @ref{Number Conversion}.
4604:
4605: A double precision number is represented by a cell pair, with the most
4606: significant cell at the TOS. It is trivial to convert an unsigned
4607: single to an (unsigned) double; simply push a @code{0} onto the
4608: TOS. Since numbers are represented by Gforth using 2's complement
4609: arithmetic, converting a signed single to a (signed) double requires
4610: sign-extension across the most significant cell. This can be achieved
4611: using @code{s>d}. The moral of the story is that you cannot convert a
4612: number without knowing whether it represents an unsigned or a
4613: signed number.
4614:
4615:
4616: doc-s>d
4617: doc-d+
4618: doc-d-
4619: doc-dnegate
4620: doc-dabs
4621: doc-dmin
4622: doc-dmax
4623:
4624:
4625: @node Numeric comparison, Mixed precision, Double precision, Arithmetic
4626: @subsection Numeric comparison
4627: @cindex numeric comparison words
4628:
4629:
4630: doc-<
4631: doc-<=
4632: doc-<>
4633: doc-=
4634: doc->
4635: doc->=
4636:
4637: doc-0<
4638: doc-0<=
4639: doc-0<>
4640: doc-0=
4641: doc-0>
4642: doc-0>=
4643:
4644: doc-u<
4645: doc-u<=
4646: @c u<> and u= exist but are the same as <> and =
4647: @c doc-u<>
4648: @c doc-u=
4649: doc-u>
4650: doc-u>=
4651:
4652: doc-within
4653:
4654: doc-d<
4655: doc-d<=
4656: doc-d<>
4657: doc-d=
4658: doc-d>
4659: doc-d>=
4660:
4661: doc-d0<
4662: doc-d0<=
4663: doc-d0<>
4664: doc-d0=
4665: doc-d0>
4666: doc-d0>=
4667:
4668: doc-du<
4669: doc-du<=
4670: @c du<> and du= exist but are the same as d<> and d=
4671: @c doc-du<>
4672: @c doc-du=
4673: doc-du>
4674: doc-du>=
4675:
4676:
4677: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
4678: @subsection Mixed precision
4679: @cindex mixed precision arithmetic words
4680:
4681:
4682: doc-m+
4683: doc-*/
4684: doc-*/mod
4685: doc-m*
4686: doc-um*
4687: doc-m*/
4688: doc-um/mod
4689: doc-fm/mod
4690: doc-sm/rem
4691:
4692:
4693: @node Floating Point, , Mixed precision, Arithmetic
4694: @subsection Floating Point
4695: @cindex floating point arithmetic words
4696:
4697: For the rules used by the text interpreter for
4698: recognising floating-point numbers see @ref{Number Conversion}.
4699:
4700: Gforth has a separate floating point
4701: stack, but the documentation uses the unified notation.
4702:
4703: @cindex floating-point arithmetic, pitfalls
4704: Floating point numbers have a number of unpleasant surprises for the
4705: unwary (e.g., floating point addition is not associative) and even a few
4706: for the wary. You should not use them unless you know what you are doing
4707: or you don't care that the results you get are totally bogus. If you
4708: want to learn about the problems of floating point numbers (and how to
4709: avoid them), you might start with @cite{David Goldberg,
4710: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
4711: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
4712: Surveys 23(1):5@minus{}48, March 1991}.
4713:
4714:
4715: doc-d>f
4716: doc-f>d
4717: doc-f+
4718: doc-f-
4719: doc-f*
4720: doc-f/
4721: doc-fnegate
4722: doc-fabs
4723: doc-fmax
4724: doc-fmin
4725: doc-floor
4726: doc-fround
4727: doc-f**
4728: doc-fsqrt
4729: doc-fexp
4730: doc-fexpm1
4731: doc-fln
4732: doc-flnp1
4733: doc-flog
4734: doc-falog
4735: doc-f2*
4736: doc-f2/
4737: doc-1/f
4738: doc-precision
4739: doc-set-precision
4740:
4741: @cindex angles in trigonometric operations
4742: @cindex trigonometric operations
4743: Angles in floating point operations are given in radians (a full circle
4744: has 2 pi radians).
4745:
4746: doc-fsin
4747: doc-fcos
4748: doc-fsincos
4749: doc-ftan
4750: doc-fasin
4751: doc-facos
4752: doc-fatan
4753: doc-fatan2
4754: doc-fsinh
4755: doc-fcosh
4756: doc-ftanh
4757: doc-fasinh
4758: doc-facosh
4759: doc-fatanh
4760: doc-pi
4761:
4762: @cindex equality of floats
4763: @cindex floating-point comparisons
4764: One particular problem with floating-point arithmetic is that comparison
4765: for equality often fails when you would expect it to succeed. For this
4766: reason approximate equality is often preferred (but you still have to
4767: know what you are doing). The comparison words are:
4768:
4769: doc-f~rel
4770: doc-f~abs
4771: doc-f=
4772: doc-f~
4773: doc-f<>
4774:
4775: doc-f<
4776: doc-f<=
4777: doc-f>
4778: doc-f>=
4779:
4780: doc-f0<
4781: doc-f0<=
4782: doc-f0<>
4783: doc-f0=
4784: doc-f0>
4785: doc-f0>=
4786:
4787:
4788: @node Stack Manipulation, Memory, Arithmetic, Words
4789: @section Stack Manipulation
4790: @cindex stack manipulation words
4791:
4792: @cindex floating-point stack in the standard
4793: Gforth maintains a number of separate stacks:
4794:
4795: @cindex data stack
4796: @cindex parameter stack
4797: @itemize @bullet
4798: @item
4799: A data stack (also known as the @dfn{parameter stack}) -- for
4800: characters, cells, addresses, and double cells.
4801:
4802: @cindex floating-point stack
4803: @item
4804: A floating point stack -- for holding floating point (FP) numbers.
4805:
4806: @cindex return stack
4807: @item
4808: A return stack -- for holding the return addresses of colon
4809: definitions and other (non-FP) data.
4810:
4811: @cindex locals stack
4812: @item
4813: A locals stack -- for holding local variables.
4814: @end itemize
4815:
4816: @menu
4817: * Data stack::
4818: * Floating point stack::
4819: * Return stack::
4820: * Locals stack::
4821: * Stack pointer manipulation::
4822: @end menu
4823:
4824: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
4825: @subsection Data stack
4826: @cindex data stack manipulation words
4827: @cindex stack manipulations words, data stack
4828:
4829:
4830: doc-drop
4831: doc-nip
4832: doc-dup
4833: doc-over
4834: doc-tuck
4835: doc-swap
4836: doc-pick
4837: doc-rot
4838: doc--rot
4839: doc-?dup
4840: doc-roll
4841: doc-2drop
4842: doc-2nip
4843: doc-2dup
4844: doc-2over
4845: doc-2tuck
4846: doc-2swap
4847: doc-2rot
4848:
4849:
4850: @node Floating point stack, Return stack, Data stack, Stack Manipulation
4851: @subsection Floating point stack
4852: @cindex floating-point stack manipulation words
4853: @cindex stack manipulation words, floating-point stack
4854:
4855: Whilst every sane Forth has a separate floating-point stack, it is not
4856: strictly required; an ANS Forth system could theoretically keep
4857: floating-point numbers on the data stack. As an additional difficulty,
4858: you don't know how many cells a floating-point number takes. It is
4859: reportedly possible to write words in a way that they work also for a
4860: unified stack model, but we do not recommend trying it. Instead, just
4861: say that your program has an environmental dependency on a separate
4862: floating-point stack.
4863:
4864: doc-floating-stack
4865:
4866: doc-fdrop
4867: doc-fnip
4868: doc-fdup
4869: doc-fover
4870: doc-ftuck
4871: doc-fswap
4872: doc-fpick
4873: doc-frot
4874:
4875:
4876: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
4877: @subsection Return stack
4878: @cindex return stack manipulation words
4879: @cindex stack manipulation words, return stack
4880:
4881: @cindex return stack and locals
4882: @cindex locals and return stack
4883: A Forth system is allowed to keep local variables on the
4884: return stack. This is reasonable, as local variables usually eliminate
4885: the need to use the return stack explicitly. So, if you want to produce
4886: a standard compliant program and you are using local variables in a
4887: word, forget about return stack manipulations in that word (refer to the
4888: standard document for the exact rules).
4889:
4890: doc->r
4891: doc-r>
4892: doc-r@
4893: doc-rdrop
4894: doc-2>r
4895: doc-2r>
4896: doc-2r@
4897: doc-2rdrop
4898:
4899:
4900: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
4901: @subsection Locals stack
4902:
4903: Gforth uses an extra locals stack. It is described, along with the
4904: reasons for its existence, in @ref{Implementation,Implementation of locals}.
4905:
4906: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
4907: @subsection Stack pointer manipulation
4908: @cindex stack pointer manipulation words
4909:
4910: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
4911: doc-sp0
4912: doc-sp@
4913: doc-sp!
4914: doc-fp0
4915: doc-fp@
4916: doc-fp!
4917: doc-rp0
4918: doc-rp@
4919: doc-rp!
4920: doc-lp0
4921: doc-lp@
4922: doc-lp!
4923:
4924:
4925: @node Memory, Control Structures, Stack Manipulation, Words
4926: @section Memory
4927: @cindex memory words
4928:
4929: @menu
4930: * Memory model::
4931: * Dictionary allocation::
4932: * Heap Allocation::
4933: * Memory Access::
4934: * Address arithmetic::
4935: * Memory Blocks::
4936: @end menu
4937:
4938: @node Memory model, Dictionary allocation, Memory, Memory
4939: @subsection ANS Forth and Gforth memory models
4940:
4941: @c The ANS Forth description is a mess (e.g., is the heap part of
4942: @c the dictionary?), so let's not stick to closely with it.
4943:
4944: ANS Forth considers a Forth system as consisting of several memories, of
4945: which only @dfn{data space} is managed and accessible with the memory
4946: words. Memory not necessarily in data space includes the stacks, the
4947: code (called code space) and the headers (called name space). In Gforth
4948: everything is in data space, but the code for the primitives is usually
4949: read-only.
4950:
4951: Data space is divided into a number of areas: The (data space portion of
4952: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
4953: refer to the search data structure embodied in word lists and headers,
4954: because it is used for looking up names, just as you would in a
4955: conventional dictionary.}, the heap, and a number of system-allocated
4956: buffers.
4957:
4958: In ANS Forth data space is also divided into contiguous regions. You
4959: can only use address arithmetic within a contiguous region, not between
4960: them. Usually each allocation gives you one contiguous region, but the
4961: dictionary allocation words have additional rules (@pxref{Dictionary
4962: allocation}).
4963:
4964: Gforth provides one big address space, and address arithmetic can be
4965: performed between any addresses. However, in the dictionary headers or
4966: code are interleaved with data, so almost the only contiguous data space
4967: regions there are those described by ANS Forth as contiguous; but you
4968: can be sure that the dictionary is allocated towards increasing
4969: addresses even between contiguous regions. The memory order of
4970: allocations in the heap is platform-dependent (and possibly different
4971: from one run to the next).
4972:
4973: @subsubsection ANS Forth dictionary details
4974:
4975: This section is just informative, you can skip it if you are in a hurry.
4976:
4977: When you create a colon definition, the text interpreter compiles the
4978: code for the definition into the code space and compiles the name
4979: of the definition into the header space, together with other
4980: information about the definition (such as its execution token).
4981:
4982: When you create a variable, the execution of @code{Variable} will
4983: compile some code, assign one cell in data space, and compile the name
4984: of the variable into the header space.
4985:
4986: @cindex memory regions - relationship between them
4987: ANS Forth does not specify the relationship between the three memory
4988: regions, and specifies that a Standard program must not access code or
4989: data space directly -- it may only access data space directly. In
4990: addition, the Standard defines what relationships you may and may not
4991: rely on when allocating regions in data space. These constraints are
4992: simply a reflection of the many diverse techniques that are used to
4993: implement Forth systems; understanding and following the requirements of
4994: the Standard allows you to write portable programs -- programs that run
4995: in the same way on any of these diverse systems. Another way of looking
4996: at this is to say that ANS Forth was designed to permit compliant Forth
4997: systems to be implemented in many diverse ways.
4998:
4999: @cindex memory regions - how they are assigned
5000: Here are some examples of ways in which name, code and data spaces
5001: might be assigned in different Forth implementations:
5002:
5003: @itemize @bullet
5004: @item
5005: For a Forth system that runs from RAM under a general-purpose operating
5006: system, it can be convenient to interleave name, code and data spaces in
5007: a single contiguous memory region. This organisation can be
5008: memory-efficient (for example, because the relationship between the name
5009: dictionary entry and the associated code space entry can be
5010: implicit, rather than requiring an explicit memory pointer to reference
5011: from the header space and the code space). This is the
5012: organisation used by Gforth, as this example@footnote{The addresses
5013: in the example have been truncated to fit it onto the page, and the
5014: addresses and data shown will not match the output from your system} shows:
5015: @example
5016: hex
5017: variable fred 123456 fred !
5018: variable jim abcd jim !
5019: : foo + / - ;
5020: ' fred 10 - 50 dump
5021: ..80: 5C 46 0E 40 84 66 72 65 - 64 20 20 20 20 20 20 20 \F.@.fred
5022: ..90: D0 9B 04 08 00 00 00 00 - 56 34 12 00 80 46 0E 40 ........V4...F.@@
5023: ..A0: 83 6A 69 6D 20 20 20 20 - D0 9B 04 08 00 00 00 00 .jim ........
5024: ..B0: CD AB 00 00 9C 46 0E 40 - 83 66 6F 6F 20 20 20 20 .....F.@.foo
5025: ..C0: 80 9B 04 08 00 00 00 00 - E4 2E 05 08 0C 2F 05 08 ............./..
5026: @end example
5027:
5028: @item
5029: For a high-performance system running on a modern RISC processor with a
5030: modified Harvard architecture (one that has a unified main memory but
5031: separate instruction and data caches), it is desirable to separate
5032: processor instructions from processor data. This encourages a high cache
5033: density and therefore a high cache hit rate. The Forth code space
5034: is not necessarily made up entirely of processor instructions; its
5035: nature is dependent upon the Forth implementation.
5036:
5037: @item
5038: A Forth compiler that runs on a segmented 8086 processor could be
5039: designed to interleave the name, code and data spaces within a single
5040: 64Kbyte segment. A more common implementation choice is to use a
5041: separate 64Kbyte segment for each region, which provides more memory
5042: overall but provides an address map in which only the data space is
5043: accessible.
5044:
5045: @item
5046: Microprocessors exist that run Forth (or many of the primitives required
5047: to implement the Forth virtual machine efficiently) directly. On these
5048: processors, the relationship between name, code and data spaces may be
5049: imposed as a side-effect of the architecture of the processor.
5050:
5051: @item
5052: A Forth compiler that executes from ROM on an embedded system needs its
5053: data space separated from the name and code spaces so that the data
5054: space can be mapped to a RAM area.
5055:
5056: @item
5057: A Forth compiler that runs on an embedded system may have a requirement
5058: for a small memory footprint. On such a system it can be useful to
5059: separate the header space from the data and code spaces; once the
5060: application has been compiled, the header space is no longer
5061: required@footnote{more strictly speaking, most applications can be
5062: designed so that this is the case}. The header space can be deleted
5063: entirely, or could be stored in memory on a remote @i{host} system for
5064: debug and development purposes. In the latter case, the compiler running
5065: on the @i{target} system could implement a protocol across a
5066: communication link that would allow it to interrogate the header space.
5067: @end itemize
5068:
5069: @node Dictionary allocation, Heap Allocation, Memory model, Memory
5070: @subsection Dictionary allocation
5071: @cindex reserving data space
5072: @cindex data space - reserving some
5073:
5074: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
5075: you want to deallocate X, you also deallocate everything
5076: allocated after X.
5077:
5078: The allocations using the words below are contiguous and grow the region
5079: towards increasing addresses. Other words that allocate dictionary
5080: memory of any kind (i.e., defining words including @code{:noname}) end
5081: the contiguous region and start a new one.
5082:
5083: In ANS Forth only @code{create}d words are guaranteed to produce an
5084: address that is the start of the following contiguous region. In
5085: particular, the cell allocated by @code{variable} is not guaranteed to
5086: be contiguous with following @code{allot}ed memory.
5087:
5088: You can deallocate memory by using @code{allot} with a negative argument
5089: (with some restrictions, see @code{allot}). For larger deallocations use
5090: @code{marker}.
5091:
5092:
5093: doc-here
5094: doc-unused
5095: doc-allot
5096: doc-c,
5097: doc-f,
5098: doc-,
5099: doc-2,
5100: @cindex user space
5101: doc-udp
5102: doc-uallot
5103:
5104: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
5105: course you should allocate memory in an aligned way, too. I.e., before
5106: allocating allocating a cell, @code{here} must be cell-aligned, etc.
5107: The words below align @code{here} if it is not already. Basically it is
5108: only already aligned for a type, if the last allocation was a multiple
5109: of the size of this type and if @code{here} was aligned for this type
5110: before.
5111:
5112: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
5113: ANS Forth (@code{maxalign}ed in Gforth).
5114:
5115: doc-align
5116: doc-falign
5117: doc-sfalign
5118: doc-dfalign
5119: doc-maxalign
5120: doc-cfalign
5121:
5122:
5123: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
5124: @subsection Heap allocation
5125: @cindex heap allocation
5126: @cindex dynamic allocation of memory
5127: @cindex memory-allocation word set
5128:
5129: Heap allocation supports deallocation of allocated memory in any
5130: order. Dictionary allocation is not affected by it (i.e., it does not
5131: end a contiguous region). In Gforth, these words are implemented using
5132: the standard C library calls malloc(), free() and resize().
5133:
5134: doc-allocate
5135: doc-free
5136: doc-resize
5137:
5138:
5139: @node Memory Access, Address arithmetic, Heap Allocation, Memory
5140: @subsection Memory Access
5141: @cindex memory access words
5142:
5143:
5144: doc-@
5145: doc-!
5146: doc-+!
5147: doc-c@
5148: doc-c!
5149: doc-2@
5150: doc-2!
5151: doc-f@
5152: doc-f!
5153: doc-sf@
5154: doc-sf!
5155: doc-df@
5156: doc-df!
5157:
5158: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5159: @subsection Address arithmetic
5160: @cindex address arithmetic words
5161:
5162: Address arithmetic is the foundation on which data structures like
5163: arrays, records (@pxref{Structures}) and objects (@pxref{Object-oriented
5164: Forth}) are built.
5165:
5166: ANS Forth does not specify the sizes of the data types. Instead, it
5167: offers a number of words for computing sizes and doing address
5168: arithmetic. Address arithmetic is performed in terms of address units
5169: (aus); on most systems the address unit is one byte. Note that a
5170: character may have more than one au, so @code{chars} is no noop (on
5171: systems where it is a noop, it compiles to nothing).
5172:
5173: @cindex alignment of addresses for types
5174: ANS Forth also defines words for aligning addresses for specific
5175: types. Many computers require that accesses to specific data types
5176: must only occur at specific addresses; e.g., that cells may only be
5177: accessed at addresses divisible by 4. Even if a machine allows unaligned
5178: accesses, it can usually perform aligned accesses faster.
5179:
5180: For the performance-conscious: alignment operations are usually only
5181: necessary during the definition of a data structure, not during the
5182: (more frequent) accesses to it.
5183:
5184: ANS Forth defines no words for character-aligning addresses. This is not
5185: an oversight, but reflects the fact that addresses that are not
5186: char-aligned have no use in the standard and therefore will not be
5187: created.
5188:
5189: @cindex @code{CREATE} and alignment
5190: ANS Forth guarantees that addresses returned by @code{CREATE}d words
5191: are cell-aligned; in addition, Gforth guarantees that these addresses
5192: are aligned for all purposes.
5193:
5194: Note that the ANS Forth word @code{char} has nothing to do with address
5195: arithmetic.
5196:
5197:
5198: doc-chars
5199: doc-char+
5200: doc-cells
5201: doc-cell+
5202: doc-cell
5203: doc-aligned
5204: doc-floats
5205: doc-float+
5206: doc-float
5207: doc-faligned
5208: doc-sfloats
5209: doc-sfloat+
5210: doc-sfaligned
5211: doc-dfloats
5212: doc-dfloat+
5213: doc-dfaligned
5214: doc-maxaligned
5215: doc-cfaligned
5216: doc-address-unit-bits
5217:
5218:
5219: @node Memory Blocks, , Address arithmetic, Memory
5220: @subsection Memory Blocks
5221: @cindex memory block words
5222: @cindex character strings - moving and copying
5223:
5224: Memory blocks often represent character strings; For ways of storing
5225: character strings in memory see @ref{String Formats}. For other
5226: string-processing words see @ref{Displaying characters and strings}.
5227:
5228: Some of these words work on address units. Others work on character
5229: units (increments of @code{CHAR}), and expect a @code{CHAR}-aligned
5230: address. Choose the correct operation depending upon your data type.
5231:
5232: When copying characters between overlapping memory regions, choose
5233: carefully between @code{cmove} and @code{cmove>}.
5234:
5235: You can only use any of these words @i{portably} to access data space.
5236:
5237: @comment TODO - think the naming of the arguments is wrong for move
5238: @comment well, really it seems to be the Standard that's wrong; it
5239: @comment describes MOVE as a word that requires a CELL-aligned source
5240: @comment and destination address but a xtranfer count that need not
5241: @comment be a multiple of CELL.
5242:
5243: doc-move
5244: doc-erase
5245: doc-cmove
5246: doc-cmove>
5247: doc-fill
5248: doc-blank
5249: doc-compare
5250: doc-search
5251: doc--trailing
5252: doc-/string
5253:
5254:
5255: @comment TODO examples
5256:
5257:
5258: @node Control Structures, Defining Words, Memory, Words
5259: @section Control Structures
5260: @cindex control structures
5261:
5262: Control structures in Forth cannot be used interpretively, only in a
5263: colon definition@footnote{To be precise, they have no interpretation
5264: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5265: not like this limitation, but have not seen a satisfying way around it
5266: yet, although many schemes have been proposed.
5267:
5268: @menu
5269: * Selection:: IF ... ELSE ... ENDIF
5270: * Simple Loops:: BEGIN ...
5271: * Counted Loops:: DO
5272: * Arbitrary control structures::
5273: * Calls and returns::
5274: * Exception Handling::
5275: @end menu
5276:
5277: @node Selection, Simple Loops, Control Structures, Control Structures
5278: @subsection Selection
5279: @cindex selection control structures
5280: @cindex control structures for selection
5281:
5282: @c what's the purpose of all these @i? Maybe we should define a macro
5283: @c so we can produce logical markup. - anton
5284:
5285: @c nac-> When I started working on the manual, a mixture of @i and @var
5286: @c were used inconsistently in code examples and \Glossary entries. These
5287: @c two behave differently in info format so I decided to standardize on @i.
5288: @c Logical markup would be better but texi isn't really upto it, and
5289: @c texi2html just ignores macros.
5290: @c nac02dec1999-> update: the latest texinfo release can spit out html
5291: @c and it handles macros, so we could do some logical markup. Unfortunately
5292: @c texinfo will not split html output, which would be a big pain if you
5293: @c wanted to put the document on the web, which would be nice.
5294:
5295: @cindex @code{IF} control structure
5296: @example
5297: @i{flag}
5298: IF
5299: @i{code}
5300: ENDIF
5301: @end example
5302: @noindent
5303:
5304: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5305: with any bit set represents truth) @i{code} is executed.
5306:
5307: @example
5308: @i{flag}
5309: IF
5310: @i{code1}
5311: ELSE
5312: @i{code2}
5313: ENDIF
5314: @end example
5315:
5316: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5317: executed.
5318:
5319: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5320: standard, and @code{ENDIF} is not, although it is quite popular. We
5321: recommend using @code{ENDIF}, because it is less confusing for people
5322: who also know other languages (and is not prone to reinforcing negative
5323: prejudices against Forth in these people). Adding @code{ENDIF} to a
5324: system that only supplies @code{THEN} is simple:
5325: @example
5326: : ENDIF POSTPONE THEN ; immediate
5327: @end example
5328:
5329: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5330: (adv.)} has the following meanings:
5331: @quotation
5332: ... 2b: following next after in order ... 3d: as a necessary consequence
5333: (if you were there, then you saw them).
5334: @end quotation
5335: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5336: and many other programming languages has the meaning 3d.]
5337:
5338: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
5339: you can avoid using @code{?dup}. Using these alternatives is also more
5340: efficient than using @code{?dup}. Definitions in ANS Forth
5341: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5342: @file{compat/control.fs}.
5343:
5344: @cindex @code{CASE} control structure
5345: @example
5346: @i{n}
5347: CASE
5348: @i{n1} OF @i{code1} ENDOF
5349: @i{n2} OF @i{code2} ENDOF
5350: @dots{}
5351: ENDCASE
5352: @end example
5353:
5354: Executes the first @i{codei}, where the @i{ni} is equal to
5355: @i{n}. A default case can be added by simply writing the code after
5356: the last @code{ENDOF}. It may use @i{n}, which is on top of the stack,
5357: but must not consume it.
5358:
5359: @node Simple Loops, Counted Loops, Selection, Control Structures
5360: @subsection Simple Loops
5361: @cindex simple loops
5362: @cindex loops without count
5363:
5364: @cindex @code{WHILE} loop
5365: @example
5366: BEGIN
5367: @i{code1}
5368: @i{flag}
5369: WHILE
5370: @i{code2}
5371: REPEAT
5372: @end example
5373:
5374: @i{code1} is executed and @i{flag} is computed. If it is true,
5375: @i{code2} is executed and the loop is restarted; If @i{flag} is
5376: false, execution continues after the @code{REPEAT}.
5377:
5378: @cindex @code{UNTIL} loop
5379: @example
5380: BEGIN
5381: @i{code}
5382: @i{flag}
5383: UNTIL
5384: @end example
5385:
5386: @i{code} is executed. The loop is restarted if @code{flag} is false.
5387:
5388: @cindex endless loop
5389: @cindex loops, endless
5390: @example
5391: BEGIN
5392: @i{code}
5393: AGAIN
5394: @end example
5395:
5396: This is an endless loop.
5397:
5398: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5399: @subsection Counted Loops
5400: @cindex counted loops
5401: @cindex loops, counted
5402: @cindex @code{DO} loops
5403:
5404: The basic counted loop is:
5405: @example
5406: @i{limit} @i{start}
5407: ?DO
5408: @i{body}
5409: LOOP
5410: @end example
5411:
5412: This performs one iteration for every integer, starting from @i{start}
5413: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
5414: accessed with @code{i}. For example, the loop:
5415: @example
5416: 10 0 ?DO
5417: i .
5418: LOOP
5419: @end example
5420: @noindent
5421: prints @code{0 1 2 3 4 5 6 7 8 9}
5422:
5423: The index of the innermost loop can be accessed with @code{i}, the index
5424: of the next loop with @code{j}, and the index of the third loop with
5425: @code{k}.
5426:
5427:
5428: doc-i
5429: doc-j
5430: doc-k
5431:
5432:
5433: The loop control data are kept on the return stack, so there are some
5434: restrictions on mixing return stack accesses and counted loop words. In
5435: particuler, if you put values on the return stack outside the loop, you
5436: cannot read them inside the loop@footnote{well, not in a way that is
5437: portable.}. If you put values on the return stack within a loop, you
5438: have to remove them before the end of the loop and before accessing the
5439: index of the loop.
5440:
5441: There are several variations on the counted loop:
5442:
5443: @itemize @bullet
5444: @item
5445: @code{LEAVE} leaves the innermost counted loop immediately; execution
5446: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5447:
5448: @example
5449: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5450: @end example
5451: prints @code{0 1 2 3}
5452:
5453:
5454: @item
5455: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5456: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5457: return stack so @code{EXIT} can get to its return address. For example:
5458:
5459: @example
5460: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5461: @end example
5462: prints @code{0 1 2 3}
5463:
5464:
5465: @item
5466: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
5467: (and @code{LOOP} iterates until they become equal by wrap-around
5468: arithmetic). This behaviour is usually not what you want. Therefore,
5469: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
5470: @code{?DO}), which do not enter the loop if @i{start} is greater than
5471: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
5472: unsigned loop parameters.
5473:
5474: @item
5475: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5476: the loop, independent of the loop parameters. Do not use @code{DO}, even
5477: if you know that the loop is entered in any case. Such knowledge tends
5478: to become invalid during maintenance of a program, and then the
5479: @code{DO} will make trouble.
5480:
5481: @item
5482: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5483: index by @i{n} instead of by 1. The loop is terminated when the border
5484: between @i{limit-1} and @i{limit} is crossed. E.g.:
5485:
5486: @example
5487: 4 0 +DO i . 2 +LOOP
5488: @end example
5489: @noindent
5490: prints @code{0 2}
5491:
5492: @example
5493: 4 1 +DO i . 2 +LOOP
5494: @end example
5495: @noindent
5496: prints @code{1 3}
5497:
5498:
5499: @cindex negative increment for counted loops
5500: @cindex counted loops with negative increment
5501: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
5502:
5503: @example
5504: -1 0 ?DO i . -1 +LOOP
5505: @end example
5506: @noindent
5507: prints @code{0 -1}
5508:
5509: @example
5510: 0 0 ?DO i . -1 +LOOP
5511: @end example
5512: prints nothing.
5513:
5514: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5515: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5516: index by @i{u} each iteration. The loop is terminated when the border
5517: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
5518: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5519:
5520: @example
5521: -2 0 -DO i . 1 -LOOP
5522: @end example
5523: @noindent
5524: prints @code{0 -1}
5525:
5526: @example
5527: -1 0 -DO i . 1 -LOOP
5528: @end example
5529: @noindent
5530: prints @code{0}
5531:
5532: @example
5533: 0 0 -DO i . 1 -LOOP
5534: @end example
5535: @noindent
5536: prints nothing.
5537:
5538: @end itemize
5539:
5540: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
5541: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5542: for these words that uses only standard words is provided in
5543: @file{compat/loops.fs}.
5544:
5545:
5546: @cindex @code{FOR} loops
5547: Another counted loop is:
5548: @example
5549: @i{n}
5550: FOR
5551: @i{body}
5552: NEXT
5553: @end example
5554: This is the preferred loop of native code compiler writers who are too
5555: lazy to optimize @code{?DO} loops properly. This loop structure is not
5556: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5557: @code{i} produces values starting with @i{n} and ending with 0. Other
5558: Forth systems may behave differently, even if they support @code{FOR}
5559: loops. To avoid problems, don't use @code{FOR} loops.
5560:
5561: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5562: @subsection Arbitrary control structures
5563: @cindex control structures, user-defined
5564:
5565: @cindex control-flow stack
5566: ANS Forth permits and supports using control structures in a non-nested
5567: way. Information about incomplete control structures is stored on the
5568: control-flow stack. This stack may be implemented on the Forth data
5569: stack, and this is what we have done in Gforth.
5570:
5571: @cindex @code{orig}, control-flow stack item
5572: @cindex @code{dest}, control-flow stack item
5573: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5574: entry represents a backward branch target. A few words are the basis for
5575: building any control structure possible (except control structures that
5576: need storage, like calls, coroutines, and backtracking).
5577:
5578:
5579: doc-if
5580: doc-ahead
5581: doc-then
5582: doc-begin
5583: doc-until
5584: doc-again
5585: doc-cs-pick
5586: doc-cs-roll
5587:
5588:
5589: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5590: manipulate the control-flow stack in a portable way. Without them, you
5591: would need to know how many stack items are occupied by a control-flow
5592: entry (many systems use one cell. In Gforth they currently take three,
5593: but this may change in the future).
5594:
5595: Some standard control structure words are built from these words:
5596:
5597:
5598: doc-else
5599: doc-while
5600: doc-repeat
5601:
5602:
5603: @noindent
5604: Gforth adds some more control-structure words:
5605:
5606:
5607: doc-endif
5608: doc-?dup-if
5609: doc-?dup-0=-if
5610:
5611:
5612: @noindent
5613: Counted loop words constitute a separate group of words:
5614:
5615:
5616: doc-?do
5617: doc-+do
5618: doc-u+do
5619: doc--do
5620: doc-u-do
5621: doc-do
5622: doc-for
5623: doc-loop
5624: doc-+loop
5625: doc--loop
5626: doc-next
5627: doc-leave
5628: doc-?leave
5629: doc-unloop
5630: doc-done
5631:
5632:
5633: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5634: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
5635: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5636: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5637: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5638: resolved (by using one of the loop-ending words or @code{DONE}).
5639:
5640: @noindent
5641: Another group of control structure words are:
5642:
5643:
5644: doc-case
5645: doc-endcase
5646: doc-of
5647: doc-endof
5648:
5649:
5650: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5651: @code{CS-ROLL}.
5652:
5653: @subsubsection Programming Style
5654: @cindex control structures programming style
5655: @cindex programming style, arbitrary control structures
5656:
5657: In order to ensure readability we recommend that you do not create
5658: arbitrary control structures directly, but define new control structure
5659: words for the control structure you want and use these words in your
5660: program. For example, instead of writing:
5661:
5662: @example
5663: BEGIN
5664: ...
5665: IF [ 1 CS-ROLL ]
5666: ...
5667: AGAIN THEN
5668: @end example
5669:
5670: @noindent
5671: we recommend defining control structure words, e.g.,
5672:
5673: @example
5674: : WHILE ( DEST -- ORIG DEST )
5675: POSTPONE IF
5676: 1 CS-ROLL ; immediate
5677:
5678: : REPEAT ( orig dest -- )
5679: POSTPONE AGAIN
5680: POSTPONE THEN ; immediate
5681: @end example
5682:
5683: @noindent
5684: and then using these to create the control structure:
5685:
5686: @example
5687: BEGIN
5688: ...
5689: WHILE
5690: ...
5691: REPEAT
5692: @end example
5693:
5694: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5695: @code{WHILE} are predefined, so in this example it would not be
5696: necessary to define them.
5697:
5698: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5699: @subsection Calls and returns
5700: @cindex calling a definition
5701: @cindex returning from a definition
5702:
5703: @cindex recursive definitions
5704: A definition can be called simply be writing the name of the definition
5705: to be called. Normally a definition is invisible during its own
5706: definition. If you want to write a directly recursive definition, you
5707: can use @code{recursive} to make the current definition visible, or
5708: @code{recurse} to call the current definition directly.
5709:
5710:
5711: doc-recursive
5712: doc-recurse
5713:
5714:
5715: @comment TODO add example of the two recursion methods
5716: @quotation
5717: @progstyle
5718: I prefer using @code{recursive} to @code{recurse}, because calling the
5719: definition by name is more descriptive (if the name is well-chosen) than
5720: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5721: implementation, it is much better to read (and think) ``now sort the
5722: partitions'' than to read ``now do a recursive call''.
5723: @end quotation
5724:
5725: For mutual recursion, use @code{Defer}red words, like this:
5726:
5727: @example
5728: Defer foo
5729:
5730: : bar ( ... -- ... )
5731: ... foo ... ;
5732:
5733: :noname ( ... -- ... )
5734: ... bar ... ;
5735: IS foo
5736: @end example
5737:
5738: Deferred words are discussed in more detail in @ref{Deferred words}.
5739:
5740: The current definition returns control to the calling definition when
5741: the end of the definition is reached or @code{EXIT} is encountered.
5742:
5743: doc-exit
5744: doc-;s
5745:
5746:
5747: @node Exception Handling, , Calls and returns, Control Structures
5748: @subsection Exception Handling
5749: @cindex exceptions
5750:
5751: If your program detects a fatal error condition, the simplest action
5752: that it can take is to @code{quit}. This resets the return stack and
5753: restarts the text interpreter, but does not print any error message.
5754:
5755: The next stage in severity is to execute @code{abort}, which has the
5756: same effect as @code{quit}, with the addition that it resets the data
5757: stack.
5758:
5759: A slightly more sophisticated approach is use use @code{abort"}, which
5760: compiles a string to be used as an error message and does a conditional
5761: @code{abort} at run-time. For example:
5762:
5763: @example
5764: @kbd{: checker abort" That flag was true" ." A false flag" ;@key{RET}} ok
5765: @kbd{0 checker@key{RET}} A false flag ok
5766: @kbd{1 checker@key{RET}}
5767: :1: That flag was true
5768: 1 checker
5769: ^^^^^^^
5770: $400D1648 throw
5771: $400E4660
5772: @end example
5773:
5774: These simple techniques allow a program to react to a fatal error
5775: condition, but they are not exactly user-friendly. The ANS Forth
5776: Exception word set provides the pair of words @code{throw} and
5777: @code{catch}, which can be used to provide sophisticated error-handling.
5778:
5779: @code{catch} has a similar behaviour to @code{execute}, in that it takes
5780: an @i{xt} as a parameter and starts execution of the xt. However,
5781: before passing control to the xt, @code{catch} pushes an
5782: @dfn{exception frame} onto the @dfn{exception stack}. This exception
5783: frame is used to restore the system to a known state if a detected error
5784: occurs during the execution of the xt. A typical way to use @code{catch}
5785: would be:
5786:
5787: @example
5788: ... ['] foo catch IF ...
5789: @end example
5790:
5791: @c TOS is undefined. - anton
5792:
5793: @c nac-> TODO -- I need to look at this example again.
5794:
5795: Whilst @code{foo} executes, it can call other words to any level of
5796: nesting, as usual. If @code{foo} (and all the words that it calls)
5797: execute successfully, control will ultimately pass to the word following
5798: the @code{catch}, and there will be a 0 at TOS. However, if any word
5799: detects an error, it can terminate the execution of @code{foo} by
5800: pushing a non-zero error code onto the stack and then performing a
5801: @code{throw}. The execution of @code{throw} will pass control to the
5802: word following the @code{catch}, but this time the TOS will hold the
5803: error code. Therefore, the @code{IF} in the example can be used to
5804: determine whether @code{foo} executed successfully.
5805:
5806: This simple example shows how you can use @code{throw} and @code{catch}
5807: to ``take over'' exception handling from the system:
5808: @example
5809: : my-div ['] / catch if ." DIVIDE ERROR" else ." OK.. " . then ;
5810: @end example
5811:
5812: The next example is more sophisticated and shows a multi-level
5813: @code{throw} and @code{catch}. To understand this example, start at the
5814: definition of @code{top-level} and work backwards:
5815:
5816: @example
5817: : lowest-level ( -- c )
5818: key dup 27 = if
5819: 1 throw \ ESCAPE key pressed
5820: else
5821: ." lowest-level successful" CR
5822: then
5823: ;
5824:
5825: : lower-level ( -- c )
5826: lowest-level
5827: \ at this level consider a CTRL-U to be a fatal error
5828: dup 21 = if \ CTRL-U
5829: 2 throw
5830: else
5831: ." lower-level successful" CR
5832: then
5833: ;
5834:
5835: : low-level ( -- c )
5836: ['] lower-level catch
5837: ?dup if
5838: \ error occurred - do we recognise it?
5839: dup 1 = if
5840: \ ESCAPE key pressed.. pretend it was an E
5841: [char] E
5842: else throw \ propogate the error upwards
5843: then
5844: then
5845: ." low-level successfull" CR
5846: ;
5847:
5848: : top-level ( -- )
5849: CR ['] low-level catch \ CATCH is used like EXECUTE
5850: ?dup if \ error occurred..
5851: ." Error " . ." occurred - contact your supplier"
5852: else
5853: ." The '" emit ." ' key was pressed" CR
5854: then
5855: ;
5856: @end example
5857:
5858: The ANS Forth document assigns @code{throw} codes thus:
5859:
5860: @itemize @bullet
5861: @item
5862: codes in the range -1 -- -255 are reserved to be assigned by the
5863: Standard. Assignments for codes in the range -1 -- -58 are currently
5864: documented in the Standard. In particular, @code{-1 throw} is equivalent
5865: to @code{abort} and @code{-2 throw} is equivalent to @code{abort"}.
5866: @item
5867: codes in the range -256 -- -4095 are reserved to be assigned by the system.
5868: @item
5869: all other codes may be assigned by programs.
5870: @end itemize
5871:
5872: Gforth provides the word @code{exception} as a mechanism for assigning
5873: system throw codes to applications. This allows multiple applications to
5874: co-exist in memory without any clash of @code{throw} codes. A definition
5875: of @code{exception} in ANS Forth is provided in
5876: @file{compat/exception.fs}.
5877:
5878:
5879: doc-quit
5880: doc-abort
5881: doc-abort"
5882:
5883: doc-catch
5884: doc-throw
5885: doc---exception-exception
5886:
5887:
5888:
5889: @c -------------------------------------------------------------
5890: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
5891: @section Defining Words
5892: @cindex defining words
5893:
5894: Defining words are used to extend Forth by creating new entries in the dictionary.
5895:
5896: @menu
5897: * CREATE::
5898: * Variables:: Variables and user variables
5899: * Constants::
5900: * Values:: Initialised variables
5901: * Colon Definitions::
5902: * Anonymous Definitions:: Definitions without names
5903: * User-defined Defining Words::
5904: * Deferred words:: Allow forward references
5905: * Aliases::
5906: * Supplying names::
5907: @end menu
5908:
5909: @node CREATE, Variables, Defining Words, Defining Words
5910: @subsection @code{CREATE}
5911: @cindex simple defining words
5912: @cindex defining words, simple
5913:
5914: Defining words are used to create new entries in the dictionary. The
5915: simplest defining word is @code{CREATE}. @code{CREATE} is used like
5916: this:
5917:
5918: @example
5919: CREATE new-word1
5920: @end example
5921:
5922: @code{CREATE} is a parsing word that generates a dictionary entry for
5923: @code{new-word1}. When @code{new-word1} is executed, all that it does is
5924: leave an address on the stack. The address represents the value of
5925: the data space pointer (@code{HERE}) at the time that @code{new-word1}
5926: was defined. Therefore, @code{CREATE} is a way of associating a name
5927: with the address of a region of memory.
5928:
5929: doc-create
5930:
5931: By extending this example to reserve some memory in data space, we end
5932: up with a @i{variable}. Here are two different ways to do it:
5933:
5934: @example
5935: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
5936: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
5937: @end example
5938:
5939: The variable can be examined and modified using @code{@@} (``fetch'') and
5940: @code{!} (``store'') like this:
5941:
5942: @example
5943: new-word2 @@ . \ get address, fetch from it and display
5944: 1234 new-word2 ! \ new value, get address, store to it
5945: @end example
5946:
5947: @cindex arrays
5948: A similar mechanism can be used to create arrays. For example, an
5949: 80-character text input buffer:
5950:
5951: @example
5952: CREATE text-buf 80 chars allot
5953:
5954: text-buf 0 chars c@@ \ the 1st character (offset 0)
5955: text-buf 3 chars c@@ \ the 4th character (offset 3)
5956: @end example
5957:
5958: You can build arbitrarily complex data structures by allocating
5959: appropriate areas of memory. For further discussions of this, and to
5960: learn about some Gforth tools that make it easier,
5961: @xref{Structures}.
5962:
5963:
5964: @node Variables, Constants, CREATE, Defining Words
5965: @subsection Variables
5966: @cindex variables
5967:
5968: The previous section showed how a sequence of commands could be used to
5969: generate a variable. As a final refinement, the whole code sequence can
5970: be wrapped up in a defining word (pre-empting the subject of the next
5971: section), making it easier to create new variables:
5972:
5973: @example
5974: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
5975: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
5976:
5977: myvariableX foo \ variable foo starts off with an unknown value
5978: myvariable0 joe \ whilst joe is initialised to 0
5979:
5980: 45 3 * foo ! \ set foo to 135
5981: 1234 joe ! \ set joe to 1234
5982: 3 joe +! \ increment joe by 3.. to 1237
5983: @end example
5984:
5985: Not surprisingly, there is no need to define @code{myvariable}, since
5986: Forth already has a definition @code{Variable}. ANS Forth does not
5987: require a @code{Variable} to be initialised when it is created (i.e., it
5988: behaves like @code{myvariableX}). In contrast, Gforth's @code{Variable}
5989: initialises the variable to 0 (i.e., it behaves exactly like
5990: @code{myvariable0}). Forth also provides @code{2Variable} and
5991: @code{fvariable} for double and floating-point variables, respectively
5992: -- both are initialised to 0 in Gforth. If you use a @code{Variable} to
5993: store a boolean, you can use @code{on} and @code{off} to toggle its
5994: state.
5995:
5996: doc-variable
5997: doc-2variable
5998: doc-fvariable
5999:
6000: @cindex user variables
6001: @cindex user space
6002: The defining word @code{User} behaves in the same way as @code{Variable}.
6003: The difference is that it reserves space in @i{user (data) space} rather
6004: than normal data space. In a Forth system that has a multi-tasker, each
6005: task has its own set of user variables.
6006:
6007: doc-user
6008:
6009: @comment TODO is that stuff about user variables strictly correct? Is it
6010: @comment just terminal tasks that have user variables?
6011: @comment should document tasker.fs (with some examples) elsewhere
6012: @comment in this manual, then expand on user space and user variables.
6013:
6014:
6015: @node Constants, Values, Variables, Defining Words
6016: @subsection Constants
6017: @cindex constants
6018:
6019: @code{Constant} allows you to declare a fixed value and refer to it by
6020: name. For example:
6021:
6022: @example
6023: 12 Constant INCHES-PER-FOOT
6024: 3E+08 fconstant SPEED-O-LIGHT
6025: @end example
6026:
6027: A @code{Variable} can be both read and written, so its run-time
6028: behaviour is to supply an address through which its current value can be
6029: manipulated. In contrast, the value of a @code{Constant} cannot be
6030: changed once it has been declared@footnote{Well, often it can be -- but
6031: not in a Standard, portable way. It's safer to use a @code{Value} (read
6032: on).} so it's not necessary to supply the address -- it is more
6033: efficient to return the value of the constant directly. That's exactly
6034: what happens; the run-time effect of a constant is to put its value on
6035: the top of the stack (You can find one
6036: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
6037:
6038: Gforth also provides @code{2Constant} and @code{fconstant} for defining
6039: double and floating-point constants, respectively.
6040:
6041: doc-constant
6042: doc-2constant
6043: doc-fconstant
6044:
6045: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
6046: @c nac-> How could that not be true in an ANS Forth? You can't define a
6047: @c constant, use it and then delete the definition of the constant..
6048: @c I agree that it's rather deep, but IMO it is an important difference
6049: @c relative to other programming languages.. often it's annoying: it
6050: @c certainly changes my programming style relative to C.
6051:
6052: Constants in Forth behave differently from their equivalents in other
6053: programming languages. In other languages, a constant (such as an EQU in
6054: assembler or a #define in C) only exists at compile-time; in the
6055: executable program the constant has been translated into an absolute
6056: number and, unless you are using a symbolic debugger, it's impossible to
6057: know what abstract thing that number represents. In Forth a constant has
6058: an entry in the header space and remains there after the code that uses
6059: it has been defined. In fact, it must remain in the dictionary since it
6060: has run-time duties to perform. For example:
6061:
6062: @example
6063: 12 Constant INCHES-PER-FOOT
6064: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
6065: @end example
6066:
6067: @cindex in-lining of constants
6068: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
6069: associated with the constant @code{INCHES-PER-FOOT}. If you use
6070: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
6071: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
6072: attempt to optimise constants by in-lining them where they are used. You
6073: can force Gforth to in-line a constant like this:
6074:
6075: @example
6076: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
6077: @end example
6078:
6079: If you use @code{see} to decompile @i{this} version of
6080: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
6081: longer present. To understand how this works, read
6082: @ref{Interpret/Compile states}, and @ref{Literals}.
6083:
6084: In-lining constants in this way might improve execution time
6085: fractionally, and can ensure that a constant is now only referenced at
6086: compile-time. However, the definition of the constant still remains in
6087: the dictionary. Some Forth compilers provide a mechanism for controlling
6088: a second dictionary for holding transient words such that this second
6089: dictionary can be deleted later in order to recover memory
6090: space. However, there is no standard way of doing this.
6091:
6092:
6093: @node Values, Colon Definitions, Constants, Defining Words
6094: @subsection Values
6095: @cindex values
6096:
6097: A @code{Value} is like a @code{Variable} but with two important
6098: differences:
6099:
6100: @itemize @bullet
6101: @item
6102: A @code{Value} is initialised when it is declared; like a
6103: @code{Constant} but unlike a @code{Variable}.
6104: @item
6105: A @code{Value} returns its value rather than its address when it is
6106: executed; i.e., it has the same run-time behaviour as @code{Constant}.
6107: @end itemize
6108:
6109: A @code{Value} needs an additional word, @code{TO} to allow its value to
6110: be changed. Here are some examples:
6111:
6112: @example
6113: 12 Value APPLES \ Define APPLES with an initial value of 12
6114: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6115: APPLES \ puts 34 on the top of the stack.
6116: @end example
6117:
6118: doc-value
6119: doc-to
6120:
6121:
6122: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6123: @subsection Colon Definitions
6124: @cindex colon definitions
6125:
6126: @example
6127: : name ( ... -- ... )
6128: word1 word2 word3 ;
6129: @end example
6130:
6131: @noindent
6132: Creates a word called @code{name} that, upon execution, executes
6133: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
6134:
6135: The explanation above is somewhat superficial. For simple examples of
6136: colon definitions see @ref{Your first definition}. For an in-depth
6137: discussion of some of the issues involved, @xref{Interpretation and
6138: Compilation Semantics}.
6139:
6140: doc-:
6141: doc-;
6142:
6143:
6144: @node Anonymous Definitions, User-defined Defining Words, Colon Definitions, Defining Words
6145: @subsection Anonymous Definitions
6146: @cindex colon definitions
6147: @cindex defining words without name
6148:
6149: Sometimes you want to define an @dfn{anonymous word}; a word without a
6150: name. You can do this with:
6151:
6152: doc-:noname
6153:
6154: This leaves the execution token for the word on the stack after the
6155: closing @code{;}. Here's an example in which a deferred word is
6156: initialised with an @code{xt} from an anonymous colon definition:
6157:
6158: @example
6159: Defer deferred
6160: :noname ( ... -- ... )
6161: ... ;
6162: IS deferred
6163: @end example
6164:
6165: @noindent
6166: Gforth provides an alternative way of doing this, using two separate
6167: words:
6168:
6169: doc-noname
6170: @cindex execution token of last defined word
6171: doc-lastxt
6172:
6173: @noindent
6174: The previous example can be rewritten using @code{noname} and
6175: @code{lastxt}:
6176:
6177: @example
6178: Defer deferred
6179: noname : ( ... -- ... )
6180: ... ;
6181: lastxt IS deferred
6182: @end example
6183:
6184: @noindent
6185: @code{noname} works with any defining word, not just @code{:}.
6186:
6187: @code{lastxt} also works when the last word was not defined as
6188: @code{noname}. It also has the useful property that is is valid as soon
6189: as the header for a definition has been built. Thus:
6190:
6191: @example
6192: lastxt . : foo [ lastxt . ] ; ' foo .
6193: @end example
6194:
6195: @noindent
6196: prints 3 numbers; the last two are the same.
6197:
6198:
6199: @node User-defined Defining Words, Deferred words, Anonymous Definitions, Defining Words
6200: @subsection User-defined Defining Words
6201: @cindex user-defined defining words
6202: @cindex defining words, user-defined
6203:
6204: You can create a new defining word by wrapping defining-time code around
6205: an existing defining word and putting the sequence in a colon
6206: definition. For example, suppose that you have a word @code{stats} that
6207: gathers statistics about colon definitions given the @i{xt} of the
6208: definition, and you want every colon definition in your application to
6209: make a call to @code{stats}. You can define and use a new version of
6210: @code{:} like this:
6211:
6212: @example
6213: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6214: ... ; \ other code
6215:
6216: : my: : lastxt postpone literal ['] stats compile, ;
6217:
6218: my: foo + - ;
6219: @end example
6220:
6221: When @code{foo} is defined using @code{my:} these steps occur:
6222:
6223: @itemize @bullet
6224: @item
6225: @code{my:} is executed.
6226: @item
6227: The @code{:} within the definition (the one between @code{my:} and
6228: @code{lastxt}) is executed, and does just what it always does; it parses
6229: the input stream for a name, builds a dictionary header for the name
6230: @code{foo} and switches @code{state} from interpret to compile.
6231: @item
6232: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
6233: being defined -- @code{foo} -- onto the stack.
6234: @item
6235: The code that was produced by @code{postpone literal} is executed; this
6236: causes the value on the stack to be compiled as a literal in the code
6237: area of @code{foo}.
6238: @item
6239: The code @code{['] stats} compiles a literal into the definition of
6240: @code{my:}. When @code{compile,} is executed, that literal -- the
6241: execution token for @code{stats} -- is layed down in the code area of
6242: @code{foo} , following the literal@footnote{Strictly speaking, the
6243: mechanism that @code{compile,} uses to convert an @i{xt} into something
6244: in the code area is implementation-dependent. A threaded implementation
6245: might spit out the execution token directly whilst another
6246: implementation might spit out a native code sequence.}.
6247: @item
6248: At this point, the execution of @code{my:} is complete, and control
6249: returns to the text interpreter. The text interpreter is in compile
6250: state, so subsequent text @code{+ -} is compiled into the definition of
6251: @code{foo} and the @code{;} terminates the definition as always.
6252: @end itemize
6253:
6254: You can use @code{see} to decompile a word that was defined using
6255: @code{my:} and see how it is different from a normal @code{:}
6256: definition. For example:
6257:
6258: @example
6259: : bar + - ; \ like foo but using : rather than my:
6260: see bar
6261: : bar
6262: + - ;
6263: see foo
6264: : foo
6265: 107645672 stats + - ;
6266:
6267: \ use ' stats . to show that 107645672 is the xt for stats
6268: @end example
6269:
6270: You can use techniques like this to make new defining words in terms of
6271: @i{any} existing defining word.
6272:
6273:
6274: @cindex defining defining words
6275: @cindex @code{CREATE} ... @code{DOES>}
6276: If you want the words defined with your defining words to behave
6277: differently from words defined with standard defining words, you can
6278: write your defining word like this:
6279:
6280: @example
6281: : def-word ( "name" -- )
6282: CREATE @i{code1}
6283: DOES> ( ... -- ... )
6284: @i{code2} ;
6285:
6286: def-word name
6287: @end example
6288:
6289: @cindex child words
6290: This fragment defines a @dfn{defining word} @code{def-word} and then
6291: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6292: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6293: is not executed at this time. The word @code{name} is sometimes called a
6294: @dfn{child} of @code{def-word}.
6295:
6296: When you execute @code{name}, the address of the body of @code{name} is
6297: put on the data stack and @i{code2} is executed (the address of the body
6298: of @code{name} is the address @code{HERE} returns immediately after the
6299: @code{CREATE}).
6300:
6301: @cindex atavism in child words
6302: You can use @code{def-word} to define a set of child words that behave
6303: differently, though atavistically; they all have a common run-time
6304: behaviour determined by @i{code2}. Typically, the @i{code1} sequence
6305: builds a data area in the body of the child word. The structure of the
6306: data is common to all children of @code{def-word}, but the data values
6307: are specific -- and private -- to each child word. When a child word is
6308: executed, the address of its private data area is passed as a parameter
6309: on TOS to be used and manipulated@footnote{It is legitimate both to read
6310: and write to this data area.} by @i{code2}.
6311:
6312: The two fragments of code that make up the defining words act (are
6313: executed) at two completely separate times:
6314:
6315: @itemize @bullet
6316: @item
6317: At @i{define time}, the defining word executes @i{code1} to generate a
6318: child word
6319: @item
6320: At @i{child execution time}, when a child word is invoked, @i{code2}
6321: is executed, using parameters (data) that are private and specific to
6322: the child word.
6323: @end itemize
6324:
6325: Another way of understanding the behaviour of @code{def-word} and
6326: @code{name} is to say that, if you make the following definitions:
6327: @example
6328: : def-word1 ( "name" -- )
6329: CREATE @i{code1} ;
6330:
6331: : action1 ( ... -- ... )
6332: @i{code2} ;
6333:
6334: def-word1 name1
6335: @end example
6336:
6337: @noindent
6338: Then using @code{name1 action1} is equivalent to using @code{name}.
6339:
6340: The classic example is that you can define @code{CONSTANT} in this way:
6341:
6342: @example
6343: : CONSTANT ( w "name" -- )
6344: CREATE ,
6345: DOES> ( -- w )
6346: @@ ;
6347: @end example
6348:
6349: @comment There is a beautiful description of how this works and what
6350: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6351: @comment commentary on the Counting Fruits problem.
6352:
6353: When you create a constant with @code{5 CONSTANT five}, a set of
6354: define-time actions take place; first a new word @code{five} is created,
6355: then the value 5 is laid down in the body of @code{five} with
6356: @code{,}. When @code{five} is executed, the address of the body is put on
6357: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6358: no code of its own; it simply contains a data field and a pointer to the
6359: code that follows @code{DOES>} in its defining word. That makes words
6360: created in this way very compact.
6361:
6362: The final example in this section is intended to remind you that space
6363: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6364: both read and written by a Standard program@footnote{Exercise: use this
6365: example as a starting point for your own implementation of @code{Value}
6366: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6367: @code{[']}.}:
6368:
6369: @example
6370: : foo ( "name" -- )
6371: CREATE -1 ,
6372: DOES> ( -- )
6373: @@ . ;
6374:
6375: foo first-word
6376: foo second-word
6377:
6378: 123 ' first-word >BODY !
6379: @end example
6380:
6381: If @code{first-word} had been a @code{CREATE}d word, we could simply
6382: have executed it to get the address of its data field. However, since it
6383: was defined to have @code{DOES>} actions, its execution semantics are to
6384: perform those @code{DOES>} actions. To get the address of its data field
6385: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6386: translate the xt into the address of the data field. When you execute
6387: @code{first-word}, it will display @code{123}. When you execute
6388: @code{second-word} it will display @code{-1}.
6389:
6390: @cindex stack effect of @code{DOES>}-parts
6391: @cindex @code{DOES>}-parts, stack effect
6392: In the examples above the stack comment after the @code{DOES>} specifies
6393: the stack effect of the defined words, not the stack effect of the
6394: following code (the following code expects the address of the body on
6395: the top of stack, which is not reflected in the stack comment). This is
6396: the convention that I use and recommend (it clashes a bit with using
6397: locals declarations for stack effect specification, though).
6398:
6399: @menu
6400: * CREATE..DOES> applications::
6401: * CREATE..DOES> details::
6402: * Advanced does> usage example::
6403: @end menu
6404:
6405: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
6406: @subsubsection Applications of @code{CREATE..DOES>}
6407: @cindex @code{CREATE} ... @code{DOES>}, applications
6408:
6409: You may wonder how to use this feature. Here are some usage patterns:
6410:
6411: @cindex factoring similar colon definitions
6412: When you see a sequence of code occurring several times, and you can
6413: identify a meaning, you will factor it out as a colon definition. When
6414: you see similar colon definitions, you can factor them using
6415: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6416: that look very similar:
6417: @example
6418: : ori, ( reg-target reg-source n -- )
6419: 0 asm-reg-reg-imm ;
6420: : andi, ( reg-target reg-source n -- )
6421: 1 asm-reg-reg-imm ;
6422: @end example
6423:
6424: @noindent
6425: This could be factored with:
6426: @example
6427: : reg-reg-imm ( op-code -- )
6428: CREATE ,
6429: DOES> ( reg-target reg-source n -- )
6430: @@ asm-reg-reg-imm ;
6431:
6432: 0 reg-reg-imm ori,
6433: 1 reg-reg-imm andi,
6434: @end example
6435:
6436: @cindex currying
6437: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6438: supply a part of the parameters for a word (known as @dfn{currying} in
6439: the functional language community). E.g., @code{+} needs two
6440: parameters. Creating versions of @code{+} with one parameter fixed can
6441: be done like this:
6442: @example
6443: : curry+ ( n1 -- )
6444: CREATE ,
6445: DOES> ( n2 -- n1+n2 )
6446: @@ + ;
6447:
6448: 3 curry+ 3+
6449: -2 curry+ 2-
6450: @end example
6451:
6452: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
6453: @subsubsection The gory details of @code{CREATE..DOES>}
6454: @cindex @code{CREATE} ... @code{DOES>}, details
6455:
6456: doc-does>
6457:
6458: @cindex @code{DOES>} in a separate definition
6459: This means that you need not use @code{CREATE} and @code{DOES>} in the
6460: same definition; you can put the @code{DOES>}-part in a separate
6461: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
6462: @example
6463: : does1
6464: DOES> ( ... -- ... )
6465: ... ;
6466:
6467: : does2
6468: DOES> ( ... -- ... )
6469: ... ;
6470:
6471: : def-word ( ... -- ... )
6472: create ...
6473: IF
6474: does1
6475: ELSE
6476: does2
6477: ENDIF ;
6478: @end example
6479:
6480: In this example, the selection of whether to use @code{does1} or
6481: @code{does2} is made at compile-time; at the time that the child word is
6482: @code{CREATE}d.
6483:
6484: @cindex @code{DOES>} in interpretation state
6485: In a standard program you can apply a @code{DOES>}-part only if the last
6486: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6487: will override the behaviour of the last word defined in any case. In a
6488: standard program, you can use @code{DOES>} only in a colon
6489: definition. In Gforth, you can also use it in interpretation state, in a
6490: kind of one-shot mode; for example:
6491: @example
6492: CREATE name ( ... -- ... )
6493: @i{initialization}
6494: DOES>
6495: @i{code} ;
6496: @end example
6497:
6498: @noindent
6499: is equivalent to the standard:
6500: @example
6501: :noname
6502: DOES>
6503: @i{code} ;
6504: CREATE name EXECUTE ( ... -- ... )
6505: @i{initialization}
6506: @end example
6507:
6508: doc->body
6509:
6510: @node Advanced does> usage example, , CREATE..DOES> details, User-defined Defining Words
6511: @subsubsection Advanced does> usage example
6512:
6513: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6514: for disassembling instructions, that follow a very repetetive scheme:
6515:
6516: @example
6517: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6518: @var{entry-num} cells @var{table} + !
6519: @end example
6520:
6521: Of course, this inspires the idea to factor out the commonalities to
6522: allow a definition like
6523:
6524: @example
6525: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6526: @end example
6527:
6528: The parameters @var{disasm-operands} and @var{table} are usually
6529: correlated. Moreover, there existed code defining instructions like
6530: this:
6531:
6532: @example
6533: @var{entry-num} @var{inst-format} @var{inst-name}
6534: @end example
6535:
6536: This code comes from the assembler and resides in
6537: @file{arch/mips/insts.fs}.
6538:
6539: So I had to define the @var{inst-format} words that performed the scheme
6540: above when executed. At first I chose to use run-time code-generation:
6541:
6542: @example
6543: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6544: :noname Postpone @var{disasm-operands}
6545: name Postpone sliteral Postpone type Postpone ;
6546: swap cells @var{table} + ! ;
6547: @end example
6548:
6549: Note that this supplies the other two parameters of the scheme above.
6550:
6551: An alternative would have been to write this using
6552: @code{create}/@code{does>}:
6553:
6554: @example
6555: : @var{inst-format} ( entry-num "name" -- )
6556: here name string, ( entry-num c-addr ) \ parse and save "name"
6557: noname create , ( entry-num )
6558: lastxt swap cells @var{table} + !
6559: does> ( addr w -- )
6560: \ disassemble instruction w at addr
6561: @@ >r
6562: @var{disasm-operands}
6563: r> count type ;
6564: @end example
6565:
6566: Somehow the first solution is simpler, mainly because it's simpler to
6567: shift a string from definition-time to use-time with @code{sliteral}
6568: than with @code{string,} and friends.
6569:
6570: I wrote a lot of words following this scheme and soon thought about
6571: factoring out the commonalities among them. Note that this uses a
6572: two-level defining word, i.e., a word that defines ordinary defining
6573: words.
6574:
6575: This time a solution involving @code{postpone} and friends seemed more
6576: difficult (try it as an exercise), so I decided to use a
6577: @code{create}/@code{does>} word; since I was already at it, I also used
6578: @code{create}/@code{does>} for the lower level (try using
6579: @code{postpone} etc. as an exercise), resulting in the following
6580: definition:
6581:
6582: @example
6583: : define-format ( disasm-xt table-xt -- )
6584: \ define an instruction format that uses disasm-xt for
6585: \ disassembling and enters the defined instructions into table
6586: \ table-xt
6587: create 2,
6588: does> ( u "inst" -- )
6589: \ defines an anonymous word for disassembling instruction inst,
6590: \ and enters it as u-th entry into table-xt
6591: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6592: noname create 2, \ define anonymous word
6593: execute lastxt swap ! \ enter xt of defined word into table-xt
6594: does> ( addr w -- )
6595: \ disassemble instruction w at addr
6596: 2@@ >r ( addr w disasm-xt R: c-addr )
6597: execute ( R: c-addr ) \ disassemble operands
6598: r> count type ; \ print name
6599: @end example
6600:
6601: Note that the tables here (in contrast to above) do the @code{cells +}
6602: by themselves (that's why you have to pass an xt). This word is used in
6603: the following way:
6604:
6605: @example
6606: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6607: @end example
6608:
6609: In terms of currying, this kind of two-level defining word provides the
6610: parameters in three stages: first @var{disasm-operands} and @var{table},
6611: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6612: the instruction to be disassembled.
6613:
6614: Of course this did not quite fit all the instruction format names used
6615: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6616: the parameters into the right form.
6617:
6618: If you have trouble following this section, don't worry. First, this is
6619: involved and takes time (and probably some playing around) to
6620: understand; second, this is the first two-level
6621: @code{create}/@code{does>} word I have written in seventeen years of
6622: Forth; and if I did not have @file{insts.fs} to start with, I may well
6623: have elected to use just a one-level defining word (with some repeating
6624: of parameters when using the defining word). So it is not necessary to
6625: understand this, but it may improve your understanding of Forth.
6626:
6627:
6628: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6629: @subsection Deferred words
6630: @cindex deferred words
6631:
6632: The defining word @code{Defer} allows you to define a word by name
6633: without defining its behaviour; the definition of its behaviour is
6634: deferred. Here are two situation where this can be useful:
6635:
6636: @itemize @bullet
6637: @item
6638: Where you want to allow the behaviour of a word to be altered later, and
6639: for all precompiled references to the word to change when its behaviour
6640: is changed.
6641: @item
6642: For mutual recursion; @xref{Calls and returns}.
6643: @end itemize
6644:
6645: In the following example, @code{foo} always invokes the version of
6646: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6647: always invokes the version that prints ``@code{Hello}''. There is no way
6648: of getting @code{foo} to use the later version without re-ordering the
6649: source code and recompiling it.
6650:
6651: @example
6652: : greet ." Good morning" ;
6653: : foo ... greet ... ;
6654: : greet ." Hello" ;
6655: : bar ... greet ... ;
6656: @end example
6657:
6658: This problem can be solved by defining @code{greet} as a @code{Defer}red
6659: word. The behaviour of a @code{Defer}red word can be defined and
6660: redefined at any time by using @code{IS} to associate the xt of a
6661: previously-defined word with it. The previous example becomes:
6662:
6663: @example
6664: Defer greet
6665: : foo ... greet ... ;
6666: : bar ... greet ... ;
6667: : greet1 ." Good morning" ;
6668: : greet2 ." Hello" ;
6669: ' greet2 <IS> greet \ make greet behave like greet2
6670: @end example
6671:
6672: A deferred word can be used to improve the statistics-gathering example
6673: from @ref{User-defined Defining Words}; rather than edit the
6674: application's source code to change every @code{:} to a @code{my:}, do
6675: this:
6676:
6677: @example
6678: : real: : ; \ retain access to the original
6679: defer : \ redefine as a deferred word
6680: ' my: IS : \ use special version of :
6681: \
6682: \ load application here
6683: \
6684: ' real: IS : \ go back to the original
6685: @end example
6686:
6687:
6688: One thing to note is that @code{<IS>} consumes its name when it is
6689: executed. If you want to specify the name at compile time, use
6690: @code{[IS]}:
6691:
6692: @example
6693: : set-greet ( xt -- )
6694: [IS] greet ;
6695:
6696: ' greet1 set-greet
6697: @end example
6698:
6699: A deferred word can only inherit default semantics from the xt (because
6700: that is all that an xt can represent -- for more discussion of this
6701: @pxref{Tokens for Words}). However, the semantics of the deferred word
6702: itself can be modified at the time that it is defined. For example:
6703:
6704: @example
6705: : bar .... ; compile-only
6706: Defer fred immediate
6707: Defer jim
6708:
6709: ' bar <IS> jim \ jim has default semantics
6710: ' bar <IS> fred \ fred is immediate
6711: @end example
6712:
6713: doc-defer
6714: doc-<is>
6715: doc-[is]
6716: doc-is
6717: @comment TODO document these: what's defers [is]
6718: doc-what's
6719: doc-defers
6720:
6721: @c Use @code{words-deferred} to see a list of deferred words.
6722:
6723: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
6724: are provided in @file{compat/defer.fs}.
6725:
6726:
6727: @node Aliases, Supplying names, Deferred words, Defining Words
6728: @subsection Aliases
6729: @cindex aliases
6730:
6731: The defining word @code{Alias} allows you to define a word by name that
6732: has the same behaviour as some other word. Here are two situation where
6733: this can be useful:
6734:
6735: @itemize @bullet
6736: @item
6737: When you want access to a word's definition from a different word list
6738: (for an example of this, see the definition of the @code{Root} word list
6739: in the Gforth source).
6740: @item
6741: When you want to create a synonym; a definition that can be known by
6742: either of two names (for example, @code{THEN} and @code{ENDIF} are
6743: aliases).
6744: @end itemize
6745:
6746: The word whose behaviour the alias is to inherit is represented by an
6747: xt. Therefore, the alias only inherits default semantics from its
6748: ancestor. The semantics of the alias itself can be modified at the time
6749: that it is defined. For example:
6750:
6751: @example
6752: : foo ... ; immediate
6753:
6754: ' foo Alias bar \ bar is not an immediate word
6755: ' foo Alias fooby immediate \ fooby is an immediate word
6756: @end example
6757:
6758: Words that are aliases have the same xt, different headers in the
6759: dictionary, and consequently different name tokens (@pxref{Tokens for
6760: Words}) and possibly different immediate flags. An alias can only have
6761: default or immediate compilation semantics; you can define aliases for
6762: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
6763:
6764: doc-alias
6765:
6766:
6767: @node Supplying names, , Aliases, Defining Words
6768: @subsection Supplying the name of a defined word
6769: @cindex names for defined words
6770: @cindex defining words, name given in a string
6771:
6772: By default, a defining word takes the name for the defined word from the
6773: input stream. Sometimes you want to supply the name from a string. You
6774: can do this with:
6775:
6776: doc-nextname
6777:
6778: For example:
6779:
6780: @example
6781: s" foo" nextname create
6782: @end example
6783:
6784: @noindent
6785: is equivalent to:
6786:
6787: @example
6788: create foo
6789: @end example
6790:
6791: @noindent
6792: @code{nextname} works with any defining word, not just @code{:}.
6793:
6794:
6795: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6796: @section Interpretation and Compilation Semantics
6797: @cindex semantics, interpretation and compilation
6798:
6799: @cindex interpretation semantics
6800: The @dfn{interpretation semantics} of a word are what the text
6801: interpreter does when it encounters the word in interpret state. It also
6802: appears in some other contexts, e.g., the execution token returned by
6803: @code{' @i{word}} identifies the interpretation semantics of
6804: @i{word} (in other words, @code{' @i{word} execute} is equivalent to
6805: interpret-state text interpretation of @code{@i{word}}).
6806:
6807: @cindex compilation semantics
6808: The @dfn{compilation semantics} of a word are what the text interpreter
6809: does when it encounters the word in compile state. It also appears in
6810: other contexts, e.g, @code{POSTPONE @i{word}} compiles@footnote{In
6811: standard terminology, ``appends to the current definition''.} the
6812: compilation semantics of @i{word}.
6813:
6814: @cindex execution semantics
6815: The standard also talks about @dfn{execution semantics}. They are used
6816: only for defining the interpretation and compilation semantics of many
6817: words. By default, the interpretation semantics of a word are to
6818: @code{execute} its execution semantics, and the compilation semantics of
6819: a word are to @code{compile,} its execution semantics.@footnote{In
6820: standard terminology: The default interpretation semantics are its
6821: execution semantics; the default compilation semantics are to append its
6822: execution semantics to the execution semantics of the current
6823: definition.}
6824:
6825: @comment TODO expand, make it co-operate with new sections on text interpreter.
6826:
6827: @cindex immediate words
6828: @cindex compile-only words
6829: You can change the semantics of the most-recently defined word:
6830:
6831:
6832: doc-immediate
6833: doc-compile-only
6834: doc-restrict
6835:
6836:
6837: Note that ticking (@code{'}) a compile-only word gives an error
6838: (``Interpreting a compile-only word'').
6839:
6840: @menu
6841: * Combined words::
6842: @end menu
6843:
6844: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
6845: @subsection Combined Words
6846: @cindex combined words
6847:
6848: Gforth allows you to define @dfn{combined words} -- words that have an
6849: arbitrary combination of interpretation and compilation semantics.
6850:
6851:
6852: doc-interpret/compile:
6853:
6854:
6855: This feature was introduced for implementing @code{TO} and @code{S"}. I
6856: recommend that you do not define such words, as cute as they may be:
6857: they make it hard to get at both parts of the word in some contexts.
6858: E.g., assume you want to get an execution token for the compilation
6859: part. Instead, define two words, one that embodies the interpretation
6860: part, and one that embodies the compilation part. Once you have done
6861: that, you can define a combined word with @code{interpret/compile:} for
6862: the convenience of your users.
6863:
6864: You might try to use this feature to provide an optimizing
6865: implementation of the default compilation semantics of a word. For
6866: example, by defining:
6867: @example
6868: :noname
6869: foo bar ;
6870: :noname
6871: POSTPONE foo POSTPONE bar ;
6872: interpret/compile: opti-foobar
6873: @end example
6874:
6875: @noindent
6876: as an optimizing version of:
6877:
6878: @example
6879: : foobar
6880: foo bar ;
6881: @end example
6882:
6883: Unfortunately, this does not work correctly with @code{[compile]},
6884: because @code{[compile]} assumes that the compilation semantics of all
6885: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
6886: opti-foobar} would compile compilation semantics, whereas
6887: @code{[compile] foobar} would compile interpretation semantics.
6888:
6889: @cindex state-smart words (are a bad idea)
6890: Some people try to use @dfn{state-smart} words to emulate the feature provided
6891: by @code{interpret/compile:} (words are state-smart if they check
6892: @code{STATE} during execution). E.g., they would try to code
6893: @code{foobar} like this:
6894:
6895: @example
6896: : foobar
6897: STATE @@
6898: IF ( compilation state )
6899: POSTPONE foo POSTPONE bar
6900: ELSE
6901: foo bar
6902: ENDIF ; immediate
6903: @end example
6904:
6905: Although this works if @code{foobar} is only processed by the text
6906: interpreter, it does not work in other contexts (like @code{'} or
6907: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
6908: for a state-smart word, not for the interpretation semantics of the
6909: original @code{foobar}; when you execute this execution token (directly
6910: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
6911: state, the result will not be what you expected (i.e., it will not
6912: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
6913: write them@footnote{For a more detailed discussion of this topic, see
6914: M. Anton Ertl,
6915: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
6916: it is Evil and How to Exorcise it}}, EuroForth '98.}!
6917:
6918: @cindex defining words with arbitrary semantics combinations
6919: It is also possible to write defining words that define words with
6920: arbitrary combinations of interpretation and compilation semantics. In
6921: general, they look like this:
6922:
6923: @example
6924: : def-word
6925: create-interpret/compile
6926: @i{code1}
6927: interpretation>
6928: @i{code2}
6929: <interpretation
6930: compilation>
6931: @i{code3}
6932: <compilation ;
6933: @end example
6934:
6935: For a @i{word} defined with @code{def-word}, the interpretation
6936: semantics are to push the address of the body of @i{word} and perform
6937: @i{code2}, and the compilation semantics are to push the address of
6938: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
6939: can also be defined like this (except that the defined constants don't
6940: behave correctly when @code{[compile]}d):
6941:
6942: @example
6943: : constant ( n "name" -- )
6944: create-interpret/compile
6945: ,
6946: interpretation> ( -- n )
6947: @@
6948: <interpretation
6949: compilation> ( compilation. -- ; run-time. -- n )
6950: @@ postpone literal
6951: <compilation ;
6952: @end example
6953:
6954:
6955: doc-create-interpret/compile
6956: doc-interpretation>
6957: doc-<interpretation
6958: doc-compilation>
6959: doc-<compilation
6960:
6961:
6962: Words defined with @code{interpret/compile:} and
6963: @code{create-interpret/compile} have an extended header structure that
6964: differs from other words; however, unless you try to access them with
6965: plain address arithmetic, you should not notice this. Words for
6966: accessing the header structure usually know how to deal with this; e.g.,
6967: @code{'} @i{word} @code{>body} also gives you the body of a word created
6968: with @code{create-interpret/compile}.
6969:
6970:
6971: doc-postpone
6972:
6973: @comment TODO -- expand glossary text for POSTPONE
6974:
6975:
6976: @c -------------------------------------------------------------
6977: @node Tokens for Words, The Text Interpreter, Interpretation and Compilation Semantics, Words
6978: @section Tokens for Words
6979: @cindex tokens for words
6980:
6981: This section describes the creation and use of tokens that represent
6982: words.
6983:
6984: Named words have information stored in their header space entries to
6985: indicate any non-default semantics (@pxref{Interpretation and
6986: Compilation Semantics}). The semantics can be modified, using
6987: @code{immediate} and/or @code{compile-only}, at the time that the words
6988: are defined. Unnamed words have (by definition) no header space
6989: entry, and therefore must have default semantics.
6990:
6991: Named words have interpretation and compilation semantics. Unnamed words
6992: just have execution semantics.
6993:
6994: @cindex xt
6995: @cindex execution token
6996: The execution semantics of an unnamed word are represented by an
6997: @dfn{execution token} (@i{xt}). As explained in @ref{Supplying names},
6998: the execution token of the last word defined can be produced with
6999: @code{lastxt}.
7000:
7001: The interpretation semantics of a named word are also represented by an
7002: execution token. You can produce the execution token using @code{'} or
7003: @code{[']}. A simple example shows the difference between the two:
7004:
7005: @example
7006: : greet ( -- ) ." Hello" ;
7007: : foo ( -- xt ) ['] greet execute ; \ ['] parses greet at compile-time
7008: : bar ( -- ) ' execute ; \ ' parses at run-time
7009:
7010: \ the next four lines all do the same thing
7011: foo
7012: bar greet
7013: greet
7014: ' greet EXECUTE
7015: @end example
7016:
7017: An execution token occupies one cell.
7018: @cindex code field address
7019: @cindex CFA
7020: In Gforth, the abstract data type @i{execution token} is implemented
7021: as a code field address (CFA).
7022: @comment TODO note that the standard does not say what it represents..
7023: @comment and you cannot necessarily compile it in all Forths (eg native
7024: @comment compilers?).
7025:
7026: For literals, use @code{'} in interpreted code and @code{[']} in
7027: compiled code. Gforth's @code{'} and @code{[']} behave somewhat
7028: unusually by complaining about compile-only words. To get the execution
7029: token for a compile-only word @i{name}, use @code{COMP' @i{name} DROP}
7030: or @code{[COMP'] @i{name} DROP}.
7031:
7032: @cindex compilation token
7033: The compilation semantics of a named word are represented by a
7034: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
7035: @i{xt} is an execution token. The compilation semantics represented by
7036: the compilation token can be performed with @code{execute}, which
7037: consumes the whole compilation token, with an additional stack effect
7038: determined by the represented compilation semantics.
7039:
7040: At present, the @i{w} part of a compilation token is an execution token,
7041: and the @i{xt} part represents either @code{execute} or
7042: @code{compile,}@footnote{Depending upon the compilation semantics of the
7043: word. If the word has default compilation semantics, the @i{xt} will
7044: represent @code{compile,}. Otherwise (e.g., for immediate words), the
7045: @i{xt} will represent @code{execute}.}. However, don't rely on that
7046: knowledge, unless necessary; future versions of Gforth may introduce
7047: unusual compilation tokens (e.g., a compilation token that represents
7048: the compilation semantics of a literal).
7049:
7050: You can compile the compilation semantics with @code{postpone,}. I.e.,
7051: @code{COMP' @i{word} postpone,} is equivalent to @code{postpone
7052: @i{word}}.
7053:
7054: @cindex name token
7055: @cindex name field address
7056: @cindex NFA
7057: Named words are also represented by the @dfn{name token}, (@i{nt}). In
7058: Gforth, the abstract data type @emph{name token} is implemented as a
7059: name field address (NFA).
7060:
7061:
7062: doc-execute
7063: doc-perform
7064: doc-compile,
7065: doc-[']
7066: doc-'
7067: doc-[comp']
7068: doc-comp'
7069: doc-postpone,
7070:
7071: doc-find-name
7072: doc-name>int
7073: doc-name?int
7074: doc-name>comp
7075: doc-name>string
7076:
7077:
7078: @c ----------------------------------------------------------
7079: @node The Text Interpreter, Word Lists, Tokens for Words, Words
7080: @section The Text Interpreter
7081: @cindex interpreter - outer
7082: @cindex text interpreter
7083: @cindex outer interpreter
7084:
7085: @c Should we really describe all these ugly details? IMO the text
7086: @c interpreter should be much cleaner, but that may not be possible within
7087: @c ANS Forth. - anton
7088: @c nac-> I wanted to explain how it works to show how you can exploit
7089: @c it in your own programs. When I was writing a cross-compiler, figuring out
7090: @c some of these gory details was very helpful to me. None of the textbooks
7091: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
7092: @c seems to positively avoid going into too much detail for some of
7093: @c the internals.
7094:
7095: The text interpreter@footnote{This is an expanded version of the
7096: material in @ref{Introducing the Text Interpreter}.} is an endless loop
7097: that processes input from the current input device. It is also called
7098: the outer interpreter, in contrast to the inner interpreter
7099: (@pxref{Engine}) which executes the compiled Forth code on interpretive
7100: implementations.
7101:
7102: @cindex interpret state
7103: @cindex compile state
7104: The text interpreter operates in one of two states: @dfn{interpret
7105: state} and @dfn{compile state}. The current state is defined by the
7106: aptly-named variable, @code{state}.
7107:
7108: This section starts by describing how the text interpreter behaves when
7109: it is in interpret state, processing input from the user input device --
7110: the keyboard. This is the mode that a Forth system is in after it starts
7111: up.
7112:
7113: @cindex input buffer
7114: @cindex terminal input buffer
7115: The text interpreter works from an area of memory called the @dfn{input
7116: buffer}@footnote{When the text interpreter is processing input from the
7117: keyboard, this area of memory is called the @dfn{terminal input buffer}
7118: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7119: @code{#TIB}.}, which stores your keyboard input when you press the
7120: @key{RET} key. Starting at the beginning of the input buffer, it skips
7121: leading spaces (called @dfn{delimiters}) then parses a string (a
7122: sequence of non-space characters) until it reaches either a space
7123: character or the end of the buffer. Having parsed a string, it makes two
7124: attempts to process it:
7125:
7126: @cindex dictionary
7127: @itemize @bullet
7128: @item
7129: It looks for the string in a @dfn{dictionary} of definitions. If the
7130: string is found, the string names a @dfn{definition} (also known as a
7131: @dfn{word}) and the dictionary search returns information that allows
7132: the text interpreter to perform the word's @dfn{interpretation
7133: semantics}. In most cases, this simply means that the word will be
7134: executed.
7135: @item
7136: If the string is not found in the dictionary, the text interpreter
7137: attempts to treat it as a number, using the rules described in
7138: @ref{Number Conversion}. If the string represents a legal number in the
7139: current radix, the number is pushed onto a parameter stack (the data
7140: stack for integers, the floating-point stack for floating-point
7141: numbers).
7142: @end itemize
7143:
7144: If both attempts fail, or if the word is found in the dictionary but has
7145: no interpretation semantics@footnote{This happens if the word was
7146: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7147: remainder of the input buffer, issues an error message and waits for
7148: more input. If one of the attempts succeeds, the text interpreter
7149: repeats the parsing process until the whole of the input buffer has been
7150: processed, at which point it prints the status message ``@code{ ok}''
7151: and waits for more input.
7152:
7153: @cindex parse area
7154: The text interpreter keeps track of its position in the input buffer by
7155: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7156: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7157: of the input buffer. The region from offset @code{>IN @@} to the end of
7158: the input buffer is called the @dfn{parse area}@footnote{In other words,
7159: the text interpreter processes the contents of the input buffer by
7160: parsing strings from the parse area until the parse area is empty.}.
7161: This example shows how @code{>IN} changes as the text interpreter parses
7162: the input buffer:
7163:
7164: @example
7165: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7166: CR ." ->" TYPE ." <-" ; IMMEDIATE
7167:
7168: 1 2 3 remaining + remaining .
7169:
7170: : foo 1 2 3 remaining SWAP remaining ;
7171: @end example
7172:
7173: @noindent
7174: The result is:
7175:
7176: @example
7177: ->+ remaining .<-
7178: ->.<-5 ok
7179:
7180: ->SWAP remaining ;-<
7181: ->;<- ok
7182: @end example
7183:
7184: @cindex parsing words
7185: The value of @code{>IN} can also be modified by a word in the input
7186: buffer that is executed by the text interpreter. This means that a word
7187: can ``trick'' the text interpreter into either skipping a section of the
7188: input buffer@footnote{This is how parsing words work.} or into parsing a
7189: section twice. For example:
7190:
7191: @example
7192: : lat ." <<lat>>" ;
7193: : flat ." <<flat>>" >IN DUP @@ 3 - SWAP ! ;
7194: @end example
7195:
7196: @noindent
7197: When @code{flat} is executed, this output is produced@footnote{Exercise
7198: for the reader: what would happen if the @code{3} were replaced with
7199: @code{4}?}:
7200:
7201: @example
7202: <<flat>><<lat>>
7203: @end example
7204:
7205: @noindent
7206: Two important notes about the behaviour of the text interpreter:
7207:
7208: @itemize @bullet
7209: @item
7210: It processes each input string to completion before parsing additional
7211: characters from the input buffer.
7212: @item
7213: It treats the input buffer as a read-only region (and so must your code).
7214: @end itemize
7215:
7216: @noindent
7217: When the text interpreter is in compile state, its behaviour changes in
7218: these ways:
7219:
7220: @itemize @bullet
7221: @item
7222: If a parsed string is found in the dictionary, the text interpreter will
7223: perform the word's @dfn{compilation semantics}. In most cases, this
7224: simply means that the execution semantics of the word will be appended
7225: to the current definition.
7226: @item
7227: When a number is encountered, it is compiled into the current definition
7228: (as a literal) rather than being pushed onto a parameter stack.
7229: @item
7230: If an error occurs, @code{state} is modified to put the text interpreter
7231: back into interpret state.
7232: @item
7233: Each time a line is entered from the keyboard, Gforth prints
7234: ``@code{ compiled}'' rather than `` @code{ok}''.
7235: @end itemize
7236:
7237: @cindex text interpreter - input sources
7238: When the text interpreter is using an input device other than the
7239: keyboard, its behaviour changes in these ways:
7240:
7241: @itemize @bullet
7242: @item
7243: When the parse area is empty, the text interpreter attempts to refill
7244: the input buffer from the input source. When the input source is
7245: exhausted, the input source is set back to the user input device.
7246: @item
7247: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7248: time the parse area is emptied.
7249: @item
7250: If an error occurs, the input source is set back to the user input
7251: device.
7252: @end itemize
7253:
7254: You can read about this in more detail in @ref{Input Sources}.
7255:
7256: doc->in
7257: doc-source
7258:
7259: doc-tib
7260: doc-#tib
7261:
7262:
7263: @menu
7264: * Input Sources::
7265: * Number Conversion::
7266: * Interpret/Compile states::
7267: * Literals::
7268: * Interpreter Directives::
7269: @end menu
7270:
7271: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7272: @subsection Input Sources
7273: @cindex input sources
7274: @cindex text interpreter - input sources
7275:
7276: By default, the text interpreter processes input from the user input
7277: device (the keyboard) when Forth starts up. The text interpreter can
7278: process input from any of these sources:
7279:
7280: @itemize @bullet
7281: @item
7282: The user input device -- the keyboard.
7283: @item
7284: A file, using the words described in @ref{Forth source files}.
7285: @item
7286: A block, using the words described in @ref{Blocks}.
7287: @item
7288: A text string, using @code{evaluate}.
7289: @end itemize
7290:
7291: A program can identify the current input device from the values of
7292: @code{source-id} and @code{blk}.
7293:
7294:
7295: doc-source-id
7296: doc-blk
7297:
7298: doc-save-input
7299: doc-restore-input
7300:
7301: doc-evaluate
7302:
7303:
7304:
7305: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
7306: @subsection Number Conversion
7307: @cindex number conversion
7308: @cindex double-cell numbers, input format
7309: @cindex input format for double-cell numbers
7310: @cindex single-cell numbers, input format
7311: @cindex input format for single-cell numbers
7312: @cindex floating-point numbers, input format
7313: @cindex input format for floating-point numbers
7314:
7315: This section describes the rules that the text interpreter uses when it
7316: tries to convert a string into a number.
7317:
7318: Let <digit> represent any character that is a legal digit in the current
7319: number base@footnote{For example, 0-9 when the number base is decimal or
7320: 0-9, A-F when the number base is hexadecimal.}.
7321:
7322: Let <decimal digit> represent any character in the range 0-9.
7323:
7324: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7325: in the braces (@i{a} or @i{b} or neither).
7326:
7327: Let * represent any number of instances of the previous character
7328: (including none).
7329:
7330: Let any other character represent itself.
7331:
7332: @noindent
7333: Now, the conversion rules are:
7334:
7335: @itemize @bullet
7336: @item
7337: A string of the form <digit><digit>* is treated as a single-precision
7338: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
7339: @item
7340: A string of the form -<digit><digit>* is treated as a single-precision
7341: (cell-sized) negative integer, and is represented using 2's-complement
7342: arithmetic. Examples are -45 -5681 -0
7343: @item
7344: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
7345: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7346: (all three of these represent the same number).
7347: @item
7348: A string of the form -<digit><digit>*.<digit>* is treated as a
7349: double-precision (double-cell-sized) negative integer, and is
7350: represented using 2's-complement arithmetic. Examples are -3465. -3.465
7351: -34.65 (all three of these represent the same number).
7352: @item
7353: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7354: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
7355: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
7356: number) +12.E-4
7357: @end itemize
7358:
7359: By default, the number base used for integer number conversion is given
7360: by the contents of the variable @code{base}. Note that a lot of
7361: confusion can result from unexpected values of @code{base}. If you
7362: change @code{base} anywhere, make sure to save the old value and restore
7363: it afterwards. In general I recommend keeping @code{base} decimal, and
7364: using the prefixes described below for the popular non-decimal bases.
7365:
7366: doc-dpl
7367: doc-base
7368: doc-hex
7369: doc-decimal
7370:
7371:
7372: @cindex '-prefix for character strings
7373: @cindex &-prefix for decimal numbers
7374: @cindex %-prefix for binary numbers
7375: @cindex $-prefix for hexadecimal numbers
7376: Gforth allows you to override the value of @code{base} by using a
7377: prefix@footnote{Some Forth implementations provide a similar scheme by
7378: implementing @code{$} etc. as parsing words that process the subsequent
7379: number in the input stream and push it onto the stack. For example, see
7380: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7381: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7382: is required between the prefix and the number.} before the first digit
7383: of an (integer) number. Four prefixes are supported:
7384:
7385: @itemize @bullet
7386: @item
7387: @code{&} -- decimal
7388: @item
7389: @code{%} -- binary
7390: @item
7391: @code{$} -- hexadecimal
7392: @item
7393: @code{'} -- base @code{max-char+1}
7394: @end itemize
7395:
7396: Here are some examples, with the equivalent decimal number shown after
7397: in braces:
7398:
7399: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7400: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
7401: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
7402: &905 (905), $abc (2478), $ABC (2478).
7403:
7404: @cindex number conversion - traps for the unwary
7405: @noindent
7406: Number conversion has a number of traps for the unwary:
7407:
7408: @itemize @bullet
7409: @item
7410: You cannot determine the current number base using the code sequence
7411: @code{base @@ .} -- the number base is always 10 in the current number
7412: base. Instead, use something like @code{base @@ dec.}
7413: @item
7414: If the number base is set to a value greater than 14 (for example,
7415: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7416: it to be intepreted as either a single-precision integer or a
7417: floating-point number (Gforth treats it as an integer). The ambiguity
7418: can be resolved by explicitly stating the sign of the mantissa and/or
7419: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7420: ambiguity arises; either representation will be treated as a
7421: floating-point number.
7422: @item
7423: There is a word @code{bin} but it does @i{not} set the number base!
7424: It is used to specify file types.
7425: @item
7426: ANS Forth requires the @code{.} of a double-precision number to
7427: be the final character in the string. Allowing the @code{.} to be
7428: anywhere after the first digit is a Gforth extension.
7429: @item
7430: The number conversion process does not check for overflow.
7431: @item
7432: In Gforth, number conversion to floating-point numbers always use base
7433: 10, irrespective of the value of @code{base}. In ANS Forth,
7434: conversion to floating-point numbers whilst the value of
7435: @code{base} is not 10 is an ambiguous condition.
7436: @end itemize
7437:
7438: You can read numbers into your programs with the words described in
7439: @ref{Input}.
7440:
7441: @node Interpret/Compile states, Literals, Number Conversion, The Text Interpreter
7442: @subsection Interpret/Compile states
7443: @cindex Interpret/Compile states
7444:
7445: A standard program is not permitted to change @code{state}
7446: explicitly. However, it can change @code{state} implicitly, using the
7447: words @code{[} and @code{]}. When @code{[} is executed it switches
7448: @code{state} to interpret state, and therefore the text interpreter
7449: starts interpreting. When @code{]} is executed it switches @code{state}
7450: to compile state and therefore the text interpreter starts
7451: compiling. The most common usage for these words is for switching into
7452: interpret state and back from within a colon definition; this technique
7453: can be used to compile a literal (for an example, @pxref{Literals}) or
7454: for conditional compilation (for an example, @pxref{Interpreter
7455: Directives}).
7456:
7457:
7458: @c This is a bad example: It's non-standard, and it's not necessary.
7459: @c However, I can't think of a good example for switching into compile
7460: @c state when there is no current word (@code{state}-smart words are not a
7461: @c good reason). So maybe we should use an example for switching into
7462: @c interpret @code{state} in a colon def. - anton
7463: @c nac-> I agree. I started out by putting in the example, then realised
7464: @c that it was non-ANS, so wrote more words around it. I hope this
7465: @c re-written version is acceptable to you. I do want to keep the example
7466: @c as it is helpful for showing what is and what is not portable, particularly
7467: @c where it outlaws a style in common use.
7468:
7469:
7470: @code{[} and @code{]} also give you the ability to switch into compile
7471: state and back, but we cannot think of any useful Standard application
7472: for this ability. Pre-ANS Forth textbooks have examples like this:
7473:
7474: @example
7475: : AA ." this is A" ;
7476: : BB ." this is B" ;
7477: : CC ." this is C" ;
7478:
7479: create table ] aa bb cc [
7480:
7481: : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7482: cells table + @ execute ;
7483: @end example
7484:
7485: This example builds a jump table; @code{0 go} will display ``@code{this
7486: is A}''. Using @code{[} and @code{]} in this example is equivalent to
7487: defining @code{table} like this:
7488:
7489: @example
7490: create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
7491: @end example
7492:
7493: The problem with this code is that the definition of @code{table} is not
7494: portable -- it @i{compile}s execution tokens into code space. Whilst it
7495: @i{may} work on systems where code space and data space co-incide, the
7496: Standard only allows data space to be assigned for a @code{CREATE}d
7497: word. In addition, the Standard only allows @code{@@} to access data
7498: space, whilst this example is using it to access code space. The only
7499: portable, Standard way to build this table is to build it in data space,
7500: like this:
7501:
7502: @example
7503: create table ' aa , ' bb , ' cc ,
7504: @end example
7505:
7506: doc-state
7507: doc-[
7508: doc-]
7509:
7510:
7511: @node Literals, Interpreter Directives, Interpret/Compile states, The Text Interpreter
7512: @subsection Literals
7513: @cindex Literals
7514:
7515: Often, you want to use a number within a colon definition. When you do
7516: this, the text interpreter automatically compiles the number as a
7517: @i{literal}. A literal is a number whose run-time effect is to be pushed
7518: onto the stack. If you had to do some maths to generate the number, you
7519: might write it like this:
7520:
7521: @example
7522: : HOUR-TO-SEC ( n1 -- n2 )
7523: 60 * \ to minutes
7524: 60 * ; \ to seconds
7525: @end example
7526:
7527: It is very clear what this definition is doing, but it's inefficient
7528: since it is performing 2 multiples at run-time. An alternative would be
7529: to write:
7530:
7531: @example
7532: : HOUR-TO-SEC ( n1 -- n2 )
7533: 3600 * ; \ to seconds
7534: @end example
7535:
7536: Which does the same thing, and has the advantage of using a single
7537: multiply. Ideally, we'd like the efficiency of the second with the
7538: readability of the first.
7539:
7540: @code{Literal} allows us to achieve that. It takes a number from the
7541: stack and lays it down in the current definition just as though the
7542: number had been typed directly into the definition. Our first attempt
7543: might look like this:
7544:
7545: @example
7546: 60 \ mins per hour
7547: 60 * \ seconds per minute
7548: : HOUR-TO-SEC ( n1 -- n2 )
7549: Literal * ; \ to seconds
7550: @end example
7551:
7552: But this produces the error message @code{unstructured}. What happened?
7553: The stack notation for @code{:} is (@i{ -- colon-sys}) and the size of
7554: @i{colon-sys} is implementation-defined. In other words, once we start a
7555: colon definition we can't portably access anything that was on the stack
7556: before the definition began@footnote{@cite{Two Problems in ANS Forth},
7557: by Thomas Worthington; Forth Dimensions 20(2) pages 32--34 describes
7558: some situations where you might want to access stack items above
7559: colon-sys, and provides a solution to the problem.}. The correct way of
7560: solving this problem in this instance is to use @code{[ ]} like this:
7561:
7562: @example
7563: : HOUR-TO-SEC ( n1 -- n2 )
7564: [ 60 \ minutes per hour
7565: 60 * ] \ seconds per minute
7566: LITERAL * ; \ to seconds
7567: @end example
7568:
7569:
7570: doc-literal
7571: doc-]L
7572: doc-2literal
7573: doc-fliteral
7574:
7575:
7576: @node Interpreter Directives, , Literals, The Text Interpreter
7577: @subsection Interpreter Directives
7578: @cindex interpreter directives
7579:
7580: These words are usually used in interpret state; typically to control
7581: which parts of a source file are processed by the text
7582: interpreter. There are only a few ANS Forth Standard words, but Gforth
7583: supplements these with a rich set of immediate control structure words
7584: to compensate for the fact that the non-immediate versions can only be
7585: used in compile state (@pxref{Control Structures}). Typical usages:
7586:
7587: @example
7588: FALSE Constant ASSEMBLER
7589: .
7590: .
7591: ASSEMBLER [IF]
7592: : ASSEMBLER-FEATURE
7593: ...
7594: ;
7595: [ENDIF]
7596: .
7597: .
7598: : SEE
7599: ... \ general-purpose SEE code
7600: [ ASSEMBLER [IF] ]
7601: ... \ assembler-specific SEE code
7602: [ [ENDIF] ]
7603: ;
7604: @end example
7605:
7606:
7607: doc-[IF]
7608: doc-[ELSE]
7609: doc-[THEN]
7610: doc-[ENDIF]
7611:
7612: doc-[IFDEF]
7613: doc-[IFUNDEF]
7614:
7615: doc-[?DO]
7616: doc-[DO]
7617: doc-[FOR]
7618: doc-[LOOP]
7619: doc-[+LOOP]
7620: doc-[NEXT]
7621:
7622: doc-[BEGIN]
7623: doc-[UNTIL]
7624: doc-[AGAIN]
7625: doc-[WHILE]
7626: doc-[REPEAT]
7627:
7628:
7629: @c -------------------------------------------------------------
7630: @node Word Lists, Environmental Queries, The Text Interpreter, Words
7631: @section Word Lists
7632: @cindex word lists
7633: @cindex header space
7634:
7635: A wordlist is a list of named words; you can add new words and look up
7636: words by name (and you can remove words in a restricted way with
7637: markers). Every named (and @code{reveal}ed) word is in one wordlist.
7638:
7639: @cindex search order stack
7640: The text interpreter searches the wordlists present in the search order
7641: (a stack of wordlists), from the top to the bottom. Within each
7642: wordlist, the search starts conceptually at the newest word; i.e., if
7643: two words in a wordlist have the same name, the newer word is found.
7644:
7645: @cindex compilation word list
7646: New words are added to the @dfn{compilation wordlist} (aka current
7647: wordlist).
7648:
7649: @cindex wid
7650: A word list is identified by a cell-sized word list identifier (@i{wid})
7651: in much the same way as a file is identified by a file handle. The
7652: numerical value of the wid has no (portable) meaning, and might change
7653: from session to session.
7654:
7655: The ANS Forth ``Search order'' word set is intended to provide a set of
7656: low-level tools that allow various different schemes to be
7657: implemented. Gforth provides @code{vocabulary}, a traditional Forth
7658: word. @file{compat/vocabulary.fs} provides an implementation in ANS
7659: Forth.
7660:
7661: @comment TODO: locals section refers to here, saying that every word list (aka
7662: @comment vocabulary) has its own methods for searching etc. Need to document that.
7663:
7664: @comment TODO: document markers, reveal, tables, mappedwordlist
7665:
7666: @comment the gforthman- prefix is used to pick out the true definition of a
7667: @comment word from the source files, rather than some alias.
7668:
7669: doc-forth-wordlist
7670: doc-definitions
7671: doc-get-current
7672: doc-set-current
7673: doc-get-order
7674: doc---gforthman-set-order
7675: doc-wordlist
7676: doc-table
7677: doc-push-order
7678: doc-previous
7679: doc-also
7680: doc---gforthman-forth
7681: doc-only
7682: doc---gforthman-order
7683:
7684: doc-find
7685: doc-search-wordlist
7686:
7687: doc-words
7688: doc-vlist
7689: @c doc-words-deferred
7690:
7691: doc-mappedwordlist
7692: doc-root
7693: doc-vocabulary
7694: doc-seal
7695: doc-vocs
7696: doc-current
7697: doc-context
7698:
7699:
7700: @menu
7701: * Why use word lists?::
7702: * Word list examples::
7703: @end menu
7704:
7705: @node Why use word lists?, Word list examples, Word Lists, Word Lists
7706: @subsection Why use word lists?
7707: @cindex word lists - why use them?
7708:
7709: Here are some reasons for using multiple word lists:
7710:
7711: @itemize @bullet
7712: @item
7713: To improve compilation speed by reducing the number of header space
7714: entries that must be searched. This is achieved by creating a new
7715: word list that contains all of the definitions that are used in the
7716: definition of a Forth system but which would not usually be used by
7717: programs running on that system. That word list would be on the search
7718: list when the Forth system was compiled but would be removed from the
7719: search list for normal operation. This can be a useful technique for
7720: low-performance systems (for example, 8-bit processors in embedded
7721: systems) but is unlikely to be necessary in high-performance desktop
7722: systems.
7723: @item
7724: To prevent a set of words from being used outside the context in which
7725: they are valid. Two classic examples of this are an integrated editor
7726: (all of the edit commands are defined in a separate word list; the
7727: search order is set to the editor word list when the editor is invoked;
7728: the old search order is restored when the editor is terminated) and an
7729: integrated assembler (the op-codes for the machine are defined in a
7730: separate word list which is used when a @code{CODE} word is defined).
7731: @item
7732: To prevent a name-space clash between multiple definitions with the same
7733: name. For example, when building a cross-compiler you might have a word
7734: @code{IF} that generates conditional code for your target system. By
7735: placing this definition in a different word list you can control whether
7736: the host system's @code{IF} or the target system's @code{IF} get used in
7737: any particular context by controlling the order of the word lists on the
7738: search order stack.
7739: @end itemize
7740:
7741: @node Word list examples, , Why use word lists?, Word Lists
7742: @subsection Word list examples
7743: @cindex word lists - examples
7744:
7745: Here is an example of creating and using a new wordlist using ANS
7746: Forth Standard words:
7747:
7748: @example
7749: wordlist constant my-new-words-wordlist
7750: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
7751:
7752: \ add it to the search order
7753: also my-new-words
7754:
7755: \ alternatively, add it to the search order and make it
7756: \ the compilation word list
7757: also my-new-words definitions
7758: \ type "order" to see the problem
7759: @end example
7760:
7761: The problem with this example is that @code{order} has no way to
7762: associate the name @code{my-new-words} with the wid of the word list (in
7763: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
7764: that has no associated name). There is no Standard way of associating a
7765: name with a wid.
7766:
7767: In Gforth, this example can be re-coded using @code{vocabulary}, which
7768: associates a name with a wid:
7769:
7770: @example
7771: vocabulary my-new-words
7772:
7773: \ add it to the search order
7774: also my-new-words
7775:
7776: \ alternatively, add it to the search order and make it
7777: \ the compilation word list
7778: my-new-words definitions
7779: \ type "order" to see that the problem is solved
7780: @end example
7781:
7782: @c -------------------------------------------------------------
7783: @node Environmental Queries, Files, Word Lists, Words
7784: @section Environmental Queries
7785: @cindex environmental queries
7786:
7787: ANS Forth introduced the idea of ``environmental queries'' as a way
7788: for a program running on a system to determine certain characteristics of the system.
7789: The Standard specifies a number of strings that might be recognised by a system.
7790:
7791: The Standard requires that the header space used for environmental queries
7792: be distinct from the header space used for definitions.
7793:
7794: Typically, environmental queries are supported by creating a set of
7795: definitions in a word list that is @i{only} used during environmental
7796: queries; that is what Gforth does. There is no Standard way of adding
7797: definitions to the set of recognised environmental queries, but any
7798: implementation that supports the loading of optional word sets must have
7799: some mechanism for doing this (after loading the word set, the
7800: associated environmental query string must return @code{true}). In
7801: Gforth, the word list used to honour environmental queries can be
7802: manipulated just like any other word list.
7803:
7804:
7805: doc-environment?
7806: doc-environment-wordlist
7807:
7808: doc-gforth
7809: doc-os-class
7810:
7811:
7812: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
7813: returning two items on the stack, querying it using @code{environment?}
7814: will return an additional item; the @code{true} flag that shows that the
7815: string was recognised.
7816:
7817: @comment TODO Document the standard strings or note where they are documented herein
7818:
7819: Here are some examples of using environmental queries:
7820:
7821: @example
7822: s" address-unit-bits" environment? 0=
7823: [IF]
7824: cr .( environmental attribute address-units-bits unknown... ) cr
7825: [THEN]
7826:
7827: s" block" environment? [IF] DROP include block.fs [THEN]
7828:
7829: s" gforth" environment? [IF] 2DROP include compat/vocabulary.fs [THEN]
7830:
7831: s" gforth" environment? [IF] .( Gforth version ) TYPE
7832: [ELSE] .( Not Gforth..) [THEN]
7833: @end example
7834:
7835:
7836: Here is an example of adding a definition to the environment word list:
7837:
7838: @example
7839: get-current environment-wordlist set-current
7840: true constant block
7841: true constant block-ext
7842: set-current
7843: @end example
7844:
7845: You can see what definitions are in the environment word list like this:
7846:
7847: @example
7848: get-order 1+ environment-wordlist swap set-order words previous
7849: @end example
7850:
7851:
7852: @c -------------------------------------------------------------
7853: @node Files, Blocks, Environmental Queries, Words
7854: @section Files
7855: @cindex files
7856: @cindex I/O - file-handling
7857:
7858: Gforth provides facilities for accessing files that are stored in the
7859: host operating system's file-system. Files that are processed by Gforth
7860: can be divided into two categories:
7861:
7862: @itemize @bullet
7863: @item
7864: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
7865: @item
7866: Files that are processed by some other program (@dfn{general files}).
7867: @end itemize
7868:
7869: doc-loadfilename
7870: doc-sourcefilename
7871: doc-sourceline#
7872:
7873: @menu
7874: * Forth source files::
7875: * General files::
7876: * Search Paths::
7877: @end menu
7878:
7879:
7880: @c -------------------------------------------------------------
7881: @node Forth source files, General files, Files, Files
7882: @subsection Forth source files
7883: @cindex including files
7884: @cindex Forth source files
7885:
7886: The simplest way to interpret the contents of a file is to use one of
7887: these two formats:
7888:
7889: @example
7890: include mysource.fs
7891: s" mysource.fs" included
7892: @end example
7893:
7894: Sometimes you want to include a file only if it is not included already
7895: (by, say, another source file). In that case, you can use one of these
7896: three formats:
7897:
7898: @example
7899: require mysource.fs
7900: needs mysource.fs
7901: s" mysource.fs" required
7902: @end example
7903:
7904: @cindex stack effect of included files
7905: @cindex including files, stack effect
7906: It is good practice to write your source files such that interpreting them
7907: does not change the stack. Source files designed in this way can be used with
7908: @code{required} and friends without complications. For example:
7909:
7910: @example
7911: 1 require foo.fs drop
7912: @end example
7913:
7914:
7915: doc-include-file
7916: doc-included
7917: doc-included?
7918: doc-include
7919: doc-required
7920: doc-require
7921: doc-needs
7922: doc-init-included-files
7923:
7924:
7925: A definition in ANS Forth for @code{required} is provided in
7926: @file{compat/required.fs}.
7927:
7928: @c -------------------------------------------------------------
7929: @node General files, Search Paths, Forth source files, Files
7930: @subsection General files
7931: @cindex general files
7932: @cindex file-handling
7933:
7934: Files are opened/created by name and type. The following types are
7935: recognised:
7936:
7937:
7938: doc-r/o
7939: doc-r/w
7940: doc-w/o
7941: doc-bin
7942:
7943:
7944: When a file is opened/created, it returns a file identifier,
7945: @i{wfileid} that is used for all other file commands. All file
7946: commands also return a status value, @i{wior}, that is 0 for a
7947: successful operation and an implementation-defined non-zero value in the
7948: case of an error.
7949:
7950:
7951: doc-open-file
7952: doc-create-file
7953:
7954: doc-close-file
7955: doc-delete-file
7956: doc-rename-file
7957: doc-read-file
7958: doc-read-line
7959: doc-write-file
7960: doc-write-line
7961: doc-emit-file
7962: doc-flush-file
7963:
7964: doc-file-status
7965: doc-file-position
7966: doc-reposition-file
7967: doc-file-size
7968: doc-resize-file
7969:
7970:
7971: @c ---------------------------------------------------------
7972: @node Search Paths, , General files, Files
7973: @subsection Search Paths
7974: @cindex path for @code{included}
7975: @cindex file search path
7976: @cindex @code{include} search path
7977: @cindex search path for files
7978:
7979: If you specify an absolute filename (i.e., a filename starting with
7980: @file{/} or @file{~}, or with @file{:} in the second position (as in
7981: @samp{C:...})) for @code{included} and friends, that file is included
7982: just as you would expect.
7983:
7984: For relative filenames, Gforth uses a search path similar to Forth's
7985: search order (@pxref{Word Lists}). It tries to find the given filename
7986: in the directories present in the path, and includes the first one it
7987: finds. There are separate search paths for Forth source files and
7988: general files.
7989:
7990: If the search path contains the directory @file{.} (as it should), this
7991: refers to the directory that the present file was @code{included}
7992: from. This allows files to include other files relative to their own
7993: position (irrespective of the current working directory or the absolute
7994: position). This feature is essential for libraries consisting of
7995: several files, where a file may include other files from the library.
7996: It corresponds to @code{#include "..."} in C. If the current input
7997: source is not a file, @file{.} refers to the directory of the innermost
7998: file being included, or, if there is no file being included, to the
7999: current working directory.
8000:
8001: Use @file{~+} to refer to the current working directory (as in the
8002: @code{bash}).
8003:
8004: If the filename starts with @file{./}, the search path is not searched
8005: (just as with absolute filenames), and the @file{.} has the same meaning
8006: as described above.
8007:
8008: @menu
8009: * Forth Search Paths::
8010: * General Search Paths::
8011: @end menu
8012:
8013: @c ---------------------------------------------------------
8014: @node Forth Search Paths, General Search Paths, Search Paths, Search Paths
8015: @subsubsection Forth Search Paths
8016: @cindex search path control - Forth
8017:
8018: The search path is initialized when you start Gforth (@pxref{Invoking
8019: Gforth}). You can display it and change it using these words:
8020:
8021:
8022: doc-.fpath
8023: doc-fpath+
8024: doc-fpath=
8025: doc-open-fpath-file
8026:
8027:
8028: @noindent
8029: Here is an example of using @code{fpath} and @code{require}:
8030:
8031: @example
8032: fpath= /usr/lib/forth/|./
8033: require timer.fs
8034: @end example
8035:
8036: @c ---------------------------------------------------------
8037: @node General Search Paths, , Forth Search Paths, Search Paths
8038: @subsubsection General Search Paths
8039: @cindex search path control - for user applications
8040:
8041: Your application may need to search files in several directories, like
8042: @code{included} does. To facilitate this, Gforth allows you to define
8043: and use your own search paths, by providing generic equivalents of the
8044: Forth search path words:
8045:
8046:
8047: doc-.path
8048: doc-path+
8049: doc-path=
8050: doc-open-path-file
8051:
8052:
8053: Here's an example of creating a search path:
8054:
8055: @example
8056: \ Make a buffer for the path:
8057: create mypath 100 chars , \ maximum length (is checked)
8058: 0 , \ real len
8059: 100 chars allot \ space for path
8060: @end example
8061:
8062: @c -------------------------------------------------------------
8063: @node Blocks, Other I/O, Files, Words
8064: @section Blocks
8065: @cindex I/O - blocks
8066: @cindex blocks
8067:
8068: When you run Gforth on a modern desk-top computer, it runs under the
8069: control of an operating system which provides certain services. One of
8070: these services is @var{file services}, which allows Forth source code
8071: and data to be stored in files and read into Gforth (@pxref{Files}).
8072:
8073: Traditionally, Forth has been an important programming language on
8074: systems where it has interfaced directly to the underlying hardware with
8075: no intervening operating system. Forth provides a mechanism, called
8076: @dfn{blocks}, for accessing mass storage on such systems.
8077:
8078: A block is a 1024-byte data area, which can be used to hold data or
8079: Forth source code. No structure is imposed on the contents of the
8080: block. A block is identified by its number; blocks are numbered
8081: contiguously from 1 to an implementation-defined maximum.
8082:
8083: A typical system that used blocks but no operating system might use a
8084: single floppy-disk drive for mass storage, with the disks formatted to
8085: provide 256-byte sectors. Blocks would be implemented by assigning the
8086: first four sectors of the disk to block 1, the second four sectors to
8087: block 2 and so on, up to the limit of the capacity of the disk. The disk
8088: would not contain any file system information, just the set of blocks.
8089:
8090: @cindex blocks file
8091: On systems that do provide file services, blocks are typically
8092: implemented by storing a sequence of blocks within a single @dfn{blocks
8093: file}. The size of the blocks file will be an exact multiple of 1024
8094: bytes, corresponding to the number of blocks it contains. This is the
8095: mechanism that Gforth uses.
8096:
8097: @cindex @file{blocks.fb}
8098: Only 1 blocks file can be open at a time. If you use block words without
8099: having specified a blocks file, Gforth defaults to the blocks file
8100: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
8101: locate a blocks file (@pxref{Forth Search Paths}).
8102:
8103: @cindex block buffers
8104: When you read and write blocks under program control, Gforth uses a
8105: number of @dfn{block buffers} as intermediate storage. These buffers are
8106: not used when you use @code{load} to interpret the contents of a block.
8107:
8108: The behaviour of the block buffers is directly analagous to that of a
8109: cache. Each block buffer has three states:
8110:
8111: @itemize @bullet
8112: @item
8113: Unassigned
8114: @item
8115: Assigned-clean
8116: @item
8117: Assigned-dirty
8118: @end itemize
8119:
8120: Initially, all block buffers are @i{unassigned}. In order to access a
8121: block, the block (specified by its block number) must be assigned to a
8122: block buffer.
8123:
8124: The assignment of a block to a block buffer is performed by @code{block}
8125: or @code{buffer}. Use @code{block} when you wish to modify the existing
8126: contents of a block. Use @code{buffer} when you don't care about the
8127: existing contents of the block@footnote{The ANS Forth definition of
8128: @code{buffer} is intended not to cause disk I/O; if the data associated
8129: with the particular block is already stored in a block buffer due to an
8130: earlier @code{block} command, @code{buffer} will return that block
8131: buffer and the existing contents of the block will be
8132: available. Otherwise, @code{buffer} will simply assign a new, empty
8133: block buffer for the block.}.
8134:
8135: Once a block has been assigned to a block buffer using @code{block} or
8136: @code{buffer}, that block buffer becomes the @i{current block buffer}
8137: and its state changes to @i{assigned-clean}. Data may only be
8138: manipulated (read or written) within the current block buffer.
8139:
8140: When the contents of the current block buffer has been modified it is
8141: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
8142: either abandon the changes (by doing nothing) or commit the changes,
8143: using @code{update}. Using @code{update} does not change the blocks
8144: file; it simply changes a block buffer's state to @i{assigned-dirty}.
8145:
8146: The word @code{flush} causes all @i{assigned-dirty} blocks to be
8147: written back to the blocks file on disk. Leaving Gforth using @code{bye}
8148: also causes a @code{flush} to be performed.
8149:
8150: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
8151: algorithm to assign a block buffer to a block. That means that any
8152: particular block can only be assigned to one specific block buffer,
8153: called (for the particular operation) the @i{victim buffer}. If the
8154: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8155: the new block immediately. If it is @i{assigned-dirty} its current
8156: contents are written back to the blocks file on disk before it is
8157: allocated to the new block.
8158:
8159: Although no structure is imposed on the contents of a block, it is
8160: traditional to display the contents as 16 lines each of 64 characters. A
8161: block provides a single, continuous stream of input (for example, it
8162: acts as a single parse area) -- there are no end-of-line characters
8163: within a block, and no end-of-file character at the end of a
8164: block. There are two consequences of this:
8165:
8166: @itemize @bullet
8167: @item
8168: The last character of one line wraps straight into the first character
8169: of the following line
8170: @item
8171: The word @code{\} -- comment to end of line -- requires special
8172: treatment; in the context of a block it causes all characters until the
8173: end of the current 64-character ``line'' to be ignored.
8174: @end itemize
8175:
8176: In Gforth, when you use @code{block} with a non-existent block number,
8177: the current blocks file will be extended to the appropriate size and the
8178: block buffer will be initialised with spaces.
8179:
8180: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8181: for details) but doesn't encourage the use of blocks; the mechanism is
8182: only provided for backward compatibility -- ANS Forth requires blocks to
8183: be available when files are.
8184:
8185: Common techniques that are used when working with blocks include:
8186:
8187: @itemize @bullet
8188: @item
8189: A screen editor that allows you to edit blocks without leaving the Forth
8190: environment.
8191: @item
8192: Shadow screens; where every code block has an associated block
8193: containing comments (for example: code in odd block numbers, comments in
8194: even block numbers). Typically, the block editor provides a convenient
8195: mechanism to toggle between code and comments.
8196: @item
8197: Load blocks; a single block (typically block 1) contains a number of
8198: @code{thru} commands which @code{load} the whole of the application.
8199: @end itemize
8200:
8201: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8202: integrated into a Forth programming environment.
8203:
8204: @comment TODO what about errors on open-blocks?
8205:
8206: doc-open-blocks
8207: doc-use
8208: doc-get-block-fid
8209: doc-block-position
8210:
8211: doc-scr
8212: doc-list
8213:
8214: doc---gforthman-block
8215: doc-buffer
8216:
8217: doc-update
8218: doc-updated?
8219: doc-save-buffers
8220: doc-empty-buffers
8221: doc-empty-buffer
8222: doc-flush
8223:
8224: doc-load
8225: doc-thru
8226: doc-+load
8227: doc-+thru
8228: doc---gforthman--->
8229: doc-block-included
8230:
8231:
8232: @c -------------------------------------------------------------
8233: @node Other I/O, Programming Tools, Blocks, Words
8234: @section Other I/O
8235: @cindex I/O - keyboard and display
8236:
8237: @menu
8238: * Simple numeric output:: Predefined formats
8239: * Formatted numeric output:: Formatted (pictured) output
8240: * String Formats:: How Forth stores strings in memory
8241: * Displaying characters and strings:: Other stuff
8242: * Input:: Input
8243: @end menu
8244:
8245: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8246: @subsection Simple numeric output
8247: @cindex numeric output - simple/free-format
8248:
8249: The simplest output functions are those that display numbers from the
8250: data or floating-point stacks. Floating-point output is always displayed
8251: using base 10. Numbers displayed from the data stack use the value stored
8252: in @code{base}.
8253:
8254:
8255: doc-.
8256: doc-dec.
8257: doc-hex.
8258: doc-u.
8259: doc-.r
8260: doc-u.r
8261: doc-d.
8262: doc-ud.
8263: doc-d.r
8264: doc-ud.r
8265: doc-f.
8266: doc-fe.
8267: doc-fs.
8268:
8269:
8270: Examples of printing the number 1234.5678E23 in the different floating-point output
8271: formats are shown below:
8272:
8273: @example
8274: f. 123456779999999000000000000.
8275: fe. 123.456779999999E24
8276: fs. 1.23456779999999E26
8277: @end example
8278:
8279:
8280: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8281: @subsection Formatted numeric output
8282: @cindex formatted numeric output
8283: @cindex pictured numeric output
8284: @cindex numeric output - formatted
8285:
8286: Forth traditionally uses a technique called @dfn{pictured numeric
8287: output} for formatted printing of integers. In this technique, digits
8288: are extracted from the number (using the current output radix defined by
8289: @code{base}), converted to ASCII codes and appended to a string that is
8290: built in a scratch-pad area of memory (@pxref{core-idef,
8291: Implementation-defined options, Implementation-defined
8292: options}). Arbitrary characters can be appended to the string during the
8293: extraction process. The completed string is specified by an address
8294: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8295: under program control.
8296:
8297: All of the words described in the previous section for simple numeric
8298: output are implemented in Gforth using pictured numeric output.
8299:
8300: Three important things to remember about pictured numeric output:
8301:
8302: @itemize @bullet
8303: @item
8304: It always operates on double-precision numbers; to display a
8305: single-precision number, convert it first (for ways of doing this
8306: @pxref{Double precision}).
8307: @item
8308: It always treats the double-precision number as though it were
8309: unsigned. The examples below show ways of printing signed numbers.
8310: @item
8311: The string is built up from right to left; least significant digit first.
8312: @end itemize
8313:
8314:
8315: doc-<#
8316: doc-<<#
8317: doc-#
8318: doc-#s
8319: doc-hold
8320: doc-sign
8321: doc-#>
8322: doc-#>>
8323:
8324: doc-represent
8325:
8326:
8327: @noindent
8328: Here are some examples of using pictured numeric output:
8329:
8330: @example
8331: : my-u. ( u -- )
8332: \ Simplest use of pns.. behaves like Standard u.
8333: 0 \ convert to unsigned double
8334: <# \ start conversion
8335: #s \ convert all digits
8336: #> \ complete conversion
8337: TYPE SPACE ; \ display, with trailing space
8338:
8339: : cents-only ( u -- )
8340: 0 \ convert to unsigned double
8341: <# \ start conversion
8342: # # \ convert two least-significant digits
8343: #> \ complete conversion, discard other digits
8344: TYPE SPACE ; \ display, with trailing space
8345:
8346: : dollars-and-cents ( u -- )
8347: 0 \ convert to unsigned double
8348: <# \ start conversion
8349: # # \ convert two least-significant digits
8350: [char] . hold \ insert decimal point
8351: #s \ convert remaining digits
8352: [char] $ hold \ append currency symbol
8353: #> \ complete conversion
8354: TYPE SPACE ; \ display, with trailing space
8355:
8356: : my-. ( n -- )
8357: \ handling negatives.. behaves like Standard .
8358: s>d \ convert to signed double
8359: swap over dabs \ leave sign byte followed by unsigned double
8360: <# \ start conversion
8361: #s \ convert all digits
8362: rot sign \ get at sign byte, append "-" if needed
8363: #> \ complete conversion
8364: TYPE SPACE ; \ display, with trailing space
8365:
8366: : account. ( n -- )
8367: \ accountants don't like minus signs, they use braces
8368: \ for negative numbers
8369: s>d \ convert to signed double
8370: swap over dabs \ leave sign byte followed by unsigned double
8371: <# \ start conversion
8372: 2 pick \ get copy of sign byte
8373: 0< IF [char] ) hold THEN \ right-most character of output
8374: #s \ convert all digits
8375: rot \ get at sign byte
8376: 0< IF [char] ( hold THEN
8377: #> \ complete conversion
8378: TYPE SPACE ; \ display, with trailing space
8379: @end example
8380:
8381: Here are some examples of using these words:
8382:
8383: @example
8384: 1 my-u. 1
8385: hex -1 my-u. decimal FFFFFFFF
8386: 1 cents-only 01
8387: 1234 cents-only 34
8388: 2 dollars-and-cents $0.02
8389: 1234 dollars-and-cents $12.34
8390: 123 my-. 123
8391: -123 my. -123
8392: 123 account. 123
8393: -456 account. (456)
8394: @end example
8395:
8396:
8397: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8398: @subsection String Formats
8399: @cindex strings - see character strings
8400: @cindex character strings - formats
8401: @cindex I/O - see character strings
8402:
8403: Forth commonly uses two different methods for representing character
8404: strings:
8405:
8406: @itemize @bullet
8407: @item
8408: @cindex address of counted string
8409: @cindex counted string
8410: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8411: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8412: string and the string occupies the subsequent @i{n} char addresses in
8413: memory.
8414: @item
8415: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8416: of the string in characters, and @i{c-addr} is the address of the
8417: first byte of the string.
8418: @end itemize
8419:
8420: ANS Forth encourages the use of the second format when representing
8421: strings on the stack, whilst conceeding that the counted string format
8422: remains useful as a way of storing strings in memory.
8423:
8424:
8425: doc-count
8426:
8427:
8428: For words that move, copy and search for strings see @ref{Memory
8429: Blocks}. For words that display characters and strings see
8430: @ref{Displaying characters and strings}.
8431:
8432: @node Displaying characters and strings, Input, String Formats, Other I/O
8433: @subsection Displaying characters and strings
8434: @cindex characters - compiling and displaying
8435: @cindex character strings - compiling and displaying
8436:
8437: This section starts with a glossary of Forth words and ends with a set
8438: of examples.
8439:
8440:
8441: doc-bl
8442: doc-space
8443: doc-spaces
8444: doc-emit
8445: doc-toupper
8446: doc-."
8447: doc-.(
8448: doc-type
8449: doc-typewhite
8450: doc-cr
8451: @cindex cursor control
8452: doc-at-xy
8453: doc-page
8454: doc-s"
8455: doc-c"
8456: doc-char
8457: doc-[char]
8458: doc-sliteral
8459:
8460:
8461: @noindent
8462: As an example, consider the following text, stored in a file @file{test.fs}:
8463:
8464: @example
8465: .( text-1)
8466: : my-word
8467: ." text-2" cr
8468: .( text-3)
8469: ;
8470:
8471: ." text-4"
8472:
8473: : my-char
8474: [char] ALPHABET emit
8475: char emit
8476: ;
8477: @end example
8478:
8479: When you load this code into Gforth, the following output is generated:
8480:
8481: @example
8482: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
8483: @end example
8484:
8485: @itemize @bullet
8486: @item
8487: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
8488: is an immediate word; it behaves in the same way whether it is used inside
8489: or outside a colon definition.
8490: @item
8491: Message @code{text-4} is displayed because of Gforth's added interpretation
8492: semantics for @code{."}.
8493: @item
8494: Message @code{text-2} is @i{not} displayed, because the text interpreter
8495: performs the compilation semantics for @code{."} within the definition of
8496: @code{my-word}.
8497: @end itemize
8498:
8499: Here are some examples of executing @code{my-word} and @code{my-char}:
8500:
8501: @example
8502: @kbd{my-word @key{RET}} text-2
8503: ok
8504: @kbd{my-char fred @key{RET}} Af ok
8505: @kbd{my-char jim @key{RET}} Aj ok
8506: @end example
8507:
8508: @itemize @bullet
8509: @item
8510: Message @code{text-2} is displayed because of the run-time behaviour of
8511: @code{."}.
8512: @item
8513: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
8514: on the stack at run-time. @code{emit} always displays the character
8515: when @code{my-char} is executed.
8516: @item
8517: @code{char} parses a string at run-time and the second @code{emit} displays
8518: the first character of the string.
8519: @item
8520: If you type @code{see my-char} you can see that @code{[char]} discarded
8521: the text ``LPHABET'' and only compiled the display code for ``A'' into the
8522: definition of @code{my-char}.
8523: @end itemize
8524:
8525:
8526:
8527: @node Input, , Displaying characters and strings, Other I/O
8528: @subsection Input
8529: @cindex input
8530: @cindex I/O - see input
8531: @cindex parsing a string
8532:
8533: For ways of storing character strings in memory see @ref{String Formats}.
8534:
8535: @comment TODO examples for >number >float accept key key? pad parse word refill
8536: @comment then index them
8537:
8538:
8539: doc-key
8540: doc-key?
8541: doc-ekey
8542: doc-ekey?
8543: doc-ekey>char
8544: doc->number
8545: doc->float
8546: doc-accept
8547: doc-pad
8548: doc-parse
8549: doc-word
8550: doc-sword
8551: doc-(name)
8552: doc-refill
8553: @comment obsolescent words..
8554: doc-convert
8555: doc-query
8556: doc-expect
8557: doc-span
8558:
8559:
8560:
8561: @c -------------------------------------------------------------
8562: @node Programming Tools, Assembler and Code Words, Other I/O, Words
8563: @section Programming Tools
8564: @cindex programming tools
8565:
8566: @menu
8567: * Debugging:: Simple and quick.
8568: * Assertions:: Making your programs self-checking.
8569: * Singlestep Debugger:: Executing your program word by word.
8570: @end menu
8571:
8572: @node Debugging, Assertions, Programming Tools, Programming Tools
8573: @subsection Debugging
8574: @cindex debugging
8575:
8576: Languages with a slow edit/compile/link/test development loop tend to
8577: require sophisticated tracing/stepping debuggers to facilate
8578: productive debugging.
8579:
8580: A much better (faster) way in fast-compiling languages is to add
8581: printing code at well-selected places, let the program run, look at
8582: the output, see where things went wrong, add more printing code, etc.,
8583: until the bug is found.
8584:
8585: The simple debugging aids provided in @file{debugs.fs}
8586: are meant to support this style of debugging. In addition, there are
8587: words for non-destructively inspecting the stack and memory:
8588:
8589:
8590: doc-.s
8591: doc-f.s
8592:
8593:
8594: There is a word @code{.r} but it does @i{not} display the return
8595: stack! It is used for formatted numeric output.
8596:
8597:
8598: doc-depth
8599: doc-fdepth
8600: doc-clearstack
8601: doc-?
8602: doc-dump
8603:
8604:
8605: The word @code{~~} prints debugging information (by default the source
8606: location and the stack contents). It is easy to insert. If you use Emacs
8607: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
8608: query-replace them with nothing). The deferred words
8609: @code{printdebugdata} and @code{printdebugline} control the output of
8610: @code{~~}. The default source location output format works well with
8611: Emacs' compilation mode, so you can step through the program at the
8612: source level using @kbd{C-x `} (the advantage over a stepping debugger
8613: is that you can step in any direction and you know where the crash has
8614: happened or where the strange data has occurred).
8615:
8616: The default actions of @code{~~} clobber the contents of the pictured
8617: numeric output string, so you should not use @code{~~}, e.g., between
8618: @code{<#} and @code{#>}.
8619:
8620:
8621: doc-~~
8622: doc-printdebugdata
8623: doc-printdebugline
8624:
8625: doc-see
8626: doc-marker
8627:
8628:
8629: Here's an example of using @code{marker} at the start of a source file
8630: that you are debugging; it ensures that you only ever have one copy of
8631: the file's definitions compiled at any time:
8632:
8633: @example
8634: [IFDEF] my-code
8635: my-code
8636: [ENDIF]
8637:
8638: marker my-code
8639: init-included-files
8640:
8641: \ .. definitions start here
8642: \ .
8643: \ .
8644: \ end
8645: @end example
8646:
8647:
8648:
8649: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
8650: @subsection Assertions
8651: @cindex assertions
8652:
8653: It is a good idea to make your programs self-checking, especially if you
8654: make an assumption that may become invalid during maintenance (for
8655: example, that a certain field of a data structure is never zero). Gforth
8656: supports @dfn{assertions} for this purpose. They are used like this:
8657:
8658: @example
8659: assert( @i{flag} )
8660: @end example
8661:
8662: The code between @code{assert(} and @code{)} should compute a flag, that
8663: should be true if everything is alright and false otherwise. It should
8664: not change anything else on the stack. The overall stack effect of the
8665: assertion is @code{( -- )}. E.g.
8666:
8667: @example
8668: assert( 1 1 + 2 = ) \ what we learn in school
8669: assert( dup 0<> ) \ assert that the top of stack is not zero
8670: assert( false ) \ this code should not be reached
8671: @end example
8672:
8673: The need for assertions is different at different times. During
8674: debugging, we want more checking, in production we sometimes care more
8675: for speed. Therefore, assertions can be turned off, i.e., the assertion
8676: becomes a comment. Depending on the importance of an assertion and the
8677: time it takes to check it, you may want to turn off some assertions and
8678: keep others turned on. Gforth provides several levels of assertions for
8679: this purpose:
8680:
8681:
8682: doc-assert0(
8683: doc-assert1(
8684: doc-assert2(
8685: doc-assert3(
8686: doc-assert(
8687: doc-)
8688:
8689:
8690: The variable @code{assert-level} specifies the highest assertions that
8691: are turned on. I.e., at the default @code{assert-level} of one,
8692: @code{assert0(} and @code{assert1(} assertions perform checking, while
8693: @code{assert2(} and @code{assert3(} assertions are treated as comments.
8694:
8695: The value of @code{assert-level} is evaluated at compile-time, not at
8696: run-time. Therefore you cannot turn assertions on or off at run-time;
8697: you have to set the @code{assert-level} appropriately before compiling a
8698: piece of code. You can compile different pieces of code at different
8699: @code{assert-level}s (e.g., a trusted library at level 1 and
8700: newly-written code at level 3).
8701:
8702:
8703: doc-assert-level
8704:
8705:
8706: If an assertion fails, a message compatible with Emacs' compilation mode
8707: is produced and the execution is aborted (currently with @code{ABORT"}.
8708: If there is interest, we will introduce a special throw code. But if you
8709: intend to @code{catch} a specific condition, using @code{throw} is
8710: probably more appropriate than an assertion).
8711:
8712: Definitions in ANS Forth for these assertion words are provided
8713: in @file{compat/assert.fs}.
8714:
8715:
8716: @node Singlestep Debugger, , Assertions, Programming Tools
8717: @subsection Singlestep Debugger
8718: @cindex singlestep Debugger
8719: @cindex debugging Singlestep
8720:
8721: When you create a new word there's often the need to check whether it
8722: behaves correctly or not. You can do this by typing @code{dbg
8723: badword}. A debug session might look like this:
8724:
8725: @example
8726: : badword 0 DO i . LOOP ; ok
8727: 2 dbg badword
8728: : badword
8729: Scanning code...
8730:
8731: Nesting debugger ready!
8732:
8733: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
8734: 400D4740 8049F68 DO -> [ 0 ]
8735: 400D4744 804A0C8 i -> [ 1 ] 00000
8736: 400D4748 400C5E60 . -> 0 [ 0 ]
8737: 400D474C 8049D0C LOOP -> [ 0 ]
8738: 400D4744 804A0C8 i -> [ 1 ] 00001
8739: 400D4748 400C5E60 . -> 1 [ 0 ]
8740: 400D474C 8049D0C LOOP -> [ 0 ]
8741: 400D4758 804B384 ; -> ok
8742: @end example
8743:
8744: Each line displayed is one step. You always have to hit return to
8745: execute the next word that is displayed. If you don't want to execute
8746: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
8747: an overview what keys are available:
8748:
8749: @table @i
8750:
8751: @item @key{RET}
8752: Next; Execute the next word.
8753:
8754: @item n
8755: Nest; Single step through next word.
8756:
8757: @item u
8758: Unnest; Stop debugging and execute rest of word. If we got to this word
8759: with nest, continue debugging with the calling word.
8760:
8761: @item d
8762: Done; Stop debugging and execute rest.
8763:
8764: @item s
8765: Stop; Abort immediately.
8766:
8767: @end table
8768:
8769: Debugging large application with this mechanism is very difficult, because
8770: you have to nest very deeply into the program before the interesting part
8771: begins. This takes a lot of time.
8772:
8773: To do it more directly put a @code{BREAK:} command into your source code.
8774: When program execution reaches @code{BREAK:} the single step debugger is
8775: invoked and you have all the features described above.
8776:
8777: If you have more than one part to debug it is useful to know where the
8778: program has stopped at the moment. You can do this by the
8779: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
8780: string is typed out when the ``breakpoint'' is reached.
8781:
8782:
8783: doc-dbg
8784: doc-break:
8785: doc-break"
8786:
8787:
8788:
8789: @c -------------------------------------------------------------
8790: @node Assembler and Code Words, Threading Words, Programming Tools, Words
8791: @section Assembler and Code Words
8792: @cindex assembler
8793: @cindex code words
8794:
8795: @menu
8796: * Code and ;code::
8797: * Common Assembler:: Assembler Syntax
8798: * Common Disassembler::
8799: * 386 Assembler:: Deviations and special cases
8800: * Alpha Assembler:: Deviations and special cases
8801: * MIPS assembler:: Deviations and special cases
8802: * Other assemblers:: How to write them
8803: @end menu
8804:
8805: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
8806: @subsection @code{Code} and @code{;code}
8807:
8808: Gforth provides some words for defining primitives (words written in
8809: machine code), and for defining the machine-code equivalent of
8810: @code{DOES>}-based defining words. However, the machine-independent
8811: nature of Gforth poses a few problems: First of all, Gforth runs on
8812: several architectures, so it can provide no standard assembler. What's
8813: worse is that the register allocation not only depends on the processor,
8814: but also on the @code{gcc} version and options used.
8815:
8816: The words that Gforth offers encapsulate some system dependences (e.g.,
8817: the header structure), so a system-independent assembler may be used in
8818: Gforth. If you do not have an assembler, you can compile machine code
8819: directly with @code{,} and @code{c,}@footnote{This isn't portable,
8820: because these words emit stuff in @i{data} space; it works because
8821: Gforth has unified code/data spaces. Assembler isn't likely to be
8822: portable anyway.}.
8823:
8824:
8825: doc-assembler
8826: doc-init-asm
8827: doc-code
8828: doc-end-code
8829: doc-;code
8830: doc-flush-icache
8831:
8832:
8833: If @code{flush-icache} does not work correctly, @code{code} words
8834: etc. will not work (reliably), either.
8835:
8836: The typical usage of these @code{code} words can be shown most easily by
8837: analogy to the equivalent high-level defining words:
8838:
8839: @example
8840: : foo code foo
8841: <high-level Forth words> <assembler>
8842: ; end-code
8843:
8844: : bar : bar
8845: <high-level Forth words> <high-level Forth words>
8846: CREATE CREATE
8847: <high-level Forth words> <high-level Forth words>
8848: DOES> ;code
8849: <high-level Forth words> <assembler>
8850: ; end-code
8851: @end example
8852:
8853: @code{flush-icache} is always present. The other words are rarely used
8854: and reside in @code{code.fs}, which is usually not loaded. You can load
8855: it with @code{require code.fs}.
8856:
8857: @cindex registers of the inner interpreter
8858: In the assembly code you will want to refer to the inner interpreter's
8859: registers (e.g., the data stack pointer) and you may want to use other
8860: registers for temporary storage. Unfortunately, the register allocation
8861: is installation-dependent.
8862:
8863: The easiest solution is to use explicit register declarations
8864: (@pxref{Explicit Reg Vars, , Variables in Specified Registers, gcc.info,
8865: GNU C Manual}) for all of the inner interpreter's registers: You have to
8866: compile Gforth with @code{-DFORCE_REG} (configure option
8867: @code{--enable-force-reg}) and the appropriate declarations must be
8868: present in the @code{machine.h} file (see @code{mips.h} for an example;
8869: you can find a full list of all declarable register symbols with
8870: @code{grep register engine.c}). If you give explicit registers to all
8871: variables that are declared at the beginning of @code{engine()}, you
8872: should be able to use the other caller-saved registers for temporary
8873: storage. Alternatively, you can use the @code{gcc} option
8874: @code{-ffixed-REG} (@pxref{Code Gen Options, , Options for Code
8875: Generation Conventions, gcc.info, GNU C Manual}) to reserve a register
8876: (however, this restriction on register allocation may slow Gforth
8877: significantly).
8878:
8879: If this solution is not viable (e.g., because @code{gcc} does not allow
8880: you to explicitly declare all the registers you need), you have to find
8881: out by looking at the code where the inner interpreter's registers
8882: reside and which registers can be used for temporary storage. You can
8883: get an assembly listing of the engine's code with @code{make engine.s}.
8884:
8885: In any case, it is good practice to abstract your assembly code from the
8886: actual register allocation. E.g., if the data stack pointer resides in
8887: register @code{$17}, create an alias for this register called @code{sp},
8888: and use that in your assembly code.
8889:
8890: @cindex code words, portable
8891: Another option for implementing normal and defining words efficiently
8892: is to add the desired functionality to the source of Gforth. For normal
8893: words you just have to edit @file{primitives} (@pxref{Automatic
8894: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
8895: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
8896: @file{prims2x.fs}, and possibly @file{cross.fs}.
8897:
8898: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
8899: @subsection Common Assembler
8900:
8901: The assemblers in Gforth generally use a postfix syntax, i.e., the
8902: instruction name follows the operands.
8903:
8904: The operands are passed in the usual order (the same that is used in the
8905: manual of the architecture). Since they all are Forth words, they have
8906: to be separated by spaces; you can also use Forth words to compute the
8907: operands.
8908:
8909: The instruction names usually end with a @code{,}. This makes it easier
8910: to visually separate instructions if you put several of them on one
8911: line; it also avoids shadowing other Forth words (e.g., @code{and}).
8912:
8913: Registers are usually specified by number; e.g., (decimal) @code{11}
8914: specifies registers R11 and F11 on the Alpha architecture (which one,
8915: depends on the instruction). The usual names are also available, e.g.,
8916: @code{s2} for R11 on Alpha.
8917:
8918: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
8919: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
8920: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
8921: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
8922: conditions are specified in a way specific to each assembler.
8923:
8924: Note that the register assignments of the Gforth engine can change
8925: between Gforth versions, or even between different compilations of the
8926: same Gforth version (e.g., if you use a different GCC version). So if
8927: you want to refer to Gforth's registers (e.g., the stack pointer or
8928: TOS), I recommend defining your own words for refering to these
8929: registers, and using them later on; then you can easily adapt to a
8930: changed register assignment. The stability of the register assignment
8931: is usually better if you build Gforth with @code{--enable-force-reg}.
8932:
8933: In particular, the resturn stack pointer and the instruction pointer are
8934: in memory in @code{gforth}, and usually in registers in
8935: @code{gforth-fast}. The most common use of these registers is to
8936: dispatch to the next word (the @code{next} routine). A portable way to
8937: do this is to jump to @code{' noop >code-address} (of course, this is
8938: less efficient than integrating the @code{next} code and scheduling it
8939: well).
8940:
8941: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
8942: @subsection Common Disassembler
8943:
8944: You can disassemble a @code{code} word with @code{see}
8945: (@pxref{Debugging}). You can disassemble a section of memory with
8946:
8947: doc-disasm
8948:
8949: The disassembler generally produces output that can be fed into the
8950: assembler (i.e., same syntax, etc.). It also includes additional
8951: information in comments. In particular, the address of the instruction
8952: is given in a comment before the instruction.
8953:
8954: @code{See} may display more or less than the actual code of the word,
8955: because the recognition of the end of the code is unreliable. You can
8956: use @code{disasm} if it did not display enough. It may display more, if
8957: the code word is not immediately followed by a named word. If you have
8958: something else there, you can follow the word with @code{align last @ ,}
8959: to ensure that the end is recognized.
8960:
8961: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
8962: @subsection 386 Assembler
8963:
8964: The 386 assembler included in Gforth was written by Bernd Paysan, it's
8965: available under GPL, and originally part of bigFORTH.
8966:
8967: The 386 disassembler included in Gforth was written by Andrew McKewan
8968: and is in the public domain.
8969:
8970: The disassembler displays code in prefix Intel syntax.
8971:
8972: The assembler uses a postfix syntax with reversed parameters.
8973:
8974: The assembler includes all instruction of the Athlon, i.e. 486 core
8975: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
8976: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
8977: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
8978:
8979: There are several prefixes to switch between different operation sizes,
8980: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
8981: double-word accesses. Addressing modes can be switched with @code{.wa}
8982: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
8983: need a prefix for byte register names (@code{AL} et al).
8984:
8985: For floating point operations, the prefixes are @code{.fs} (IEEE
8986: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
8987: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
8988:
8989: The MMX opcodes don't have size prefixes, they are spelled out like in
8990: the Intel assembler. Instead of move from and to memory, there are
8991: PLDQ/PLDD and PSTQ/PSTD.
8992:
8993: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
8994: ax. Immediate values are indicated by postfixing them with @code{#},
8995: e.g., @code{3 #}. Here are some examples of addressing modes:
8996:
8997: @example
8998: 3 # \ immediate
8999: ax \ register
9000: 100 di d) \ 100[edi]
9001: 4 bx cx di) \ 4[ebx][ecx]
9002: di ax *4 i) \ [edi][eax*4]
9003: 20 ax *4 i#) \ 20[eax*4]
9004: @end example
9005:
9006: Some example of instructions are:
9007:
9008: @example
9009: ax bx mov \ move ebx,eax
9010: 3 # ax mov \ mov eax,3
9011: 100 di ) ax mov \ mov eax,100[edi]
9012: 4 bx cx di) ax mov \ mov eax,4[ebx][ecx]
9013: .w ax bx mov \ mov bx,ax
9014: @end example
9015:
9016: The following forms are supported for binary instructions:
9017:
9018: @example
9019: <reg> <reg> <inst>
9020: <n> # <reg> <inst>
9021: <mem> <reg> <inst>
9022: <reg> <mem> <inst>
9023: @end example
9024:
9025: Immediate to memory is not supported. The shift/rotate syntax is:
9026:
9027: @example
9028: <reg/mem> 1 # shl \ shortens to shift without immediate
9029: <reg/mem> 4 # shl
9030: <reg/mem> cl shl
9031: @end example
9032:
9033: Precede string instructions (@code{movs} etc.) with @code{.b} to get
9034: the byte version.
9035:
9036: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
9037: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
9038: pc < >= <= >}. (Note that most of these words shadow some Forth words
9039: when @code{assembler} is in front of @code{forth} in the search path,
9040: e.g., in @code{code} words). Currently the control structure words use
9041: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
9042: to shuffle them (you can also use @code{swap} etc.).
9043:
9044: Here is an example of a @code{code} word (assumes that the stack pointer
9045: is in esi and the TOS is in ebx):
9046:
9047: @example
9048: code my+ ( n1 n2 -- n )
9049: 4 si D) bx add
9050: 4 # si add
9051: Next
9052: end-code
9053: @end example
9054:
9055: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
9056: @subsection Alpha Assembler
9057:
9058: The Alpha assembler and disassembler were originally written by Bernd
9059: Thallner.
9060:
9061: The register names @code{a0}--@code{a5} are not available to avoid
9062: shadowing hex numbers.
9063:
9064: Immediate forms of arithmetic instructions are distinguished by a
9065: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
9066: does not count as arithmetic instruction).
9067:
9068: You have to specify all operands to an instruction, even those that
9069: other assemblers consider optional, e.g., the destination register for
9070: @code{br,}, or the destination register and hint for @code{jmp,}.
9071:
9072: You can specify conditions for @code{if,} by removing the first @code{b}
9073: and the trailing @code{,} from a branch with a corresponding name; e.g.,
9074:
9075: @example
9076: 11 fgt if, \ if F11>0e
9077: ...
9078: endif,
9079: @end example
9080:
9081: @code{fbgt,} gives @code{fgt}.
9082:
9083: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
9084: @subsection MIPS assembler
9085:
9086: The MIPS assembler was originally written by Christian Pirker.
9087:
9088: Currently the assembler and disassembler only cover the MIPS-I
9089: architecture (R3000), and don't support FP instructions.
9090:
9091: The register names @code{$a0}--@code{$a3} are not available to avoid
9092: shadowing hex numbers.
9093:
9094: Because there is no way to distinguish registers from immediate values,
9095: you have to explicitly use the immediate forms of instructions, i.e.,
9096: @code{addiu,}, not just @code{addu,} (@command{as} does this
9097: implicitly).
9098:
9099: If the architecture manual specifies several formats for the instruction
9100: (e.g., for @code{jalr,}), you usually have to use the one with more
9101: arguments (i.e., two for @code{jalr,}). When in doubt, see
9102: @code{arch/mips/testasm.fs} for an example of correct use.
9103:
9104: Branches and jumps in the MIPS architecture have a delay slot. You have
9105: to fill it yourself (the simplest way is to use @code{nop,}), the
9106: assembler does not do it for you (unlike @command{as}). Even
9107: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
9108: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
9109: and @code{then,} just specify branch targets, they are not affected.
9110:
9111: Note that you must not put branches, jumps, or @code{li,} into the delay
9112: slot: @code{li,} may expand to several instructions, and control flow
9113: instructions may not be put into the branch delay slot in any case.
9114:
9115: For branches the argument specifying the target is a relative address;
9116: You have to add the address of the delay slot to get the absolute
9117: address.
9118:
9119: The MIPS architecture also has load delay slots and restrictions on
9120: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
9121: yourself to satisfy these restrictions, the assembler does not do it for
9122: you.
9123:
9124: You can specify the conditions for @code{if,} etc. by taking a
9125: conditional branch and leaving away the @code{b} at the start and the
9126: @code{,} at the end. E.g.,
9127:
9128: @example
9129: 4 5 eq if,
9130: ... \ do something if $4 equals $5
9131: then,
9132: @end example
9133:
9134: @node Other assemblers, , MIPS assembler, Assembler and Code Words
9135: @subsection Other assemblers
9136:
9137: If you want to contribute another assembler/disassembler, please contact
9138: us (@email{bug-gforth@@gnu.org}) to check if we have such an assembler
9139: already. If you are writing them from scratch, please use a similar
9140: syntax style as the one we use (i.e., postfix, commas at the end of the
9141: instruction names, @pxref{Common Assembler}); make the output of the
9142: disassembler be valid input for the assembler, and keep the style
9143: similar to the style we used.
9144:
9145: Hints on implementation: The most important part is to have a good test
9146: suite that contains all instructions. Once you have that, the rest is
9147: easy. For actual coding you can take a look at
9148: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
9149: the assembler and disassembler, avoiding redundancy and some potential
9150: bugs. You can also look at that file (and @pxref{Advanced does> usage
9151: example}) to get ideas how to factor a disassembler.
9152:
9153: Start with the disassembler, because it's easier to reuse data from the
9154: disassembler for the assembler than the other way round.
9155:
9156: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
9157: how simple it can be.
9158:
9159: @c -------------------------------------------------------------
9160: @node Threading Words, Locals, Assembler and Code Words, Words
9161: @section Threading Words
9162: @cindex threading words
9163:
9164: @cindex code address
9165: These words provide access to code addresses and other threading stuff
9166: in Gforth (and, possibly, other interpretive Forths). It more or less
9167: abstracts away the differences between direct and indirect threading
9168: (and, for direct threading, the machine dependences). However, at
9169: present this wordset is still incomplete. It is also pretty low-level;
9170: some day it will hopefully be made unnecessary by an internals wordset
9171: that abstracts implementation details away completely.
9172:
9173:
9174: doc-threading-method
9175: doc->code-address
9176: doc->does-code
9177: doc-code-address!
9178: doc-does-code!
9179: doc-does-handler!
9180: doc-/does-handler
9181:
9182:
9183: The code addresses produced by various defining words are produced by
9184: the following words:
9185:
9186:
9187: doc-docol:
9188: doc-docon:
9189: doc-dovar:
9190: doc-douser:
9191: doc-dodefer:
9192: doc-dofield:
9193:
9194:
9195: You can recognize words defined by a @code{CREATE}...@code{DOES>} word
9196: with @code{>does-code}. If the word was defined in that way, the value
9197: returned is non-zero and identifies the @code{DOES>} used by the
9198: defining word.
9199: @comment TODO should that be ``identifies the xt of the DOES> ??''
9200:
9201: @c -------------------------------------------------------------
9202: @node Locals, Structures, Threading Words, Words
9203: @section Locals
9204: @cindex locals
9205:
9206: Local variables can make Forth programming more enjoyable and Forth
9207: programs easier to read. Unfortunately, the locals of ANS Forth are
9208: laden with restrictions. Therefore, we provide not only the ANS Forth
9209: locals wordset, but also our own, more powerful locals wordset (we
9210: implemented the ANS Forth locals wordset through our locals wordset).
9211:
9212: The ideas in this section have also been published in M. Anton Ertl,
9213: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
9214: Automatic Scoping of Local Variables}}, EuroForth '94.
9215:
9216: @menu
9217: * Gforth locals::
9218: * ANS Forth locals::
9219: @end menu
9220:
9221: @node Gforth locals, ANS Forth locals, Locals, Locals
9222: @subsection Gforth locals
9223: @cindex Gforth locals
9224: @cindex locals, Gforth style
9225:
9226: Locals can be defined with
9227:
9228: @example
9229: @{ local1 local2 ... -- comment @}
9230: @end example
9231: or
9232: @example
9233: @{ local1 local2 ... @}
9234: @end example
9235:
9236: E.g.,
9237: @example
9238: : max @{ n1 n2 -- n3 @}
9239: n1 n2 > if
9240: n1
9241: else
9242: n2
9243: endif ;
9244: @end example
9245:
9246: The similarity of locals definitions with stack comments is intended. A
9247: locals definition often replaces the stack comment of a word. The order
9248: of the locals corresponds to the order in a stack comment and everything
9249: after the @code{--} is really a comment.
9250:
9251: This similarity has one disadvantage: It is too easy to confuse locals
9252: declarations with stack comments, causing bugs and making them hard to
9253: find. However, this problem can be avoided by appropriate coding
9254: conventions: Do not use both notations in the same program. If you do,
9255: they should be distinguished using additional means, e.g. by position.
9256:
9257: @cindex types of locals
9258: @cindex locals types
9259: The name of the local may be preceded by a type specifier, e.g.,
9260: @code{F:} for a floating point value:
9261:
9262: @example
9263: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
9264: \ complex multiplication
9265: Ar Br f* Ai Bi f* f-
9266: Ar Bi f* Ai Br f* f+ ;
9267: @end example
9268:
9269: @cindex flavours of locals
9270: @cindex locals flavours
9271: @cindex value-flavoured locals
9272: @cindex variable-flavoured locals
9273: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9274: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9275: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9276: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9277: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9278: produces its address (which becomes invalid when the variable's scope is
9279: left). E.g., the standard word @code{emit} can be defined in terms of
9280: @code{type} like this:
9281:
9282: @example
9283: : emit @{ C^ char* -- @}
9284: char* 1 type ;
9285: @end example
9286:
9287: @cindex default type of locals
9288: @cindex locals, default type
9289: A local without type specifier is a @code{W:} local. Both flavours of
9290: locals are initialized with values from the data or FP stack.
9291:
9292: Currently there is no way to define locals with user-defined data
9293: structures, but we are working on it.
9294:
9295: Gforth allows defining locals everywhere in a colon definition. This
9296: poses the following questions:
9297:
9298: @menu
9299: * Where are locals visible by name?::
9300: * How long do locals live?::
9301: * Programming Style::
9302: * Implementation::
9303: @end menu
9304:
9305: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9306: @subsubsection Where are locals visible by name?
9307: @cindex locals visibility
9308: @cindex visibility of locals
9309: @cindex scope of locals
9310:
9311: Basically, the answer is that locals are visible where you would expect
9312: it in block-structured languages, and sometimes a little longer. If you
9313: want to restrict the scope of a local, enclose its definition in
9314: @code{SCOPE}...@code{ENDSCOPE}.
9315:
9316:
9317: doc-scope
9318: doc-endscope
9319:
9320:
9321: These words behave like control structure words, so you can use them
9322: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9323: arbitrary ways.
9324:
9325: If you want a more exact answer to the visibility question, here's the
9326: basic principle: A local is visible in all places that can only be
9327: reached through the definition of the local@footnote{In compiler
9328: construction terminology, all places dominated by the definition of the
9329: local.}. In other words, it is not visible in places that can be reached
9330: without going through the definition of the local. E.g., locals defined
9331: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9332: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9333: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
9334:
9335: The reasoning behind this solution is: We want to have the locals
9336: visible as long as it is meaningful. The user can always make the
9337: visibility shorter by using explicit scoping. In a place that can
9338: only be reached through the definition of a local, the meaning of a
9339: local name is clear. In other places it is not: How is the local
9340: initialized at the control flow path that does not contain the
9341: definition? Which local is meant, if the same name is defined twice in
9342: two independent control flow paths?
9343:
9344: This should be enough detail for nearly all users, so you can skip the
9345: rest of this section. If you really must know all the gory details and
9346: options, read on.
9347:
9348: In order to implement this rule, the compiler has to know which places
9349: are unreachable. It knows this automatically after @code{AHEAD},
9350: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9351: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9352: compiler that the control flow never reaches that place. If
9353: @code{UNREACHABLE} is not used where it could, the only consequence is
9354: that the visibility of some locals is more limited than the rule above
9355: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9356: lie to the compiler), buggy code will be produced.
9357:
9358:
9359: doc-unreachable
9360:
9361:
9362: Another problem with this rule is that at @code{BEGIN}, the compiler
9363: does not know which locals will be visible on the incoming
9364: back-edge. All problems discussed in the following are due to this
9365: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9366: loops as examples; the discussion also applies to @code{?DO} and other
9367: loops). Perhaps the most insidious example is:
9368: @example
9369: AHEAD
9370: BEGIN
9371: x
9372: [ 1 CS-ROLL ] THEN
9373: @{ x @}
9374: ...
9375: UNTIL
9376: @end example
9377:
9378: This should be legal according to the visibility rule. The use of
9379: @code{x} can only be reached through the definition; but that appears
9380: textually below the use.
9381:
9382: From this example it is clear that the visibility rules cannot be fully
9383: implemented without major headaches. Our implementation treats common
9384: cases as advertised and the exceptions are treated in a safe way: The
9385: compiler makes a reasonable guess about the locals visible after a
9386: @code{BEGIN}; if it is too pessimistic, the
9387: user will get a spurious error about the local not being defined; if the
9388: compiler is too optimistic, it will notice this later and issue a
9389: warning. In the case above the compiler would complain about @code{x}
9390: being undefined at its use. You can see from the obscure examples in
9391: this section that it takes quite unusual control structures to get the
9392: compiler into trouble, and even then it will often do fine.
9393:
9394: If the @code{BEGIN} is reachable from above, the most optimistic guess
9395: is that all locals visible before the @code{BEGIN} will also be
9396: visible after the @code{BEGIN}. This guess is valid for all loops that
9397: are entered only through the @code{BEGIN}, in particular, for normal
9398: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9399: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9400: compiler. When the branch to the @code{BEGIN} is finally generated by
9401: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9402: warns the user if it was too optimistic:
9403: @example
9404: IF
9405: @{ x @}
9406: BEGIN
9407: \ x ?
9408: [ 1 cs-roll ] THEN
9409: ...
9410: UNTIL
9411: @end example
9412:
9413: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9414: optimistically assumes that it lives until the @code{THEN}. It notices
9415: this difference when it compiles the @code{UNTIL} and issues a
9416: warning. The user can avoid the warning, and make sure that @code{x}
9417: is not used in the wrong area by using explicit scoping:
9418: @example
9419: IF
9420: SCOPE
9421: @{ x @}
9422: ENDSCOPE
9423: BEGIN
9424: [ 1 cs-roll ] THEN
9425: ...
9426: UNTIL
9427: @end example
9428:
9429: Since the guess is optimistic, there will be no spurious error messages
9430: about undefined locals.
9431:
9432: If the @code{BEGIN} is not reachable from above (e.g., after
9433: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9434: optimistic guess, as the locals visible after the @code{BEGIN} may be
9435: defined later. Therefore, the compiler assumes that no locals are
9436: visible after the @code{BEGIN}. However, the user can use
9437: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9438: visible at the BEGIN as at the point where the top control-flow stack
9439: item was created.
9440:
9441:
9442: doc-assume-live
9443:
9444:
9445: @noindent
9446: E.g.,
9447: @example
9448: @{ x @}
9449: AHEAD
9450: ASSUME-LIVE
9451: BEGIN
9452: x
9453: [ 1 CS-ROLL ] THEN
9454: ...
9455: UNTIL
9456: @end example
9457:
9458: Other cases where the locals are defined before the @code{BEGIN} can be
9459: handled by inserting an appropriate @code{CS-ROLL} before the
9460: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9461: behind the @code{ASSUME-LIVE}).
9462:
9463: Cases where locals are defined after the @code{BEGIN} (but should be
9464: visible immediately after the @code{BEGIN}) can only be handled by
9465: rearranging the loop. E.g., the ``most insidious'' example above can be
9466: arranged into:
9467: @example
9468: BEGIN
9469: @{ x @}
9470: ... 0=
9471: WHILE
9472: x
9473: REPEAT
9474: @end example
9475:
9476: @node How long do locals live?, Programming Style, Where are locals visible by name?, Gforth locals
9477: @subsubsection How long do locals live?
9478: @cindex locals lifetime
9479: @cindex lifetime of locals
9480:
9481: The right answer for the lifetime question would be: A local lives at
9482: least as long as it can be accessed. For a value-flavoured local this
9483: means: until the end of its visibility. However, a variable-flavoured
9484: local could be accessed through its address far beyond its visibility
9485: scope. Ultimately, this would mean that such locals would have to be
9486: garbage collected. Since this entails un-Forth-like implementation
9487: complexities, I adopted the same cowardly solution as some other
9488: languages (e.g., C): The local lives only as long as it is visible;
9489: afterwards its address is invalid (and programs that access it
9490: afterwards are erroneous).
9491:
9492: @node Programming Style, Implementation, How long do locals live?, Gforth locals
9493: @subsubsection Programming Style
9494: @cindex locals programming style
9495: @cindex programming style, locals
9496:
9497: The freedom to define locals anywhere has the potential to change
9498: programming styles dramatically. In particular, the need to use the
9499: return stack for intermediate storage vanishes. Moreover, all stack
9500: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9501: determined arguments) can be eliminated: If the stack items are in the
9502: wrong order, just write a locals definition for all of them; then
9503: write the items in the order you want.
9504:
9505: This seems a little far-fetched and eliminating stack manipulations is
9506: unlikely to become a conscious programming objective. Still, the number
9507: of stack manipulations will be reduced dramatically if local variables
9508: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
9509: a traditional implementation of @code{max}).
9510:
9511: This shows one potential benefit of locals: making Forth programs more
9512: readable. Of course, this benefit will only be realized if the
9513: programmers continue to honour the principle of factoring instead of
9514: using the added latitude to make the words longer.
9515:
9516: @cindex single-assignment style for locals
9517: Using @code{TO} can and should be avoided. Without @code{TO},
9518: every value-flavoured local has only a single assignment and many
9519: advantages of functional languages apply to Forth. I.e., programs are
9520: easier to analyse, to optimize and to read: It is clear from the
9521: definition what the local stands for, it does not turn into something
9522: different later.
9523:
9524: E.g., a definition using @code{TO} might look like this:
9525: @example
9526: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9527: u1 u2 min 0
9528: ?do
9529: addr1 c@@ addr2 c@@ -
9530: ?dup-if
9531: unloop exit
9532: then
9533: addr1 char+ TO addr1
9534: addr2 char+ TO addr2
9535: loop
9536: u1 u2 - ;
9537: @end example
9538: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9539: every loop iteration. @code{strcmp} is a typical example of the
9540: readability problems of using @code{TO}. When you start reading
9541: @code{strcmp}, you think that @code{addr1} refers to the start of the
9542: string. Only near the end of the loop you realize that it is something
9543: else.
9544:
9545: This can be avoided by defining two locals at the start of the loop that
9546: are initialized with the right value for the current iteration.
9547: @example
9548: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9549: addr1 addr2
9550: u1 u2 min 0
9551: ?do @{ s1 s2 @}
9552: s1 c@@ s2 c@@ -
9553: ?dup-if
9554: unloop exit
9555: then
9556: s1 char+ s2 char+
9557: loop
9558: 2drop
9559: u1 u2 - ;
9560: @end example
9561: Here it is clear from the start that @code{s1} has a different value
9562: in every loop iteration.
9563:
9564: @node Implementation, , Programming Style, Gforth locals
9565: @subsubsection Implementation
9566: @cindex locals implementation
9567: @cindex implementation of locals
9568:
9569: @cindex locals stack
9570: Gforth uses an extra locals stack. The most compelling reason for
9571: this is that the return stack is not float-aligned; using an extra stack
9572: also eliminates the problems and restrictions of using the return stack
9573: as locals stack. Like the other stacks, the locals stack grows toward
9574: lower addresses. A few primitives allow an efficient implementation:
9575:
9576:
9577: doc-@local#
9578: doc-f@local#
9579: doc-laddr#
9580: doc-lp+!#
9581: doc-lp!
9582: doc->l
9583: doc-f>l
9584:
9585:
9586: In addition to these primitives, some specializations of these
9587: primitives for commonly occurring inline arguments are provided for
9588: efficiency reasons, e.g., @code{@@local0} as specialization of
9589: @code{@@local#} for the inline argument 0. The following compiling words
9590: compile the right specialized version, or the general version, as
9591: appropriate:
9592:
9593:
9594: doc-compile-@local
9595: doc-compile-f@local
9596: doc-compile-lp+!
9597:
9598:
9599: Combinations of conditional branches and @code{lp+!#} like
9600: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9601: is taken) are provided for efficiency and correctness in loops.
9602:
9603: A special area in the dictionary space is reserved for keeping the
9604: local variable names. @code{@{} switches the dictionary pointer to this
9605: area and @code{@}} switches it back and generates the locals
9606: initializing code. @code{W:} etc.@ are normal defining words. This
9607: special area is cleared at the start of every colon definition.
9608:
9609: @cindex word list for defining locals
9610: A special feature of Gforth's dictionary is used to implement the
9611: definition of locals without type specifiers: every word list (aka
9612: vocabulary) has its own methods for searching
9613: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9614: with a special search method: When it is searched for a word, it
9615: actually creates that word using @code{W:}. @code{@{} changes the search
9616: order to first search the word list containing @code{@}}, @code{W:} etc.,
9617: and then the word list for defining locals without type specifiers.
9618:
9619: The lifetime rules support a stack discipline within a colon
9620: definition: The lifetime of a local is either nested with other locals
9621: lifetimes or it does not overlap them.
9622:
9623: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9624: pointer manipulation is generated. Between control structure words
9625: locals definitions can push locals onto the locals stack. @code{AGAIN}
9626: is the simplest of the other three control flow words. It has to
9627: restore the locals stack depth of the corresponding @code{BEGIN}
9628: before branching. The code looks like this:
9629: @format
9630: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9631: @code{branch} <begin>
9632: @end format
9633:
9634: @code{UNTIL} is a little more complicated: If it branches back, it
9635: must adjust the stack just like @code{AGAIN}. But if it falls through,
9636: the locals stack must not be changed. The compiler generates the
9637: following code:
9638: @format
9639: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9640: @end format
9641: The locals stack pointer is only adjusted if the branch is taken.
9642:
9643: @code{THEN} can produce somewhat inefficient code:
9644: @format
9645: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9646: <orig target>:
9647: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9648: @end format
9649: The second @code{lp+!#} adjusts the locals stack pointer from the
9650: level at the @i{orig} point to the level after the @code{THEN}. The
9651: first @code{lp+!#} adjusts the locals stack pointer from the current
9652: level to the level at the orig point, so the complete effect is an
9653: adjustment from the current level to the right level after the
9654: @code{THEN}.
9655:
9656: @cindex locals information on the control-flow stack
9657: @cindex control-flow stack items, locals information
9658: In a conventional Forth implementation a dest control-flow stack entry
9659: is just the target address and an orig entry is just the address to be
9660: patched. Our locals implementation adds a word list to every orig or dest
9661: item. It is the list of locals visible (or assumed visible) at the point
9662: described by the entry. Our implementation also adds a tag to identify
9663: the kind of entry, in particular to differentiate between live and dead
9664: (reachable and unreachable) orig entries.
9665:
9666: A few unusual operations have to be performed on locals word lists:
9667:
9668:
9669: doc-common-list
9670: doc-sub-list?
9671: doc-list-size
9672:
9673:
9674: Several features of our locals word list implementation make these
9675: operations easy to implement: The locals word lists are organised as
9676: linked lists; the tails of these lists are shared, if the lists
9677: contain some of the same locals; and the address of a name is greater
9678: than the address of the names behind it in the list.
9679:
9680: Another important implementation detail is the variable
9681: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9682: determine if they can be reached directly or only through the branch
9683: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9684: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9685: definition, by @code{BEGIN} and usually by @code{THEN}.
9686:
9687: Counted loops are similar to other loops in most respects, but
9688: @code{LEAVE} requires special attention: It performs basically the same
9689: service as @code{AHEAD}, but it does not create a control-flow stack
9690: entry. Therefore the information has to be stored elsewhere;
9691: traditionally, the information was stored in the target fields of the
9692: branches created by the @code{LEAVE}s, by organizing these fields into a
9693: linked list. Unfortunately, this clever trick does not provide enough
9694: space for storing our extended control flow information. Therefore, we
9695: introduce another stack, the leave stack. It contains the control-flow
9696: stack entries for all unresolved @code{LEAVE}s.
9697:
9698: Local names are kept until the end of the colon definition, even if
9699: they are no longer visible in any control-flow path. In a few cases
9700: this may lead to increased space needs for the locals name area, but
9701: usually less than reclaiming this space would cost in code size.
9702:
9703:
9704: @node ANS Forth locals, , Gforth locals, Locals
9705: @subsection ANS Forth locals
9706: @cindex locals, ANS Forth style
9707:
9708: The ANS Forth locals wordset does not define a syntax for locals, but
9709: words that make it possible to define various syntaxes. One of the
9710: possible syntaxes is a subset of the syntax we used in the Gforth locals
9711: wordset, i.e.:
9712:
9713: @example
9714: @{ local1 local2 ... -- comment @}
9715: @end example
9716: @noindent
9717: or
9718: @example
9719: @{ local1 local2 ... @}
9720: @end example
9721:
9722: The order of the locals corresponds to the order in a stack comment. The
9723: restrictions are:
9724:
9725: @itemize @bullet
9726: @item
9727: Locals can only be cell-sized values (no type specifiers are allowed).
9728: @item
9729: Locals can be defined only outside control structures.
9730: @item
9731: Locals can interfere with explicit usage of the return stack. For the
9732: exact (and long) rules, see the standard. If you don't use return stack
9733: accessing words in a definition using locals, you will be all right. The
9734: purpose of this rule is to make locals implementation on the return
9735: stack easier.
9736: @item
9737: The whole definition must be in one line.
9738: @end itemize
9739:
9740: Locals defined in this way behave like @code{VALUE}s
9741: (@pxref{Values}). I.e., they are initialized from the stack. Using their
9742: name produces their value. Their value can be changed using @code{TO}.
9743:
9744: Since this syntax is supported by Gforth directly, you need not do
9745: anything to use it. If you want to port a program using this syntax to
9746: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9747: syntax on the other system.
9748:
9749: Note that a syntax shown in the standard, section A.13 looks
9750: similar, but is quite different in having the order of locals
9751: reversed. Beware!
9752:
9753: The ANS Forth locals wordset itself consists of a word:
9754:
9755:
9756: doc-(local)
9757:
9758:
9759: The ANS Forth locals extension wordset defines a syntax using @code{locals|}, but it is so
9760: awful that we strongly recommend not to use it. We have implemented this
9761: syntax to make porting to Gforth easy, but do not document it here. The
9762: problem with this syntax is that the locals are defined in an order
9763: reversed with respect to the standard stack comment notation, making
9764: programs harder to read, and easier to misread and miswrite. The only
9765: merit of this syntax is that it is easy to implement using the ANS Forth
9766: locals wordset.
9767:
9768:
9769: @c ----------------------------------------------------------
9770: @node Structures, Object-oriented Forth, Locals, Words
9771: @section Structures
9772: @cindex structures
9773: @cindex records
9774:
9775: This section presents the structure package that comes with Gforth. A
9776: version of the package implemented in ANS Forth is available in
9777: @file{compat/struct.fs}. This package was inspired by a posting on
9778: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9779: possibly John Hayes). A version of this section has been published in
9780: ???. Marcel Hendrix provided helpful comments.
9781:
9782: @menu
9783: * Why explicit structure support?::
9784: * Structure Usage::
9785: * Structure Naming Convention::
9786: * Structure Implementation::
9787: * Structure Glossary::
9788: @end menu
9789:
9790: @node Why explicit structure support?, Structure Usage, Structures, Structures
9791: @subsection Why explicit structure support?
9792:
9793: @cindex address arithmetic for structures
9794: @cindex structures using address arithmetic
9795: If we want to use a structure containing several fields, we could simply
9796: reserve memory for it, and access the fields using address arithmetic
9797: (@pxref{Address arithmetic}). As an example, consider a structure with
9798: the following fields
9799:
9800: @table @code
9801: @item a
9802: is a float
9803: @item b
9804: is a cell
9805: @item c
9806: is a float
9807: @end table
9808:
9809: Given the (float-aligned) base address of the structure we get the
9810: address of the field
9811:
9812: @table @code
9813: @item a
9814: without doing anything further.
9815: @item b
9816: with @code{float+}
9817: @item c
9818: with @code{float+ cell+ faligned}
9819: @end table
9820:
9821: It is easy to see that this can become quite tiring.
9822:
9823: Moreover, it is not very readable, because seeing a
9824: @code{cell+} tells us neither which kind of structure is
9825: accessed nor what field is accessed; we have to somehow infer the kind
9826: of structure, and then look up in the documentation, which field of
9827: that structure corresponds to that offset.
9828:
9829: Finally, this kind of address arithmetic also causes maintenance
9830: troubles: If you add or delete a field somewhere in the middle of the
9831: structure, you have to find and change all computations for the fields
9832: afterwards.
9833:
9834: So, instead of using @code{cell+} and friends directly, how
9835: about storing the offsets in constants:
9836:
9837: @example
9838: 0 constant a-offset
9839: 0 float+ constant b-offset
9840: 0 float+ cell+ faligned c-offset
9841: @end example
9842:
9843: Now we can get the address of field @code{x} with @code{x-offset
9844: +}. This is much better in all respects. Of course, you still
9845: have to change all later offset definitions if you add a field. You can
9846: fix this by declaring the offsets in the following way:
9847:
9848: @example
9849: 0 constant a-offset
9850: a-offset float+ constant b-offset
9851: b-offset cell+ faligned constant c-offset
9852: @end example
9853:
9854: Since we always use the offsets with @code{+}, we could use a defining
9855: word @code{cfield} that includes the @code{+} in the action of the
9856: defined word:
9857:
9858: @example
9859: : cfield ( n "name" -- )
9860: create ,
9861: does> ( name execution: addr1 -- addr2 )
9862: @@ + ;
9863:
9864: 0 cfield a
9865: 0 a float+ cfield b
9866: 0 b cell+ faligned cfield c
9867: @end example
9868:
9869: Instead of @code{x-offset +}, we now simply write @code{x}.
9870:
9871: The structure field words now can be used quite nicely. However,
9872: their definition is still a bit cumbersome: We have to repeat the
9873: name, the information about size and alignment is distributed before
9874: and after the field definitions etc. The structure package presented
9875: here addresses these problems.
9876:
9877: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9878: @subsection Structure Usage
9879: @cindex structure usage
9880:
9881: @cindex @code{field} usage
9882: @cindex @code{struct} usage
9883: @cindex @code{end-struct} usage
9884: You can define a structure for a (data-less) linked list with:
9885: @example
9886: struct
9887: cell% field list-next
9888: end-struct list%
9889: @end example
9890:
9891: With the address of the list node on the stack, you can compute the
9892: address of the field that contains the address of the next node with
9893: @code{list-next}. E.g., you can determine the length of a list
9894: with:
9895:
9896: @example
9897: : list-length ( list -- n )
9898: \ "list" is a pointer to the first element of a linked list
9899: \ "n" is the length of the list
9900: 0 BEGIN ( list1 n1 )
9901: over
9902: WHILE ( list1 n1 )
9903: 1+ swap list-next @@ swap
9904: REPEAT
9905: nip ;
9906: @end example
9907:
9908: You can reserve memory for a list node in the dictionary with
9909: @code{list% %allot}, which leaves the address of the list node on the
9910: stack. For the equivalent allocation on the heap you can use @code{list%
9911: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9912: use @code{list% %allocate}). You can get the the size of a list
9913: node with @code{list% %size} and its alignment with @code{list%
9914: %alignment}.
9915:
9916: Note that in ANS Forth the body of a @code{create}d word is
9917: @code{aligned} but not necessarily @code{faligned};
9918: therefore, if you do a:
9919: @example
9920: create @emph{name} foo% %allot
9921: @end example
9922:
9923: @noindent
9924: then the memory alloted for @code{foo%} is
9925: guaranteed to start at the body of @code{@emph{name}} only if
9926: @code{foo%} contains only character, cell and double fields.
9927:
9928: @cindex structures containing structures
9929: You can include a structure @code{foo%} as a field of
9930: another structure, like this:
9931: @example
9932: struct
9933: ...
9934: foo% field ...
9935: ...
9936: end-struct ...
9937: @end example
9938:
9939: @cindex structure extension
9940: @cindex extended records
9941: Instead of starting with an empty structure, you can extend an
9942: existing structure. E.g., a plain linked list without data, as defined
9943: above, is hardly useful; You can extend it to a linked list of integers,
9944: like this:@footnote{This feature is also known as @emph{extended
9945: records}. It is the main innovation in the Oberon language; in other
9946: words, adding this feature to Modula-2 led Wirth to create a new
9947: language, write a new compiler etc. Adding this feature to Forth just
9948: required a few lines of code.}
9949:
9950: @example
9951: list%
9952: cell% field intlist-int
9953: end-struct intlist%
9954: @end example
9955:
9956: @code{intlist%} is a structure with two fields:
9957: @code{list-next} and @code{intlist-int}.
9958:
9959: @cindex structures containing arrays
9960: You can specify an array type containing @emph{n} elements of
9961: type @code{foo%} like this:
9962:
9963: @example
9964: foo% @emph{n} *
9965: @end example
9966:
9967: You can use this array type in any place where you can use a normal
9968: type, e.g., when defining a @code{field}, or with
9969: @code{%allot}.
9970:
9971: @cindex first field optimization
9972: The first field is at the base address of a structure and the word
9973: for this field (e.g., @code{list-next}) actually does not change
9974: the address on the stack. You may be tempted to leave it away in the
9975: interest of run-time and space efficiency. This is not necessary,
9976: because the structure package optimizes this case and compiling such
9977: words does not generate any code. So, in the interest of readability
9978: and maintainability you should include the word for the field when
9979: accessing the field.
9980:
9981: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
9982: @subsection Structure Naming Convention
9983: @cindex structure naming convention
9984:
9985: The field names that come to (my) mind are often quite generic, and,
9986: if used, would cause frequent name clashes. E.g., many structures
9987: probably contain a @code{counter} field. The structure names
9988: that come to (my) mind are often also the logical choice for the names
9989: of words that create such a structure.
9990:
9991: Therefore, I have adopted the following naming conventions:
9992:
9993: @itemize @bullet
9994: @cindex field naming convention
9995: @item
9996: The names of fields are of the form
9997: @code{@emph{struct}-@emph{field}}, where
9998: @code{@emph{struct}} is the basic name of the structure, and
9999: @code{@emph{field}} is the basic name of the field. You can
10000: think of field words as converting the (address of the)
10001: structure into the (address of the) field.
10002:
10003: @cindex structure naming convention
10004: @item
10005: The names of structures are of the form
10006: @code{@emph{struct}%}, where
10007: @code{@emph{struct}} is the basic name of the structure.
10008: @end itemize
10009:
10010: This naming convention does not work that well for fields of extended
10011: structures; e.g., the integer list structure has a field
10012: @code{intlist-int}, but has @code{list-next}, not
10013: @code{intlist-next}.
10014:
10015: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
10016: @subsection Structure Implementation
10017: @cindex structure implementation
10018: @cindex implementation of structures
10019:
10020: The central idea in the implementation is to pass the data about the
10021: structure being built on the stack, not in some global
10022: variable. Everything else falls into place naturally once this design
10023: decision is made.
10024:
10025: The type description on the stack is of the form @emph{align
10026: size}. Keeping the size on the top-of-stack makes dealing with arrays
10027: very simple.
10028:
10029: @code{field} is a defining word that uses @code{Create}
10030: and @code{DOES>}. The body of the field contains the offset
10031: of the field, and the normal @code{DOES>} action is simply:
10032:
10033: @example
10034: @@ +
10035: @end example
10036:
10037: @noindent
10038: i.e., add the offset to the address, giving the stack effect
10039: @i{addr1 -- addr2} for a field.
10040:
10041: @cindex first field optimization, implementation
10042: This simple structure is slightly complicated by the optimization
10043: for fields with offset 0, which requires a different
10044: @code{DOES>}-part (because we cannot rely on there being
10045: something on the stack if such a field is invoked during
10046: compilation). Therefore, we put the different @code{DOES>}-parts
10047: in separate words, and decide which one to invoke based on the
10048: offset. For a zero offset, the field is basically a noop; it is
10049: immediate, and therefore no code is generated when it is compiled.
10050:
10051: @node Structure Glossary, , Structure Implementation, Structures
10052: @subsection Structure Glossary
10053: @cindex structure glossary
10054:
10055:
10056: doc-%align
10057: doc-%alignment
10058: doc-%alloc
10059: doc-%allocate
10060: doc-%allot
10061: doc-cell%
10062: doc-char%
10063: doc-dfloat%
10064: doc-double%
10065: doc-end-struct
10066: doc-field
10067: doc-float%
10068: doc-naligned
10069: doc-sfloat%
10070: doc-%size
10071: doc-struct
10072:
10073:
10074: @c -------------------------------------------------------------
10075: @node Object-oriented Forth, Passing Commands to the OS, Structures, Words
10076: @section Object-oriented Forth
10077:
10078: Gforth comes with three packages for object-oriented programming:
10079: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
10080: is preloaded, so you have to @code{include} them before use. The most
10081: important differences between these packages (and others) are discussed
10082: in @ref{Comparison with other object models}. All packages are written
10083: in ANS Forth and can be used with any other ANS Forth.
10084:
10085: @menu
10086: * Why object-oriented programming?::
10087: * Object-Oriented Terminology::
10088: * Objects::
10089: * OOF::
10090: * Mini-OOF::
10091: * Comparison with other object models::
10092: @end menu
10093:
10094: @c ----------------------------------------------------------------
10095: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
10096: @subsection Why object-oriented programming?
10097: @cindex object-oriented programming motivation
10098: @cindex motivation for object-oriented programming
10099:
10100: Often we have to deal with several data structures (@emph{objects}),
10101: that have to be treated similarly in some respects, but differently in
10102: others. Graphical objects are the textbook example: circles, triangles,
10103: dinosaurs, icons, and others, and we may want to add more during program
10104: development. We want to apply some operations to any graphical object,
10105: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
10106: has to do something different for every kind of object.
10107: @comment TODO add some other operations eg perimeter, area
10108: @comment and tie in to concrete examples later..
10109:
10110: We could implement @code{draw} as a big @code{CASE}
10111: control structure that executes the appropriate code depending on the
10112: kind of object to be drawn. This would be not be very elegant, and,
10113: moreover, we would have to change @code{draw} every time we add
10114: a new kind of graphical object (say, a spaceship).
10115:
10116: What we would rather do is: When defining spaceships, we would tell
10117: the system: ``Here's how you @code{draw} a spaceship; you figure
10118: out the rest''.
10119:
10120: This is the problem that all systems solve that (rightfully) call
10121: themselves object-oriented; the object-oriented packages presented here
10122: solve this problem (and not much else).
10123: @comment TODO ?list properties of oo systems.. oo vs o-based?
10124:
10125: @c ------------------------------------------------------------------------
10126: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
10127: @subsection Object-Oriented Terminology
10128: @cindex object-oriented terminology
10129: @cindex terminology for object-oriented programming
10130:
10131: This section is mainly for reference, so you don't have to understand
10132: all of it right away. The terminology is mainly Smalltalk-inspired. In
10133: short:
10134:
10135: @table @emph
10136: @cindex class
10137: @item class
10138: a data structure definition with some extras.
10139:
10140: @cindex object
10141: @item object
10142: an instance of the data structure described by the class definition.
10143:
10144: @cindex instance variables
10145: @item instance variables
10146: fields of the data structure.
10147:
10148: @cindex selector
10149: @cindex method selector
10150: @cindex virtual function
10151: @item selector
10152: (or @emph{method selector}) a word (e.g.,
10153: @code{draw}) that performs an operation on a variety of data
10154: structures (classes). A selector describes @emph{what} operation to
10155: perform. In C++ terminology: a (pure) virtual function.
10156:
10157: @cindex method
10158: @item method
10159: the concrete definition that performs the operation
10160: described by the selector for a specific class. A method specifies
10161: @emph{how} the operation is performed for a specific class.
10162:
10163: @cindex selector invocation
10164: @cindex message send
10165: @cindex invoking a selector
10166: @item selector invocation
10167: a call of a selector. One argument of the call (the TOS (top-of-stack))
10168: is used for determining which method is used. In Smalltalk terminology:
10169: a message (consisting of the selector and the other arguments) is sent
10170: to the object.
10171:
10172: @cindex receiving object
10173: @item receiving object
10174: the object used for determining the method executed by a selector
10175: invocation. In the @file{objects.fs} model, it is the object that is on
10176: the TOS when the selector is invoked. (@emph{Receiving} comes from
10177: the Smalltalk @emph{message} terminology.)
10178:
10179: @cindex child class
10180: @cindex parent class
10181: @cindex inheritance
10182: @item child class
10183: a class that has (@emph{inherits}) all properties (instance variables,
10184: selectors, methods) from a @emph{parent class}. In Smalltalk
10185: terminology: The subclass inherits from the superclass. In C++
10186: terminology: The derived class inherits from the base class.
10187:
10188: @end table
10189:
10190: @c If you wonder about the message sending terminology, it comes from
10191: @c a time when each object had it's own task and objects communicated via
10192: @c message passing; eventually the Smalltalk developers realized that
10193: @c they can do most things through simple (indirect) calls. They kept the
10194: @c terminology.
10195:
10196: @c --------------------------------------------------------------
10197: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
10198: @subsection The @file{objects.fs} model
10199: @cindex objects
10200: @cindex object-oriented programming
10201:
10202: @cindex @file{objects.fs}
10203: @cindex @file{oof.fs}
10204:
10205: This section describes the @file{objects.fs} package. This material also
10206: has been published in M. Anton Ertl,
10207: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
10208: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
10209: 37--43.
10210: @c McKewan's and Zsoter's packages
10211:
10212: This section assumes that you have read @ref{Structures}.
10213:
10214: The techniques on which this model is based have been used to implement
10215: the parser generator, Gray, and have also been used in Gforth for
10216: implementing the various flavours of word lists (hashed or not,
10217: case-sensitive or not, special-purpose word lists for locals etc.).
10218:
10219:
10220: @menu
10221: * Properties of the Objects model::
10222: * Basic Objects Usage::
10223: * The Objects base class::
10224: * Creating objects::
10225: * Object-Oriented Programming Style::
10226: * Class Binding::
10227: * Method conveniences::
10228: * Classes and Scoping::
10229: * Dividing classes::
10230: * Object Interfaces::
10231: * Objects Implementation::
10232: * Objects Glossary::
10233: @end menu
10234:
10235: Marcel Hendrix provided helpful comments on this section. Andras Zsoter
10236: and Bernd Paysan helped me with the related works section.
10237:
10238: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
10239: @subsubsection Properties of the @file{objects.fs} model
10240: @cindex @file{objects.fs} properties
10241:
10242: @itemize @bullet
10243: @item
10244: It is straightforward to pass objects on the stack. Passing
10245: selectors on the stack is a little less convenient, but possible.
10246:
10247: @item
10248: Objects are just data structures in memory, and are referenced by their
10249: address. You can create words for objects with normal defining words
10250: like @code{constant}. Likewise, there is no difference between instance
10251: variables that contain objects and those that contain other data.
10252:
10253: @item
10254: Late binding is efficient and easy to use.
10255:
10256: @item
10257: It avoids parsing, and thus avoids problems with state-smartness
10258: and reduced extensibility; for convenience there are a few parsing
10259: words, but they have non-parsing counterparts. There are also a few
10260: defining words that parse. This is hard to avoid, because all standard
10261: defining words parse (except @code{:noname}); however, such
10262: words are not as bad as many other parsing words, because they are not
10263: state-smart.
10264:
10265: @item
10266: It does not try to incorporate everything. It does a few things and does
10267: them well (IMO). In particular, this model was not designed to support
10268: information hiding (although it has features that may help); you can use
10269: a separate package for achieving this.
10270:
10271: @item
10272: It is layered; you don't have to learn and use all features to use this
10273: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10274: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
10275: are optional and independent of each other.
10276:
10277: @item
10278: An implementation in ANS Forth is available.
10279:
10280: @end itemize
10281:
10282:
10283: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10284: @subsubsection Basic @file{objects.fs} Usage
10285: @cindex basic objects usage
10286: @cindex objects, basic usage
10287:
10288: You can define a class for graphical objects like this:
10289:
10290: @cindex @code{class} usage
10291: @cindex @code{end-class} usage
10292: @cindex @code{selector} usage
10293: @example
10294: object class \ "object" is the parent class
10295: selector draw ( x y graphical -- )
10296: end-class graphical
10297: @end example
10298:
10299: This code defines a class @code{graphical} with an
10300: operation @code{draw}. We can perform the operation
10301: @code{draw} on any @code{graphical} object, e.g.:
10302:
10303: @example
10304: 100 100 t-rex draw
10305: @end example
10306:
10307: @noindent
10308: where @code{t-rex} is a word (say, a constant) that produces a
10309: graphical object.
10310:
10311: @comment TODO add a 2nd operation eg perimeter.. and use for
10312: @comment a concrete example
10313:
10314: @cindex abstract class
10315: How do we create a graphical object? With the present definitions,
10316: we cannot create a useful graphical object. The class
10317: @code{graphical} describes graphical objects in general, but not
10318: any concrete graphical object type (C++ users would call it an
10319: @emph{abstract class}); e.g., there is no method for the selector
10320: @code{draw} in the class @code{graphical}.
10321:
10322: For concrete graphical objects, we define child classes of the
10323: class @code{graphical}, e.g.:
10324:
10325: @cindex @code{overrides} usage
10326: @cindex @code{field} usage in class definition
10327: @example
10328: graphical class \ "graphical" is the parent class
10329: cell% field circle-radius
10330:
10331: :noname ( x y circle -- )
10332: circle-radius @@ draw-circle ;
10333: overrides draw
10334:
10335: :noname ( n-radius circle -- )
10336: circle-radius ! ;
10337: overrides construct
10338:
10339: end-class circle
10340: @end example
10341:
10342: Here we define a class @code{circle} as a child of @code{graphical},
10343: with field @code{circle-radius} (which behaves just like a field
10344: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10345: for the selectors @code{draw} and @code{construct} (@code{construct} is
10346: defined in @code{object}, the parent class of @code{graphical}).
10347:
10348: Now we can create a circle on the heap (i.e.,
10349: @code{allocate}d memory) with:
10350:
10351: @cindex @code{heap-new} usage
10352: @example
10353: 50 circle heap-new constant my-circle
10354: @end example
10355:
10356: @noindent
10357: @code{heap-new} invokes @code{construct}, thus
10358: initializing the field @code{circle-radius} with 50. We can draw
10359: this new circle at (100,100) with:
10360:
10361: @example
10362: 100 100 my-circle draw
10363: @end example
10364:
10365: @cindex selector invocation, restrictions
10366: @cindex class definition, restrictions
10367: Note: You can only invoke a selector if the object on the TOS
10368: (the receiving object) belongs to the class where the selector was
10369: defined or one of its descendents; e.g., you can invoke
10370: @code{draw} only for objects belonging to @code{graphical}
10371: or its descendents (e.g., @code{circle}). Immediately before
10372: @code{end-class}, the search order has to be the same as
10373: immediately after @code{class}.
10374:
10375: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10376: @subsubsection The @file{object.fs} base class
10377: @cindex @code{object} class
10378:
10379: When you define a class, you have to specify a parent class. So how do
10380: you start defining classes? There is one class available from the start:
10381: @code{object}. It is ancestor for all classes and so is the
10382: only class that has no parent. It has two selectors: @code{construct}
10383: and @code{print}.
10384:
10385: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10386: @subsubsection Creating objects
10387: @cindex creating objects
10388: @cindex object creation
10389: @cindex object allocation options
10390:
10391: @cindex @code{heap-new} discussion
10392: @cindex @code{dict-new} discussion
10393: @cindex @code{construct} discussion
10394: You can create and initialize an object of a class on the heap with
10395: @code{heap-new} ( ... class -- object ) and in the dictionary
10396: (allocation with @code{allot}) with @code{dict-new} (
10397: ... class -- object ). Both words invoke @code{construct}, which
10398: consumes the stack items indicated by "..." above.
10399:
10400: @cindex @code{init-object} discussion
10401: @cindex @code{class-inst-size} discussion
10402: If you want to allocate memory for an object yourself, you can get its
10403: alignment and size with @code{class-inst-size 2@@} ( class --
10404: align size ). Once you have memory for an object, you can initialize
10405: it with @code{init-object} ( ... class object -- );
10406: @code{construct} does only a part of the necessary work.
10407:
10408: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10409: @subsubsection Object-Oriented Programming Style
10410: @cindex object-oriented programming style
10411: @cindex programming style, object-oriented
10412:
10413: This section is not exhaustive.
10414:
10415: @cindex stack effects of selectors
10416: @cindex selectors and stack effects
10417: In general, it is a good idea to ensure that all methods for the
10418: same selector have the same stack effect: when you invoke a selector,
10419: you often have no idea which method will be invoked, so, unless all
10420: methods have the same stack effect, you will not know the stack effect
10421: of the selector invocation.
10422:
10423: One exception to this rule is methods for the selector
10424: @code{construct}. We know which method is invoked, because we
10425: specify the class to be constructed at the same place. Actually, I
10426: defined @code{construct} as a selector only to give the users a
10427: convenient way to specify initialization. The way it is used, a
10428: mechanism different from selector invocation would be more natural
10429: (but probably would take more code and more space to explain).
10430:
10431: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10432: @subsubsection Class Binding
10433: @cindex class binding
10434: @cindex early binding
10435:
10436: @cindex late binding
10437: Normal selector invocations determine the method at run-time depending
10438: on the class of the receiving object. This run-time selection is called
10439: @i{late binding}.
10440:
10441: Sometimes it's preferable to invoke a different method. For example,
10442: you might want to use the simple method for @code{print}ing
10443: @code{object}s instead of the possibly long-winded @code{print} method
10444: of the receiver class. You can achieve this by replacing the invocation
10445: of @code{print} with:
10446:
10447: @cindex @code{[bind]} usage
10448: @example
10449: [bind] object print
10450: @end example
10451:
10452: @noindent
10453: in compiled code or:
10454:
10455: @cindex @code{bind} usage
10456: @example
10457: bind object print
10458: @end example
10459:
10460: @cindex class binding, alternative to
10461: @noindent
10462: in interpreted code. Alternatively, you can define the method with a
10463: name (e.g., @code{print-object}), and then invoke it through the
10464: name. Class binding is just a (often more convenient) way to achieve
10465: the same effect; it avoids name clutter and allows you to invoke
10466: methods directly without naming them first.
10467:
10468: @cindex superclass binding
10469: @cindex parent class binding
10470: A frequent use of class binding is this: When we define a method
10471: for a selector, we often want the method to do what the selector does
10472: in the parent class, and a little more. There is a special word for
10473: this purpose: @code{[parent]}; @code{[parent]
10474: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10475: selector}}, where @code{@emph{parent}} is the parent
10476: class of the current class. E.g., a method definition might look like:
10477:
10478: @cindex @code{[parent]} usage
10479: @example
10480: :noname
10481: dup [parent] foo \ do parent's foo on the receiving object
10482: ... \ do some more
10483: ; overrides foo
10484: @end example
10485:
10486: @cindex class binding as optimization
10487: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10488: March 1997), Andrew McKewan presents class binding as an optimization
10489: technique. I recommend not using it for this purpose unless you are in
10490: an emergency. Late binding is pretty fast with this model anyway, so the
10491: benefit of using class binding is small; the cost of using class binding
10492: where it is not appropriate is reduced maintainability.
10493:
10494: While we are at programming style questions: You should bind
10495: selectors only to ancestor classes of the receiving object. E.g., say,
10496: you know that the receiving object is of class @code{foo} or its
10497: descendents; then you should bind only to @code{foo} and its
10498: ancestors.
10499:
10500: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10501: @subsubsection Method conveniences
10502: @cindex method conveniences
10503:
10504: In a method you usually access the receiving object pretty often. If
10505: you define the method as a plain colon definition (e.g., with
10506: @code{:noname}), you may have to do a lot of stack
10507: gymnastics. To avoid this, you can define the method with @code{m:
10508: ... ;m}. E.g., you could define the method for
10509: @code{draw}ing a @code{circle} with
10510:
10511: @cindex @code{this} usage
10512: @cindex @code{m:} usage
10513: @cindex @code{;m} usage
10514: @example
10515: m: ( x y circle -- )
10516: ( x y ) this circle-radius @@ draw-circle ;m
10517: @end example
10518:
10519: @cindex @code{exit} in @code{m: ... ;m}
10520: @cindex @code{exitm} discussion
10521: @cindex @code{catch} in @code{m: ... ;m}
10522: When this method is executed, the receiver object is removed from the
10523: stack; you can access it with @code{this} (admittedly, in this
10524: example the use of @code{m: ... ;m} offers no advantage). Note
10525: that I specify the stack effect for the whole method (i.e. including
10526: the receiver object), not just for the code between @code{m:}
10527: and @code{;m}. You cannot use @code{exit} in
10528: @code{m:...;m}; instead, use
10529: @code{exitm}.@footnote{Moreover, for any word that calls
10530: @code{catch} and was defined before loading
10531: @code{objects.fs}, you have to redefine it like I redefined
10532: @code{catch}: @code{: catch this >r catch r> to-this ;}}
10533:
10534: @cindex @code{inst-var} usage
10535: You will frequently use sequences of the form @code{this
10536: @emph{field}} (in the example above: @code{this
10537: circle-radius}). If you use the field only in this way, you can
10538: define it with @code{inst-var} and eliminate the
10539: @code{this} before the field name. E.g., the @code{circle}
10540: class above could also be defined with:
10541:
10542: @example
10543: graphical class
10544: cell% inst-var radius
10545:
10546: m: ( x y circle -- )
10547: radius @@ draw-circle ;m
10548: overrides draw
10549:
10550: m: ( n-radius circle -- )
10551: radius ! ;m
10552: overrides construct
10553:
10554: end-class circle
10555: @end example
10556:
10557: @code{radius} can only be used in @code{circle} and its
10558: descendent classes and inside @code{m:...;m}.
10559:
10560: @cindex @code{inst-value} usage
10561: You can also define fields with @code{inst-value}, which is
10562: to @code{inst-var} what @code{value} is to
10563: @code{variable}. You can change the value of such a field with
10564: @code{[to-inst]}. E.g., we could also define the class
10565: @code{circle} like this:
10566:
10567: @example
10568: graphical class
10569: inst-value radius
10570:
10571: m: ( x y circle -- )
10572: radius draw-circle ;m
10573: overrides draw
10574:
10575: m: ( n-radius circle -- )
10576: [to-inst] radius ;m
10577: overrides construct
10578:
10579: end-class circle
10580: @end example
10581:
10582: Finally, you can define named methods with @code{:m}. One use of this
10583: feature is the definition of words that occur only in one class and are
10584: not intended to be overridden, but which still need method context
10585: (e.g., for accessing @code{inst-var}s). Another use is for methods that
10586: would be bound frequently, if defined anonymously.
10587:
10588:
10589: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
10590: @subsubsection Classes and Scoping
10591: @cindex classes and scoping
10592: @cindex scoping and classes
10593:
10594: Inheritance is frequent, unlike structure extension. This exacerbates
10595: the problem with the field name convention (@pxref{Structure Naming
10596: Convention}): One always has to remember in which class the field was
10597: originally defined; changing a part of the class structure would require
10598: changes for renaming in otherwise unaffected code.
10599:
10600: @cindex @code{inst-var} visibility
10601: @cindex @code{inst-value} visibility
10602: To solve this problem, I added a scoping mechanism (which was not in my
10603: original charter): A field defined with @code{inst-var} (or
10604: @code{inst-value}) is visible only in the class where it is defined and in
10605: the descendent classes of this class. Using such fields only makes
10606: sense in @code{m:}-defined methods in these classes anyway.
10607:
10608: This scoping mechanism allows us to use the unadorned field name,
10609: because name clashes with unrelated words become much less likely.
10610:
10611: @cindex @code{protected} discussion
10612: @cindex @code{private} discussion
10613: Once we have this mechanism, we can also use it for controlling the
10614: visibility of other words: All words defined after
10615: @code{protected} are visible only in the current class and its
10616: descendents. @code{public} restores the compilation
10617: (i.e. @code{current}) word list that was in effect before. If you
10618: have several @code{protected}s without an intervening
10619: @code{public} or @code{set-current}, @code{public}
10620: will restore the compilation word list in effect before the first of
10621: these @code{protected}s.
10622:
10623: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10624: @subsubsection Dividing classes
10625: @cindex Dividing classes
10626: @cindex @code{methods}...@code{end-methods}
10627:
10628: You may want to do the definition of methods separate from the
10629: definition of the class, its selectors, fields, and instance variables,
10630: i.e., separate the implementation from the definition. You can do this
10631: in the following way:
10632:
10633: @example
10634: graphical class
10635: inst-value radius
10636: end-class circle
10637:
10638: ... \ do some other stuff
10639:
10640: circle methods \ now we are ready
10641:
10642: m: ( x y circle -- )
10643: radius draw-circle ;m
10644: overrides draw
10645:
10646: m: ( n-radius circle -- )
10647: [to-inst] radius ;m
10648: overrides construct
10649:
10650: end-methods
10651: @end example
10652:
10653: You can use several @code{methods}...@code{end-methods} sections. The
10654: only things you can do to the class in these sections are: defining
10655: methods, and overriding the class's selectors. You must not define new
10656: selectors or fields.
10657:
10658: Note that you often have to override a selector before using it. In
10659: particular, you usually have to override @code{construct} with a new
10660: method before you can invoke @code{heap-new} and friends. E.g., you
10661: must not create a circle before the @code{overrides construct} sequence
10662: in the example above.
10663:
10664: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
10665: @subsubsection Object Interfaces
10666: @cindex object interfaces
10667: @cindex interfaces for objects
10668:
10669: In this model you can only call selectors defined in the class of the
10670: receiving objects or in one of its ancestors. If you call a selector
10671: with a receiving object that is not in one of these classes, the
10672: result is undefined; if you are lucky, the program crashes
10673: immediately.
10674:
10675: @cindex selectors common to hardly-related classes
10676: Now consider the case when you want to have a selector (or several)
10677: available in two classes: You would have to add the selector to a
10678: common ancestor class, in the worst case to @code{object}. You
10679: may not want to do this, e.g., because someone else is responsible for
10680: this ancestor class.
10681:
10682: The solution for this problem is interfaces. An interface is a
10683: collection of selectors. If a class implements an interface, the
10684: selectors become available to the class and its descendents. A class
10685: can implement an unlimited number of interfaces. For the problem
10686: discussed above, we would define an interface for the selector(s), and
10687: both classes would implement the interface.
10688:
10689: As an example, consider an interface @code{storage} for
10690: writing objects to disk and getting them back, and a class
10691: @code{foo} that implements it. The code would look like this:
10692:
10693: @cindex @code{interface} usage
10694: @cindex @code{end-interface} usage
10695: @cindex @code{implementation} usage
10696: @example
10697: interface
10698: selector write ( file object -- )
10699: selector read1 ( file object -- )
10700: end-interface storage
10701:
10702: bar class
10703: storage implementation
10704:
10705: ... overrides write
10706: ... overrides read1
10707: ...
10708: end-class foo
10709: @end example
10710:
10711: @noindent
10712: (I would add a word @code{read} @i{( file -- object )} that uses
10713: @code{read1} internally, but that's beyond the point illustrated
10714: here.)
10715:
10716: Note that you cannot use @code{protected} in an interface; and
10717: of course you cannot define fields.
10718:
10719: In the Neon model, all selectors are available for all classes;
10720: therefore it does not need interfaces. The price you pay in this model
10721: is slower late binding, and therefore, added complexity to avoid late
10722: binding.
10723:
10724: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10725: @subsubsection @file{objects.fs} Implementation
10726: @cindex @file{objects.fs} implementation
10727:
10728: @cindex @code{object-map} discussion
10729: An object is a piece of memory, like one of the data structures
10730: described with @code{struct...end-struct}. It has a field
10731: @code{object-map} that points to the method map for the object's
10732: class.
10733:
10734: @cindex method map
10735: @cindex virtual function table
10736: The @emph{method map}@footnote{This is Self terminology; in C++
10737: terminology: virtual function table.} is an array that contains the
10738: execution tokens (@i{xt}s) of the methods for the object's class. Each
10739: selector contains an offset into a method map.
10740:
10741: @cindex @code{selector} implementation, class
10742: @code{selector} is a defining word that uses
10743: @code{CREATE} and @code{DOES>}. The body of the
10744: selector contains the offset; the @code{DOES>} action for a
10745: class selector is, basically:
10746:
10747: @example
10748: ( object addr ) @@ over object-map @@ + @@ execute
10749: @end example
10750:
10751: Since @code{object-map} is the first field of the object, it
10752: does not generate any code. As you can see, calling a selector has a
10753: small, constant cost.
10754:
10755: @cindex @code{current-interface} discussion
10756: @cindex class implementation and representation
10757: A class is basically a @code{struct} combined with a method
10758: map. During the class definition the alignment and size of the class
10759: are passed on the stack, just as with @code{struct}s, so
10760: @code{field} can also be used for defining class
10761: fields. However, passing more items on the stack would be
10762: inconvenient, so @code{class} builds a data structure in memory,
10763: which is accessed through the variable
10764: @code{current-interface}. After its definition is complete, the
10765: class is represented on the stack by a pointer (e.g., as parameter for
10766: a child class definition).
10767:
10768: A new class starts off with the alignment and size of its parent,
10769: and a copy of the parent's method map. Defining new fields extends the
10770: size and alignment; likewise, defining new selectors extends the
10771: method map. @code{overrides} just stores a new @i{xt} in the method
10772: map at the offset given by the selector.
10773:
10774: @cindex class binding, implementation
10775: Class binding just gets the @i{xt} at the offset given by the selector
10776: from the class's method map and @code{compile,}s (in the case of
10777: @code{[bind]}) it.
10778:
10779: @cindex @code{this} implementation
10780: @cindex @code{catch} and @code{this}
10781: @cindex @code{this} and @code{catch}
10782: I implemented @code{this} as a @code{value}. At the
10783: start of an @code{m:...;m} method the old @code{this} is
10784: stored to the return stack and restored at the end; and the object on
10785: the TOS is stored @code{TO this}. This technique has one
10786: disadvantage: If the user does not leave the method via
10787: @code{;m}, but via @code{throw} or @code{exit},
10788: @code{this} is not restored (and @code{exit} may
10789: crash). To deal with the @code{throw} problem, I have redefined
10790: @code{catch} to save and restore @code{this}; the same
10791: should be done with any word that can catch an exception. As for
10792: @code{exit}, I simply forbid it (as a replacement, there is
10793: @code{exitm}).
10794:
10795: @cindex @code{inst-var} implementation
10796: @code{inst-var} is just the same as @code{field}, with
10797: a different @code{DOES>} action:
10798: @example
10799: @@ this +
10800: @end example
10801: Similar for @code{inst-value}.
10802:
10803: @cindex class scoping implementation
10804: Each class also has a word list that contains the words defined with
10805: @code{inst-var} and @code{inst-value}, and its protected
10806: words. It also has a pointer to its parent. @code{class} pushes
10807: the word lists of the class and all its ancestors onto the search order stack,
10808: and @code{end-class} drops them.
10809:
10810: @cindex interface implementation
10811: An interface is like a class without fields, parent and protected
10812: words; i.e., it just has a method map. If a class implements an
10813: interface, its method map contains a pointer to the method map of the
10814: interface. The positive offsets in the map are reserved for class
10815: methods, therefore interface map pointers have negative
10816: offsets. Interfaces have offsets that are unique throughout the
10817: system, unlike class selectors, whose offsets are only unique for the
10818: classes where the selector is available (invokable).
10819:
10820: This structure means that interface selectors have to perform one
10821: indirection more than class selectors to find their method. Their body
10822: contains the interface map pointer offset in the class method map, and
10823: the method offset in the interface method map. The
10824: @code{does>} action for an interface selector is, basically:
10825:
10826: @example
10827: ( object selector-body )
10828: 2dup selector-interface @@ ( object selector-body object interface-offset )
10829: swap object-map @@ + @@ ( object selector-body map )
10830: swap selector-offset @@ + @@ execute
10831: @end example
10832:
10833: where @code{object-map} and @code{selector-offset} are
10834: first fields and generate no code.
10835:
10836: As a concrete example, consider the following code:
10837:
10838: @example
10839: interface
10840: selector if1sel1
10841: selector if1sel2
10842: end-interface if1
10843:
10844: object class
10845: if1 implementation
10846: selector cl1sel1
10847: cell% inst-var cl1iv1
10848:
10849: ' m1 overrides construct
10850: ' m2 overrides if1sel1
10851: ' m3 overrides if1sel2
10852: ' m4 overrides cl1sel2
10853: end-class cl1
10854:
10855: create obj1 object dict-new drop
10856: create obj2 cl1 dict-new drop
10857: @end example
10858:
10859: The data structure created by this code (including the data structure
10860: for @code{object}) is shown in the <a
10861: href="objects-implementation.eps">figure</a>, assuming a cell size of 4.
10862: @comment TODO add this diagram..
10863:
10864: @node Objects Glossary, , Objects Implementation, Objects
10865: @subsubsection @file{objects.fs} Glossary
10866: @cindex @file{objects.fs} Glossary
10867:
10868:
10869: doc---objects-bind
10870: doc---objects-<bind>
10871: doc---objects-bind'
10872: doc---objects-[bind]
10873: doc---objects-class
10874: doc---objects-class->map
10875: doc---objects-class-inst-size
10876: doc---objects-class-override!
10877: doc---objects-construct
10878: doc---objects-current'
10879: doc---objects-[current]
10880: doc---objects-current-interface
10881: doc---objects-dict-new
10882: doc---objects-drop-order
10883: doc---objects-end-class
10884: doc---objects-end-class-noname
10885: doc---objects-end-interface
10886: doc---objects-end-interface-noname
10887: doc---objects-end-methods
10888: doc---objects-exitm
10889: doc---objects-heap-new
10890: doc---objects-implementation
10891: doc---objects-init-object
10892: doc---objects-inst-value
10893: doc---objects-inst-var
10894: doc---objects-interface
10895: doc---objects-m:
10896: doc---objects-:m
10897: doc---objects-;m
10898: doc---objects-method
10899: doc---objects-methods
10900: doc---objects-object
10901: doc---objects-overrides
10902: doc---objects-[parent]
10903: doc---objects-print
10904: doc---objects-protected
10905: doc---objects-public
10906: doc---objects-push-order
10907: doc---objects-selector
10908: doc---objects-this
10909: doc---objects-<to-inst>
10910: doc---objects-[to-inst]
10911: doc---objects-to-this
10912: doc---objects-xt-new
10913:
10914:
10915: @c -------------------------------------------------------------
10916: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10917: @subsection The @file{oof.fs} model
10918: @cindex oof
10919: @cindex object-oriented programming
10920:
10921: @cindex @file{objects.fs}
10922: @cindex @file{oof.fs}
10923:
10924: This section describes the @file{oof.fs} package.
10925:
10926: The package described in this section has been used in bigFORTH since 1991, and
10927: used for two large applications: a chromatographic system used to
10928: create new medicaments, and a graphic user interface library (MINOS).
10929:
10930: You can find a description (in German) of @file{oof.fs} in @cite{Object
10931: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10932: 10(2), 1994.
10933:
10934: @menu
10935: * Properties of the OOF model::
10936: * Basic OOF Usage::
10937: * The OOF base class::
10938: * Class Declaration::
10939: * Class Implementation::
10940: @end menu
10941:
10942: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
10943: @subsubsection Properties of the @file{oof.fs} model
10944: @cindex @file{oof.fs} properties
10945:
10946: @itemize @bullet
10947: @item
10948: This model combines object oriented programming with information
10949: hiding. It helps you writing large application, where scoping is
10950: necessary, because it provides class-oriented scoping.
10951:
10952: @item
10953: Named objects, object pointers, and object arrays can be created,
10954: selector invocation uses the ``object selector'' syntax. Selector invocation
10955: to objects and/or selectors on the stack is a bit less convenient, but
10956: possible.
10957:
10958: @item
10959: Selector invocation and instance variable usage of the active object is
10960: straightforward, since both make use of the active object.
10961:
10962: @item
10963: Late binding is efficient and easy to use.
10964:
10965: @item
10966: State-smart objects parse selectors. However, extensibility is provided
10967: using a (parsing) selector @code{postpone} and a selector @code{'}.
10968:
10969: @item
10970: An implementation in ANS Forth is available.
10971:
10972: @end itemize
10973:
10974:
10975: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
10976: @subsubsection Basic @file{oof.fs} Usage
10977: @cindex @file{oof.fs} usage
10978:
10979: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
10980:
10981: You can define a class for graphical objects like this:
10982:
10983: @cindex @code{class} usage
10984: @cindex @code{class;} usage
10985: @cindex @code{method} usage
10986: @example
10987: object class graphical \ "object" is the parent class
10988: method draw ( x y graphical -- )
10989: class;
10990: @end example
10991:
10992: This code defines a class @code{graphical} with an
10993: operation @code{draw}. We can perform the operation
10994: @code{draw} on any @code{graphical} object, e.g.:
10995:
10996: @example
10997: 100 100 t-rex draw
10998: @end example
10999:
11000: @noindent
11001: where @code{t-rex} is an object or object pointer, created with e.g.
11002: @code{graphical : t-rex}.
11003:
11004: @cindex abstract class
11005: How do we create a graphical object? With the present definitions,
11006: we cannot create a useful graphical object. The class
11007: @code{graphical} describes graphical objects in general, but not
11008: any concrete graphical object type (C++ users would call it an
11009: @emph{abstract class}); e.g., there is no method for the selector
11010: @code{draw} in the class @code{graphical}.
11011:
11012: For concrete graphical objects, we define child classes of the
11013: class @code{graphical}, e.g.:
11014:
11015: @example
11016: graphical class circle \ "graphical" is the parent class
11017: cell var circle-radius
11018: how:
11019: : draw ( x y -- )
11020: circle-radius @@ draw-circle ;
11021:
11022: : init ( n-radius -- (
11023: circle-radius ! ;
11024: class;
11025: @end example
11026:
11027: Here we define a class @code{circle} as a child of @code{graphical},
11028: with a field @code{circle-radius}; it defines new methods for the
11029: selectors @code{draw} and @code{init} (@code{init} is defined in
11030: @code{object}, the parent class of @code{graphical}).
11031:
11032: Now we can create a circle in the dictionary with:
11033:
11034: @example
11035: 50 circle : my-circle
11036: @end example
11037:
11038: @noindent
11039: @code{:} invokes @code{init}, thus initializing the field
11040: @code{circle-radius} with 50. We can draw this new circle at (100,100)
11041: with:
11042:
11043: @example
11044: 100 100 my-circle draw
11045: @end example
11046:
11047: @cindex selector invocation, restrictions
11048: @cindex class definition, restrictions
11049: Note: You can only invoke a selector if the receiving object belongs to
11050: the class where the selector was defined or one of its descendents;
11051: e.g., you can invoke @code{draw} only for objects belonging to
11052: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
11053: mechanism will check if you try to invoke a selector that is not
11054: defined in this class hierarchy, so you'll get an error at compilation
11055: time.
11056:
11057:
11058: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
11059: @subsubsection The @file{oof.fs} base class
11060: @cindex @file{oof.fs} base class
11061:
11062: When you define a class, you have to specify a parent class. So how do
11063: you start defining classes? There is one class available from the start:
11064: @code{object}. You have to use it as ancestor for all classes. It is the
11065: only class that has no parent. Classes are also objects, except that
11066: they don't have instance variables; class manipulation such as
11067: inheritance or changing definitions of a class is handled through
11068: selectors of the class @code{object}.
11069:
11070: @code{object} provides a number of selectors:
11071:
11072: @itemize @bullet
11073: @item
11074: @code{class} for subclassing, @code{definitions} to add definitions
11075: later on, and @code{class?} to get type informations (is the class a
11076: subclass of the class passed on the stack?).
11077:
11078: doc---object-class
11079: doc---object-definitions
11080: doc---object-class?
11081:
11082:
11083: @item
11084: @code{init} and @code{dispose} as constructor and destructor of the
11085: object. @code{init} is invocated after the object's memory is allocated,
11086: while @code{dispose} also handles deallocation. Thus if you redefine
11087: @code{dispose}, you have to call the parent's dispose with @code{super
11088: dispose}, too.
11089:
11090: doc---object-init
11091: doc---object-dispose
11092:
11093:
11094: @item
11095: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
11096: @code{[]} to create named and unnamed objects and object arrays or
11097: object pointers.
11098:
11099: doc---object-new
11100: doc---object-new[]
11101: doc---object-:
11102: doc---object-ptr
11103: doc---object-asptr
11104: doc---object-[]
11105:
11106:
11107: @item
11108: @code{::} and @code{super} for explicit scoping. You should use explicit
11109: scoping only for super classes or classes with the same set of instance
11110: variables. Explicitly-scoped selectors use early binding.
11111:
11112: doc---object-::
11113: doc---object-super
11114:
11115:
11116: @item
11117: @code{self} to get the address of the object
11118:
11119: doc---object-self
11120:
11121:
11122: @item
11123: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
11124: pointers and instance defers.
11125:
11126: doc---object-bind
11127: doc---object-bound
11128: doc---object-link
11129: doc---object-is
11130:
11131:
11132: @item
11133: @code{'} to obtain selector tokens, @code{send} to invocate selectors
11134: form the stack, and @code{postpone} to generate selector invocation code.
11135:
11136: doc---object-'
11137: doc---object-postpone
11138:
11139:
11140: @item
11141: @code{with} and @code{endwith} to select the active object from the
11142: stack, and enable its scope. Using @code{with} and @code{endwith}
11143: also allows you to create code using selector @code{postpone} without being
11144: trapped by the state-smart objects.
11145:
11146: doc---object-with
11147: doc---object-endwith
11148:
11149:
11150: @end itemize
11151:
11152: @node Class Declaration, Class Implementation, The OOF base class, OOF
11153: @subsubsection Class Declaration
11154: @cindex class declaration
11155:
11156: @itemize @bullet
11157: @item
11158: Instance variables
11159:
11160: doc---oof-var
11161:
11162:
11163: @item
11164: Object pointers
11165:
11166: doc---oof-ptr
11167: doc---oof-asptr
11168:
11169:
11170: @item
11171: Instance defers
11172:
11173: doc---oof-defer
11174:
11175:
11176: @item
11177: Method selectors
11178:
11179: doc---oof-early
11180: doc---oof-method
11181:
11182:
11183: @item
11184: Class-wide variables
11185:
11186: doc---oof-static
11187:
11188:
11189: @item
11190: End declaration
11191:
11192: doc---oof-how:
11193: doc---oof-class;
11194:
11195:
11196: @end itemize
11197:
11198: @c -------------------------------------------------------------
11199: @node Class Implementation, , Class Declaration, OOF
11200: @subsubsection Class Implementation
11201: @cindex class implementation
11202:
11203: @c -------------------------------------------------------------
11204: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
11205: @subsection The @file{mini-oof.fs} model
11206: @cindex mini-oof
11207:
11208: Gforth's third object oriented Forth package is a 12-liner. It uses a
11209: mixture of the @file{object.fs} and the @file{oof.fs} syntax,
11210: and reduces to the bare minimum of features. This is based on a posting
11211: of Bernd Paysan in comp.arch.
11212:
11213: @menu
11214: * Basic Mini-OOF Usage::
11215: * Mini-OOF Example::
11216: * Mini-OOF Implementation::
11217: * Comparison with other object models::
11218: @end menu
11219:
11220: @c -------------------------------------------------------------
11221: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
11222: @subsubsection Basic @file{mini-oof.fs} Usage
11223: @cindex mini-oof usage
11224:
11225: There is a base class (@code{class}, which allocates one cell for the
11226: object pointer) plus seven other words: to define a method, a variable,
11227: a class; to end a class, to resolve binding, to allocate an object and
11228: to compile a class method.
11229: @comment TODO better description of the last one
11230:
11231:
11232: doc-object
11233: doc-method
11234: doc-var
11235: doc-class
11236: doc-end-class
11237: doc-defines
11238: doc-new
11239: doc-::
11240:
11241:
11242:
11243: @c -------------------------------------------------------------
11244: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
11245: @subsubsection Mini-OOF Example
11246: @cindex mini-oof example
11247:
11248: A short example shows how to use this package. This example, in slightly
11249: extended form, is supplied as @file{moof-exm.fs}
11250: @comment TODO could flesh this out with some comments from the Forthwrite article
11251:
11252: @example
11253: object class
11254: method init
11255: method draw
11256: end-class graphical
11257: @end example
11258:
11259: This code defines a class @code{graphical} with an
11260: operation @code{draw}. We can perform the operation
11261: @code{draw} on any @code{graphical} object, e.g.:
11262:
11263: @example
11264: 100 100 t-rex draw
11265: @end example
11266:
11267: where @code{t-rex} is an object or object pointer, created with e.g.
11268: @code{graphical new Constant t-rex}.
11269:
11270: For concrete graphical objects, we define child classes of the
11271: class @code{graphical}, e.g.:
11272:
11273: @example
11274: graphical class
11275: cell var circle-radius
11276: end-class circle \ "graphical" is the parent class
11277:
11278: :noname ( x y -- )
11279: circle-radius @@ draw-circle ; circle defines draw
11280: :noname ( r -- )
11281: circle-radius ! ; circle defines init
11282: @end example
11283:
11284: There is no implicit init method, so we have to define one. The creation
11285: code of the object now has to call init explicitely.
11286:
11287: @example
11288: circle new Constant my-circle
11289: 50 my-circle init
11290: @end example
11291:
11292: It is also possible to add a function to create named objects with
11293: automatic call of @code{init}, given that all objects have @code{init}
11294: on the same place:
11295:
11296: @example
11297: : new: ( .. o "name" -- )
11298: new dup Constant init ;
11299: 80 circle new: large-circle
11300: @end example
11301:
11302: We can draw this new circle at (100,100) with:
11303:
11304: @example
11305: 100 100 my-circle draw
11306: @end example
11307:
11308: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
11309: @subsubsection @file{mini-oof.fs} Implementation
11310:
11311: Object-oriented systems with late binding typically use a
11312: ``vtable''-approach: the first variable in each object is a pointer to a
11313: table, which contains the methods as function pointers. The vtable
11314: may also contain other information.
11315:
11316: So first, let's declare methods:
11317:
11318: @example
11319: : method ( m v -- m' v ) Create over , swap cell+ swap
11320: DOES> ( ... o -- ... ) @ over @ + @ execute ;
11321: @end example
11322:
11323: During method declaration, the number of methods and instance
11324: variables is on the stack (in address units). @code{method} creates
11325: one method and increments the method number. To execute a method, it
11326: takes the object, fetches the vtable pointer, adds the offset, and
11327: executes the @i{xt} stored there. Each method takes the object it is
11328: invoked from as top of stack parameter. The method itself should
11329: consume that object.
11330:
11331: Now, we also have to declare instance variables
11332:
11333: @example
11334: : var ( m v size -- m v' ) Create over , +
11335: DOES> ( o -- addr ) @ + ;
11336: @end example
11337:
11338: As before, a word is created with the current offset. Instance
11339: variables can have different sizes (cells, floats, doubles, chars), so
11340: all we do is take the size and add it to the offset. If your machine
11341: has alignment restrictions, put the proper @code{aligned} or
11342: @code{faligned} before the variable, to adjust the variable
11343: offset. That's why it is on the top of stack.
11344:
11345: We need a starting point (the base object) and some syntactic sugar:
11346:
11347: @example
11348: Create object 1 cells , 2 cells ,
11349: : class ( class -- class methods vars ) dup 2@ ;
11350: @end example
11351:
11352: For inheritance, the vtable of the parent object has to be
11353: copied when a new, derived class is declared. This gives all the
11354: methods of the parent class, which can be overridden, though.
11355:
11356: @example
11357: : end-class ( class methods vars -- )
11358: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
11359: cell+ dup cell+ r> rot @ 2 cells /string move ;
11360: @end example
11361:
11362: The first line creates the vtable, initialized with
11363: @code{noop}s. The second line is the inheritance mechanism, it
11364: copies the xts from the parent vtable.
11365:
11366: We still have no way to define new methods, let's do that now:
11367:
11368: @example
11369: : defines ( xt class -- ) ' >body @ + ! ;
11370: @end example
11371:
11372: To allocate a new object, we need a word, too:
11373:
11374: @example
11375: : new ( class -- o ) here over @ allot swap over ! ;
11376: @end example
11377:
11378: Sometimes derived classes want to access the method of the
11379: parent object. There are two ways to achieve this with Mini-OOF:
11380: first, you could use named words, and second, you could look up the
11381: vtable of the parent object.
11382:
11383: @example
11384: : :: ( class "name" -- ) ' >body @ + @ compile, ;
11385: @end example
11386:
11387:
11388: Nothing can be more confusing than a good example, so here is
11389: one. First let's declare a text object (called
11390: @code{button}), that stores text and position:
11391:
11392: @example
11393: object class
11394: cell var text
11395: cell var len
11396: cell var x
11397: cell var y
11398: method init
11399: method draw
11400: end-class button
11401: @end example
11402:
11403: @noindent
11404: Now, implement the two methods, @code{draw} and @code{init}:
11405:
11406: @example
11407: :noname ( o -- )
11408: >r r@ x @ r@ y @ at-xy r@ text @ r> len @ type ;
11409: button defines draw
11410: :noname ( addr u o -- )
11411: >r 0 r@ x ! 0 r@ y ! r@ len ! r> text ! ;
11412: button defines init
11413: @end example
11414:
11415: @noindent
11416: To demonstrate inheritance, we define a class @code{bold-button}, with no
11417: new data and no new methods:
11418:
11419: @example
11420: button class
11421: end-class bold-button
11422:
11423: : bold 27 emit ." [1m" ;
11424: : normal 27 emit ." [0m" ;
11425: @end example
11426:
11427: @noindent
11428: The class @code{bold-button} has a different draw method to
11429: @code{button}, but the new method is defined in terms of the draw method
11430: for @code{button}:
11431:
11432: @example
11433: :noname bold [ button :: draw ] normal ; bold-button defines draw
11434: @end example
11435:
11436: @noindent
11437: Finally, create two objects and apply methods:
11438:
11439: @example
11440: button new Constant foo
11441: s" thin foo" foo init
11442: page
11443: foo draw
11444: bold-button new Constant bar
11445: s" fat bar" bar init
11446: 1 bar y !
11447: bar draw
11448: @end example
11449:
11450:
11451: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11452: @subsection Comparison with other object models
11453: @cindex comparison of object models
11454: @cindex object models, comparison
11455:
11456: Many object-oriented Forth extensions have been proposed (@cite{A survey
11457: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11458: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11459: relation of the object models described here to two well-known and two
11460: closely-related (by the use of method maps) models.
11461:
11462: @cindex Neon model
11463: The most popular model currently seems to be the Neon model (see
11464: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11465: 1997) by Andrew McKewan) but this model has a number of limitations
11466: @footnote{A longer version of this critique can be
11467: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11468: Dimensions, May 1997) by Anton Ertl.}:
11469:
11470: @itemize @bullet
11471: @item
11472: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11473: to pass objects on the stack.
11474:
11475: @item
11476: It requires that the selector parses the input stream (at
11477: compile time); this leads to reduced extensibility and to bugs that are+
11478: hard to find.
11479:
11480: @item
11481: It allows using every selector to every object;
11482: this eliminates the need for classes, but makes it harder to create
11483: efficient implementations.
11484: @end itemize
11485:
11486: @cindex Pountain's object-oriented model
11487: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11488: Press, London, 1987) by Dick Pountain. However, it is not really about
11489: object-oriented programming, because it hardly deals with late
11490: binding. Instead, it focuses on features like information hiding and
11491: overloading that are characteristic of modular languages like Ada (83).
11492:
11493: @cindex Zsoter's object-oriented model
11494: In @cite{Does late binding have to be slow?} (Forth Dimensions 18(1)
11495: 1996, pages 31-35) Andras Zsoter describes a model that makes heavy use
11496: of an active object (like @code{this} in @file{objects.fs}): The active
11497: object is not only used for accessing all fields, but also specifies the
11498: receiving object of every selector invocation; you have to change the
11499: active object explicitly with @code{@{ ... @}}, whereas in
11500: @file{objects.fs} it changes more or less implicitly at @code{m:
11501: ... ;m}. Such a change at the method entry point is unnecessary with the
11502: Zsoter's model, because the receiving object is the active object
11503: already. On the other hand, the explicit change is absolutely necessary
11504: in that model, because otherwise no one could ever change the active
11505: object. An ANS Forth implementation of this model is available at
11506: @uref{http://www.forth.org/fig/oopf.html}.
11507:
11508: @cindex @file{oof.fs}, differences to other models
11509: The @file{oof.fs} model combines information hiding and overloading
11510: resolution (by keeping names in various word lists) with object-oriented
11511: programming. It sets the active object implicitly on method entry, but
11512: also allows explicit changing (with @code{>o...o>} or with
11513: @code{with...endwith}). It uses parsing and state-smart objects and
11514: classes for resolving overloading and for early binding: the object or
11515: class parses the selector and determines the method from this. If the
11516: selector is not parsed by an object or class, it performs a call to the
11517: selector for the active object (late binding), like Zsoter's model.
11518: Fields are always accessed through the active object. The big
11519: disadvantage of this model is the parsing and the state-smartness, which
11520: reduces extensibility and increases the opportunities for subtle bugs;
11521: essentially, you are only safe if you never tick or @code{postpone} an
11522: object or class (Bernd disagrees, but I (Anton) am not convinced).
11523:
11524: @cindex @file{mini-oof.fs}, differences to other models
11525: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11526: version of the @file{objects.fs} model, but syntactically it is a
11527: mixture of the @file{objects.fs} and @file{oof.fs} models.
11528:
11529: @c -------------------------------------------------------------
11530: @node Passing Commands to the OS, Keeping track of Time, Object-oriented Forth, Words
11531: @section Passing Commands to the Operating System
11532: @cindex operating system - passing commands
11533: @cindex shell commands
11534:
11535: Gforth allows you to pass an arbitrary string to the host operating
11536: system shell (if such a thing exists) for execution.
11537:
11538:
11539: doc-sh
11540: doc-system
11541: doc-$?
11542: doc-getenv
11543:
11544:
11545: @c -------------------------------------------------------------
11546: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
11547: @section Keeping track of Time
11548: @cindex time-related words
11549:
11550: Gforth implements time-related operations by making calls to the C
11551: library function, @code{gettimeofday}.
11552:
11553: doc-ms
11554: doc-time&date
11555:
11556:
11557:
11558: @c -------------------------------------------------------------
11559: @node Miscellaneous Words, , Keeping track of Time, Words
11560: @section Miscellaneous Words
11561: @cindex miscellaneous words
11562:
11563: @comment TODO find homes for these
11564:
11565: These section lists the ANS Forth words that are not documented
11566: elsewhere in this manual. Ultimately, they all need proper homes.
11567:
11568: doc-[compile]
11569:
11570:
11571: The following ANS Forth words are not currently supported by Gforth
11572: (@pxref{ANS conformance}):
11573:
11574: @code{EDITOR}
11575: @code{EMIT?}
11576: @code{FORGET}
11577:
11578: @c ******************************************************************
11579: @node Error messages, Tools, Words, Top
11580: @chapter Error messages
11581: @cindex error messages
11582: @cindex backtrace
11583:
11584: A typical Gforth error message looks like this:
11585:
11586: @example
11587: in file included from :-1
11588: in file included from ./yyy.fs:1
11589: ./xxx.fs:4: Invalid memory address
11590: bar
11591: ^^^
11592: $400E664C @@
11593: $400E6664 foo
11594: @end example
11595:
11596: The message identifying the error is @code{Invalid memory address}. The
11597: error happened when text-interpreting line 4 of the file
11598: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
11599: word on the line where the error happened, is pointed out (with
11600: @code{^^^}).
11601:
11602: The file containing the error was included in line 1 of @file{./yyy.fs},
11603: and @file{yyy.fs} was included from a non-file (in this case, by giving
11604: @file{yyy.fs} as command-line parameter to Gforth).
11605:
11606: At the end of the error message you find a return stack dump that can be
11607: interpreted as a backtrace (possibly empty). On top you find the top of
11608: the return stack when the @code{throw} happened, and at the bottom you
11609: find the return stack entry just above the return stack of the topmost
11610: text interpreter.
11611:
11612: To the right of most return stack entries you see a guess for the word
11613: that pushed that return stack entry as its return address. This gives a
11614: backtrace. In our case we see that @code{bar} called @code{foo}, and
11615: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
11616: address} exception).
11617:
11618: Note that the backtrace is not perfect: We don't know which return stack
11619: entries are return addresses (so we may get false positives); and in
11620: some cases (e.g., for @code{abort"}) we cannot determine from the return
11621: address the word that pushed the return address, so for some return
11622: addresses you see no names in the return stack dump.
11623:
11624: @cindex @code{catch} and backtraces
11625: The return stack dump represents the return stack at the time when a
11626: specific @code{throw} was executed. In programs that make use of
11627: @code{catch}, it is not necessarily clear which @code{throw} should be
11628: used for the return stack dump (e.g., consider one @code{throw} that
11629: indicates an error, which is caught, and during recovery another error
11630: happens; which @code{throw} should be used for the stack dump?). Gforth
11631: presents the return stack dump for the first @code{throw} after the last
11632: executed (not returned-to) @code{catch}; this works well in the usual
11633: case.
11634:
11635: @cindex @code{gforth-fast} and backtraces
11636: @cindex @code{gforth-fast}, difference from @code{gforth}
11637: @cindex backtraces with @code{gforth-fast}
11638: @cindex return stack dump with @code{gforth-fast}
11639: @code{gforth} is able to do a return stack dump for throws generated
11640: from primitives (e.g., invalid memory address, stack empty etc.);
11641: @code{gforth-fast} is only able to do a return stack dump from a
11642: directly called @code{throw} (including @code{abort} etc.). This is the
11643: only difference (apart from a speed factor of between 1.15 (K6-2) and
11644: 1.6 (21164A)) between @code{gforth} and @code{gforth-fast}. Given an
11645: exception caused by a primitive in @code{gforth-fast}, you will
11646: typically see no return stack dump at all; however, if the exception is
11647: caught by @code{catch} (e.g., for restoring some state), and then
11648: @code{throw}n again, the return stack dump will be for the first such
11649: @code{throw}.
11650:
11651: @c ******************************************************************
11652: @node Tools, ANS conformance, Error messages, Top
11653: @chapter Tools
11654:
11655: @menu
11656: * ANS Report:: Report the words used, sorted by wordset.
11657: @end menu
11658:
11659: See also @ref{Emacs and Gforth}.
11660:
11661: @node ANS Report, , Tools, Tools
11662: @section @file{ans-report.fs}: Report the words used, sorted by wordset
11663: @cindex @file{ans-report.fs}
11664: @cindex report the words used in your program
11665: @cindex words used in your program
11666:
11667: If you want to label a Forth program as ANS Forth Program, you must
11668: document which wordsets the program uses; for extension wordsets, it is
11669: helpful to list the words the program requires from these wordsets
11670: (because Forth systems are allowed to provide only some words of them).
11671:
11672: The @file{ans-report.fs} tool makes it easy for you to determine which
11673: words from which wordset and which non-ANS words your application
11674: uses. You simply have to include @file{ans-report.fs} before loading the
11675: program you want to check. After loading your program, you can get the
11676: report with @code{print-ans-report}. A typical use is to run this as
11677: batch job like this:
11678: @example
11679: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
11680: @end example
11681:
11682: The output looks like this (for @file{compat/control.fs}):
11683: @example
11684: The program uses the following words
11685: from CORE :
11686: : POSTPONE THEN ; immediate ?dup IF 0=
11687: from BLOCK-EXT :
11688: \
11689: from FILE :
11690: (
11691: @end example
11692:
11693: @subsection Caveats
11694:
11695: Note that @file{ans-report.fs} just checks which words are used, not whether
11696: they are used in an ANS Forth conforming way!
11697:
11698: Some words are defined in several wordsets in the
11699: standard. @file{ans-report.fs} reports them for only one of the
11700: wordsets, and not necessarily the one you expect. It depends on usage
11701: which wordset is the right one to specify. E.g., if you only use the
11702: compilation semantics of @code{S"}, it is a Core word; if you also use
11703: its interpretation semantics, it is a File word.
11704:
11705: @c ******************************************************************
11706: @node ANS conformance, Standard vs Extensions, Tools, Top
11707: @chapter ANS conformance
11708: @cindex ANS conformance of Gforth
11709:
11710: To the best of our knowledge, Gforth is an
11711:
11712: ANS Forth System
11713: @itemize @bullet
11714: @item providing the Core Extensions word set
11715: @item providing the Block word set
11716: @item providing the Block Extensions word set
11717: @item providing the Double-Number word set
11718: @item providing the Double-Number Extensions word set
11719: @item providing the Exception word set
11720: @item providing the Exception Extensions word set
11721: @item providing the Facility word set
11722: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
11723: @item providing the File Access word set
11724: @item providing the File Access Extensions word set
11725: @item providing the Floating-Point word set
11726: @item providing the Floating-Point Extensions word set
11727: @item providing the Locals word set
11728: @item providing the Locals Extensions word set
11729: @item providing the Memory-Allocation word set
11730: @item providing the Memory-Allocation Extensions word set (that one's easy)
11731: @item providing the Programming-Tools word set
11732: @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
11733: @item providing the Search-Order word set
11734: @item providing the Search-Order Extensions word set
11735: @item providing the String word set
11736: @item providing the String Extensions word set (another easy one)
11737: @end itemize
11738:
11739: @cindex system documentation
11740: In addition, ANS Forth systems are required to document certain
11741: implementation choices. This chapter tries to meet these
11742: requirements. In many cases it gives a way to ask the system for the
11743: information instead of providing the information directly, in
11744: particular, if the information depends on the processor, the operating
11745: system or the installation options chosen, or if they are likely to
11746: change during the maintenance of Gforth.
11747:
11748: @comment The framework for the rest has been taken from pfe.
11749:
11750: @menu
11751: * The Core Words::
11752: * The optional Block word set::
11753: * The optional Double Number word set::
11754: * The optional Exception word set::
11755: * The optional Facility word set::
11756: * The optional File-Access word set::
11757: * The optional Floating-Point word set::
11758: * The optional Locals word set::
11759: * The optional Memory-Allocation word set::
11760: * The optional Programming-Tools word set::
11761: * The optional Search-Order word set::
11762: @end menu
11763:
11764:
11765: @c =====================================================================
11766: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
11767: @comment node-name, next, previous, up
11768: @section The Core Words
11769: @c =====================================================================
11770: @cindex core words, system documentation
11771: @cindex system documentation, core words
11772:
11773: @menu
11774: * core-idef:: Implementation Defined Options
11775: * core-ambcond:: Ambiguous Conditions
11776: * core-other:: Other System Documentation
11777: @end menu
11778:
11779: @c ---------------------------------------------------------------------
11780: @node core-idef, core-ambcond, The Core Words, The Core Words
11781: @subsection Implementation Defined Options
11782: @c ---------------------------------------------------------------------
11783: @cindex core words, implementation-defined options
11784: @cindex implementation-defined options, core words
11785:
11786:
11787: @table @i
11788: @item (Cell) aligned addresses:
11789: @cindex cell-aligned addresses
11790: @cindex aligned addresses
11791: processor-dependent. Gforth's alignment words perform natural alignment
11792: (e.g., an address aligned for a datum of size 8 is divisible by
11793: 8). Unaligned accesses usually result in a @code{-23 THROW}.
11794:
11795: @item @code{EMIT} and non-graphic characters:
11796: @cindex @code{EMIT} and non-graphic characters
11797: @cindex non-graphic characters and @code{EMIT}
11798: The character is output using the C library function (actually, macro)
11799: @code{putc}.
11800:
11801: @item character editing of @code{ACCEPT} and @code{EXPECT}:
11802: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
11803: @cindex editing in @code{ACCEPT} and @code{EXPECT}
11804: @cindex @code{ACCEPT}, editing
11805: @cindex @code{EXPECT}, editing
11806: This is modeled on the GNU readline library (@pxref{Readline
11807: Interaction, , Command Line Editing, readline, The GNU Readline
11808: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
11809: producing a full word completion every time you type it (instead of
11810: producing the common prefix of all completions). @xref{Command-line editing}.
11811:
11812: @item character set:
11813: @cindex character set
11814: The character set of your computer and display device. Gforth is
11815: 8-bit-clean (but some other component in your system may make trouble).
11816:
11817: @item Character-aligned address requirements:
11818: @cindex character-aligned address requirements
11819: installation-dependent. Currently a character is represented by a C
11820: @code{unsigned char}; in the future we might switch to @code{wchar_t}
11821: (Comments on that requested).
11822:
11823: @item character-set extensions and matching of names:
11824: @cindex character-set extensions and matching of names
11825: @cindex case-sensitivity for name lookup
11826: @cindex name lookup, case-sensitivity
11827: @cindex locale and case-sensitivity
11828: Any character except the ASCII NUL character can be used in a
11829: name. Matching is case-insensitive (except in @code{TABLE}s). The
11830: matching is performed using the C library function @code{strncasecmp}, whose
11831: function is probably influenced by the locale. E.g., the @code{C} locale
11832: does not know about accents and umlauts, so they are matched
11833: case-sensitively in that locale. For portability reasons it is best to
11834: write programs such that they work in the @code{C} locale. Then one can
11835: use libraries written by a Polish programmer (who might use words
11836: containing ISO Latin-2 encoded characters) and by a French programmer
11837: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
11838: funny results for some of the words (which ones, depends on the font you
11839: are using)). Also, the locale you prefer may not be available in other
11840: operating systems. Hopefully, Unicode will solve these problems one day.
11841:
11842: @item conditions under which control characters match a space delimiter:
11843: @cindex space delimiters
11844: @cindex control characters as delimiters
11845: If @code{WORD} is called with the space character as a delimiter, all
11846: white-space characters (as identified by the C macro @code{isspace()})
11847: are delimiters. @code{PARSE}, on the other hand, treats space like other
11848: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
11849: like @code{PARSE} otherwise. @code{(NAME)}, which is used by the outer
11850: interpreter (aka text interpreter) by default, treats all white-space
11851: characters as delimiters.
11852:
11853: @item format of the control-flow stack:
11854: @cindex control-flow stack, format
11855: The data stack is used as control-flow stack. The size of a control-flow
11856: stack item in cells is given by the constant @code{cs-item-size}. At the
11857: time of this writing, an item consists of a (pointer to a) locals list
11858: (third), an address in the code (second), and a tag for identifying the
11859: item (TOS). The following tags are used: @code{defstart},
11860: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
11861: @code{scopestart}.
11862:
11863: @item conversion of digits > 35
11864: @cindex digits > 35
11865: The characters @code{[\]^_'} are the digits with the decimal value
11866: 36@minus{}41. There is no way to input many of the larger digits.
11867:
11868: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
11869: @cindex @code{EXPECT}, display after end of input
11870: @cindex @code{ACCEPT}, display after end of input
11871: The cursor is moved to the end of the entered string. If the input is
11872: terminated using the @kbd{Return} key, a space is typed.
11873:
11874: @item exception abort sequence of @code{ABORT"}:
11875: @cindex exception abort sequence of @code{ABORT"}
11876: @cindex @code{ABORT"}, exception abort sequence
11877: The error string is stored into the variable @code{"error} and a
11878: @code{-2 throw} is performed.
11879:
11880: @item input line terminator:
11881: @cindex input line terminator
11882: @cindex line terminator on input
11883: @cindex newline character on input
11884: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
11885: lines. One of these characters is typically produced when you type the
11886: @kbd{Enter} or @kbd{Return} key.
11887:
11888: @item maximum size of a counted string:
11889: @cindex maximum size of a counted string
11890: @cindex counted string, maximum size
11891: @code{s" /counted-string" environment? drop .}. Currently 255 characters
11892: on all ports, but this may change.
11893:
11894: @item maximum size of a parsed string:
11895: @cindex maximum size of a parsed string
11896: @cindex parsed string, maximum size
11897: Given by the constant @code{/line}. Currently 255 characters.
11898:
11899: @item maximum size of a definition name, in characters:
11900: @cindex maximum size of a definition name, in characters
11901: @cindex name, maximum length
11902: 31
11903:
11904: @item maximum string length for @code{ENVIRONMENT?}, in characters:
11905: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
11906: @cindex @code{ENVIRONMENT?} string length, maximum
11907: 31
11908:
11909: @item method of selecting the user input device:
11910: @cindex user input device, method of selecting
11911: The user input device is the standard input. There is currently no way to
11912: change it from within Gforth. However, the input can typically be
11913: redirected in the command line that starts Gforth.
11914:
11915: @item method of selecting the user output device:
11916: @cindex user output device, method of selecting
11917: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
11918: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
11919: output when the user output device is a terminal, otherwise the output
11920: is buffered.
11921:
11922: @item methods of dictionary compilation:
11923: What are we expected to document here?
11924:
11925: @item number of bits in one address unit:
11926: @cindex number of bits in one address unit
11927: @cindex address unit, size in bits
11928: @code{s" address-units-bits" environment? drop .}. 8 in all current
11929: ports.
11930:
11931: @item number representation and arithmetic:
11932: @cindex number representation and arithmetic
11933: Processor-dependent. Binary two's complement on all current ports.
11934:
11935: @item ranges for integer types:
11936: @cindex ranges for integer types
11937: @cindex integer types, ranges
11938: Installation-dependent. Make environmental queries for @code{MAX-N},
11939: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
11940: unsigned (and positive) types is 0. The lower bound for signed types on
11941: two's complement and one's complement machines machines can be computed
11942: by adding 1 to the upper bound.
11943:
11944: @item read-only data space regions:
11945: @cindex read-only data space regions
11946: @cindex data-space, read-only regions
11947: The whole Forth data space is writable.
11948:
11949: @item size of buffer at @code{WORD}:
11950: @cindex size of buffer at @code{WORD}
11951: @cindex @code{WORD} buffer size
11952: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
11953: shared with the pictured numeric output string. If overwriting
11954: @code{PAD} is acceptable, it is as large as the remaining dictionary
11955: space, although only as much can be sensibly used as fits in a counted
11956: string.
11957:
11958: @item size of one cell in address units:
11959: @cindex cell size
11960: @code{1 cells .}.
11961:
11962: @item size of one character in address units:
11963: @cindex char size
11964: @code{1 chars .}. 1 on all current ports.
11965:
11966: @item size of the keyboard terminal buffer:
11967: @cindex size of the keyboard terminal buffer
11968: @cindex terminal buffer, size
11969: Varies. You can determine the size at a specific time using @code{lp@@
11970: tib - .}. It is shared with the locals stack and TIBs of files that
11971: include the current file. You can change the amount of space for TIBs
11972: and locals stack at Gforth startup with the command line option
11973: @code{-l}.
11974:
11975: @item size of the pictured numeric output buffer:
11976: @cindex size of the pictured numeric output buffer
11977: @cindex pictured numeric output buffer, size
11978: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
11979: shared with @code{WORD}.
11980:
11981: @item size of the scratch area returned by @code{PAD}:
11982: @cindex size of the scratch area returned by @code{PAD}
11983: @cindex @code{PAD} size
11984: The remainder of dictionary space. @code{unused pad here - - .}.
11985:
11986: @item system case-sensitivity characteristics:
11987: @cindex case-sensitivity characteristics
11988: Dictionary searches are case-insensitive (except in
11989: @code{TABLE}s). However, as explained above under @i{character-set
11990: extensions}, the matching for non-ASCII characters is determined by the
11991: locale you are using. In the default @code{C} locale all non-ASCII
11992: characters are matched case-sensitively.
11993:
11994: @item system prompt:
11995: @cindex system prompt
11996: @cindex prompt
11997: @code{ ok} in interpret state, @code{ compiled} in compile state.
11998:
11999: @item division rounding:
12000: @cindex division rounding
12001: installation dependent. @code{s" floored" environment? drop .}. We leave
12002: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
12003: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
12004:
12005: @item values of @code{STATE} when true:
12006: @cindex @code{STATE} values
12007: -1.
12008:
12009: @item values returned after arithmetic overflow:
12010: On two's complement machines, arithmetic is performed modulo
12011: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12012: arithmetic (with appropriate mapping for signed types). Division by zero
12013: typically results in a @code{-55 throw} (Floating-point unidentified
12014: fault), although a @code{-10 throw} (divide by zero) would be more
12015: appropriate.
12016:
12017: @item whether the current definition can be found after @t{DOES>}:
12018: @cindex @t{DOES>}, visibility of current definition
12019: No.
12020:
12021: @end table
12022:
12023: @c ---------------------------------------------------------------------
12024: @node core-ambcond, core-other, core-idef, The Core Words
12025: @subsection Ambiguous conditions
12026: @c ---------------------------------------------------------------------
12027: @cindex core words, ambiguous conditions
12028: @cindex ambiguous conditions, core words
12029:
12030: @table @i
12031:
12032: @item a name is neither a word nor a number:
12033: @cindex name not found
12034: @cindex undefined word
12035: @code{-13 throw} (Undefined word). Actually, @code{-13 bounce}, which
12036: preserves the data and FP stack, so you don't lose more work than
12037: necessary.
12038:
12039: @item a definition name exceeds the maximum length allowed:
12040: @cindex word name too long
12041: @code{-19 throw} (Word name too long)
12042:
12043: @item addressing a region not inside the various data spaces of the forth system:
12044: @cindex Invalid memory address
12045: The stacks, code space and header space are accessible. Machine code space is
12046: typically readable. Accessing other addresses gives results dependent on
12047: the operating system. On decent systems: @code{-9 throw} (Invalid memory
12048: address).
12049:
12050: @item argument type incompatible with parameter:
12051: @cindex argument type mismatch
12052: This is usually not caught. Some words perform checks, e.g., the control
12053: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
12054: mismatch).
12055:
12056: @item attempting to obtain the execution token of a word with undefined execution semantics:
12057: @cindex Interpreting a compile-only word, for @code{'} etc.
12058: @cindex execution token of words with undefined execution semantics
12059: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
12060: get an execution token for @code{compile-only-error} (which performs a
12061: @code{-14 throw} when executed).
12062:
12063: @item dividing by zero:
12064: @cindex dividing by zero
12065: @cindex floating point unidentified fault, integer division
12066: On better platforms, this produces a @code{-10 throw} (Division by
12067: zero); on other systems, this typically results in a @code{-55 throw}
12068: (Floating-point unidentified fault).
12069:
12070: @item insufficient data stack or return stack space:
12071: @cindex insufficient data stack or return stack space
12072: @cindex stack overflow
12073: @cindex address alignment exception, stack overflow
12074: @cindex Invalid memory address, stack overflow
12075: Depending on the operating system, the installation, and the invocation
12076: of Gforth, this is either checked by the memory management hardware, or
12077: it is not checked. If it is checked, you typically get a @code{-3 throw}
12078: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
12079: throw} (Invalid memory address) (depending on the platform and how you
12080: achieved the overflow) as soon as the overflow happens. If it is not
12081: checked, overflows typically result in mysterious illegal memory
12082: accesses, producing @code{-9 throw} (Invalid memory address) or
12083: @code{-23 throw} (Address alignment exception); they might also destroy
12084: the internal data structure of @code{ALLOCATE} and friends, resulting in
12085: various errors in these words.
12086:
12087: @item insufficient space for loop control parameters:
12088: @cindex insufficient space for loop control parameters
12089: like other return stack overflows.
12090:
12091: @item insufficient space in the dictionary:
12092: @cindex insufficient space in the dictionary
12093: @cindex dictionary overflow
12094: If you try to allot (either directly with @code{allot}, or indirectly
12095: with @code{,}, @code{create} etc.) more memory than available in the
12096: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
12097: to access memory beyond the end of the dictionary, the results are
12098: similar to stack overflows.
12099:
12100: @item interpreting a word with undefined interpretation semantics:
12101: @cindex interpreting a word with undefined interpretation semantics
12102: @cindex Interpreting a compile-only word
12103: For some words, we have defined interpretation semantics. For the
12104: others: @code{-14 throw} (Interpreting a compile-only word).
12105:
12106: @item modifying the contents of the input buffer or a string literal:
12107: @cindex modifying the contents of the input buffer or a string literal
12108: These are located in writable memory and can be modified.
12109:
12110: @item overflow of the pictured numeric output string:
12111: @cindex overflow of the pictured numeric output string
12112: @cindex pictured numeric output string, overflow
12113: @code{-17 throw} (Pictured numeric ouput string overflow).
12114:
12115: @item parsed string overflow:
12116: @cindex parsed string overflow
12117: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
12118:
12119: @item producing a result out of range:
12120: @cindex result out of range
12121: On two's complement machines, arithmetic is performed modulo
12122: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12123: arithmetic (with appropriate mapping for signed types). Division by zero
12124: typically results in a @code{-10 throw} (divide by zero) or @code{-55
12125: throw} (floating point unidentified fault). @code{convert} and
12126: @code{>number} currently overflow silently.
12127:
12128: @item reading from an empty data or return stack:
12129: @cindex stack empty
12130: @cindex stack underflow
12131: @cindex return stack underflow
12132: The data stack is checked by the outer (aka text) interpreter after
12133: every word executed. If it has underflowed, a @code{-4 throw} (Stack
12134: underflow) is performed. Apart from that, stacks may be checked or not,
12135: depending on operating system, installation, and invocation. If they are
12136: caught by a check, they typically result in @code{-4 throw} (Stack
12137: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
12138: (Invalid memory address), depending on the platform and which stack
12139: underflows and by how much. Note that even if the system uses checking
12140: (through the MMU), your program may have to underflow by a significant
12141: number of stack items to trigger the reaction (the reason for this is
12142: that the MMU, and therefore the checking, works with a page-size
12143: granularity). If there is no checking, the symptoms resulting from an
12144: underflow are similar to those from an overflow. Unbalanced return
12145: stack errors result in a variaty of symptoms, including @code{-9 throw}
12146: (Invalid memory address) and Illegal Instruction (typically @code{-260
12147: throw}).
12148:
12149: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
12150: @cindex unexpected end of the input buffer
12151: @cindex zero-length string as a name
12152: @cindex Attempt to use zero-length string as a name
12153: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
12154: use zero-length string as a name). Words like @code{'} probably will not
12155: find what they search. Note that it is possible to create zero-length
12156: names with @code{nextname} (should it not?).
12157:
12158: @item @code{>IN} greater than input buffer:
12159: @cindex @code{>IN} greater than input buffer
12160: The next invocation of a parsing word returns a string with length 0.
12161:
12162: @item @code{RECURSE} appears after @code{DOES>}:
12163: @cindex @code{RECURSE} appears after @code{DOES>}
12164: Compiles a recursive call to the defining word, not to the defined word.
12165:
12166: @item argument input source different than current input source for @code{RESTORE-INPUT}:
12167: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
12168: @cindex argument type mismatch, @code{RESTORE-INPUT}
12169: @cindex @code{RESTORE-INPUT}, Argument type mismatch
12170: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
12171: the end of the file was reached), its source-id may be
12172: reused. Therefore, restoring an input source specification referencing a
12173: closed file may lead to unpredictable results instead of a @code{-12
12174: THROW}.
12175:
12176: In the future, Gforth may be able to restore input source specifications
12177: from other than the current input source.
12178:
12179: @item data space containing definitions gets de-allocated:
12180: @cindex data space containing definitions gets de-allocated
12181: Deallocation with @code{allot} is not checked. This typically results in
12182: memory access faults or execution of illegal instructions.
12183:
12184: @item data space read/write with incorrect alignment:
12185: @cindex data space read/write with incorrect alignment
12186: @cindex alignment faults
12187: @cindex address alignment exception
12188: Processor-dependent. Typically results in a @code{-23 throw} (Address
12189: alignment exception). Under Linux-Intel on a 486 or later processor with
12190: alignment turned on, incorrect alignment results in a @code{-9 throw}
12191: (Invalid memory address). There are reportedly some processors with
12192: alignment restrictions that do not report violations.
12193:
12194: @item data space pointer not properly aligned, @code{,}, @code{C,}:
12195: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
12196: Like other alignment errors.
12197:
12198: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
12199: Like other stack underflows.
12200:
12201: @item loop control parameters not available:
12202: @cindex loop control parameters not available
12203: Not checked. The counted loop words simply assume that the top of return
12204: stack items are loop control parameters and behave accordingly.
12205:
12206: @item most recent definition does not have a name (@code{IMMEDIATE}):
12207: @cindex most recent definition does not have a name (@code{IMMEDIATE})
12208: @cindex last word was headerless
12209: @code{abort" last word was headerless"}.
12210:
12211: @item name not defined by @code{VALUE} used by @code{TO}:
12212: @cindex name not defined by @code{VALUE} used by @code{TO}
12213: @cindex @code{TO} on non-@code{VALUE}s
12214: @cindex Invalid name argument, @code{TO}
12215: @code{-32 throw} (Invalid name argument) (unless name is a local or was
12216: defined by @code{CONSTANT}; in the latter case it just changes the constant).
12217:
12218: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
12219: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
12220: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
12221: @code{-13 throw} (Undefined word)
12222:
12223: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
12224: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
12225: Gforth behaves as if they were of the same type. I.e., you can predict
12226: the behaviour by interpreting all parameters as, e.g., signed.
12227:
12228: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
12229: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
12230: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
12231: compilation semantics of @code{TO}.
12232:
12233: @item String longer than a counted string returned by @code{WORD}:
12234: @cindex string longer than a counted string returned by @code{WORD}
12235: @cindex @code{WORD}, string overflow
12236: Not checked. The string will be ok, but the count will, of course,
12237: contain only the least significant bits of the length.
12238:
12239: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
12240: @cindex @code{LSHIFT}, large shift counts
12241: @cindex @code{RSHIFT}, large shift counts
12242: Processor-dependent. Typical behaviours are returning 0 and using only
12243: the low bits of the shift count.
12244:
12245: @item word not defined via @code{CREATE}:
12246: @cindex @code{>BODY} of non-@code{CREATE}d words
12247: @code{>BODY} produces the PFA of the word no matter how it was defined.
12248:
12249: @cindex @code{DOES>} of non-@code{CREATE}d words
12250: @code{DOES>} changes the execution semantics of the last defined word no
12251: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
12252: @code{CREATE , DOES>}.
12253:
12254: @item words improperly used outside @code{<#} and @code{#>}:
12255: Not checked. As usual, you can expect memory faults.
12256:
12257: @end table
12258:
12259:
12260: @c ---------------------------------------------------------------------
12261: @node core-other, , core-ambcond, The Core Words
12262: @subsection Other system documentation
12263: @c ---------------------------------------------------------------------
12264: @cindex other system documentation, core words
12265: @cindex core words, other system documentation
12266:
12267: @table @i
12268: @item nonstandard words using @code{PAD}:
12269: @cindex @code{PAD} use by nonstandard words
12270: None.
12271:
12272: @item operator's terminal facilities available:
12273: @cindex operator's terminal facilities available
12274: After processing the command line, Gforth goes into interactive mode,
12275: and you can give commands to Gforth interactively. The actual facilities
12276: available depend on how you invoke Gforth.
12277:
12278: @item program data space available:
12279: @cindex program data space available
12280: @cindex data space available
12281: @code{UNUSED .} gives the remaining dictionary space. The total
12282: dictionary space can be specified with the @code{-m} switch
12283: (@pxref{Invoking Gforth}) when Gforth starts up.
12284:
12285: @item return stack space available:
12286: @cindex return stack space available
12287: You can compute the total return stack space in cells with
12288: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
12289: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
12290:
12291: @item stack space available:
12292: @cindex stack space available
12293: You can compute the total data stack space in cells with
12294: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
12295: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
12296:
12297: @item system dictionary space required, in address units:
12298: @cindex system dictionary space required, in address units
12299: Type @code{here forthstart - .} after startup. At the time of this
12300: writing, this gives 80080 (bytes) on a 32-bit system.
12301: @end table
12302:
12303:
12304: @c =====================================================================
12305: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
12306: @section The optional Block word set
12307: @c =====================================================================
12308: @cindex system documentation, block words
12309: @cindex block words, system documentation
12310:
12311: @menu
12312: * block-idef:: Implementation Defined Options
12313: * block-ambcond:: Ambiguous Conditions
12314: * block-other:: Other System Documentation
12315: @end menu
12316:
12317:
12318: @c ---------------------------------------------------------------------
12319: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
12320: @subsection Implementation Defined Options
12321: @c ---------------------------------------------------------------------
12322: @cindex implementation-defined options, block words
12323: @cindex block words, implementation-defined options
12324:
12325: @table @i
12326: @item the format for display by @code{LIST}:
12327: @cindex @code{LIST} display format
12328: First the screen number is displayed, then 16 lines of 64 characters,
12329: each line preceded by the line number.
12330:
12331: @item the length of a line affected by @code{\}:
12332: @cindex length of a line affected by @code{\}
12333: @cindex @code{\}, line length in blocks
12334: 64 characters.
12335: @end table
12336:
12337:
12338: @c ---------------------------------------------------------------------
12339: @node block-ambcond, block-other, block-idef, The optional Block word set
12340: @subsection Ambiguous conditions
12341: @c ---------------------------------------------------------------------
12342: @cindex block words, ambiguous conditions
12343: @cindex ambiguous conditions, block words
12344:
12345: @table @i
12346: @item correct block read was not possible:
12347: @cindex block read not possible
12348: Typically results in a @code{throw} of some OS-derived value (between
12349: -512 and -2048). If the blocks file was just not long enough, blanks are
12350: supplied for the missing portion.
12351:
12352: @item I/O exception in block transfer:
12353: @cindex I/O exception in block transfer
12354: @cindex block transfer, I/O exception
12355: Typically results in a @code{throw} of some OS-derived value (between
12356: -512 and -2048).
12357:
12358: @item invalid block number:
12359: @cindex invalid block number
12360: @cindex block number invalid
12361: @code{-35 throw} (Invalid block number)
12362:
12363: @item a program directly alters the contents of @code{BLK}:
12364: @cindex @code{BLK}, altering @code{BLK}
12365: The input stream is switched to that other block, at the same
12366: position. If the storing to @code{BLK} happens when interpreting
12367: non-block input, the system will get quite confused when the block ends.
12368:
12369: @item no current block buffer for @code{UPDATE}:
12370: @cindex @code{UPDATE}, no current block buffer
12371: @code{UPDATE} has no effect.
12372:
12373: @end table
12374:
12375: @c ---------------------------------------------------------------------
12376: @node block-other, , block-ambcond, The optional Block word set
12377: @subsection Other system documentation
12378: @c ---------------------------------------------------------------------
12379: @cindex other system documentation, block words
12380: @cindex block words, other system documentation
12381:
12382: @table @i
12383: @item any restrictions a multiprogramming system places on the use of buffer addresses:
12384: No restrictions (yet).
12385:
12386: @item the number of blocks available for source and data:
12387: depends on your disk space.
12388:
12389: @end table
12390:
12391:
12392: @c =====================================================================
12393: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
12394: @section The optional Double Number word set
12395: @c =====================================================================
12396: @cindex system documentation, double words
12397: @cindex double words, system documentation
12398:
12399: @menu
12400: * double-ambcond:: Ambiguous Conditions
12401: @end menu
12402:
12403:
12404: @c ---------------------------------------------------------------------
12405: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
12406: @subsection Ambiguous conditions
12407: @c ---------------------------------------------------------------------
12408: @cindex double words, ambiguous conditions
12409: @cindex ambiguous conditions, double words
12410:
12411: @table @i
12412: @item @i{d} outside of range of @i{n} in @code{D>S}:
12413: @cindex @code{D>S}, @i{d} out of range of @i{n}
12414: The least significant cell of @i{d} is produced.
12415:
12416: @end table
12417:
12418:
12419: @c =====================================================================
12420: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
12421: @section The optional Exception word set
12422: @c =====================================================================
12423: @cindex system documentation, exception words
12424: @cindex exception words, system documentation
12425:
12426: @menu
12427: * exception-idef:: Implementation Defined Options
12428: @end menu
12429:
12430:
12431: @c ---------------------------------------------------------------------
12432: @node exception-idef, , The optional Exception word set, The optional Exception word set
12433: @subsection Implementation Defined Options
12434: @c ---------------------------------------------------------------------
12435: @cindex implementation-defined options, exception words
12436: @cindex exception words, implementation-defined options
12437:
12438: @table @i
12439: @item @code{THROW}-codes used in the system:
12440: @cindex @code{THROW}-codes used in the system
12441: The codes -256@minus{}-511 are used for reporting signals. The mapping
12442: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
12443: codes -512@minus{}-2047 are used for OS errors (for file and memory
12444: allocation operations). The mapping from OS error numbers to throw codes
12445: is -512@minus{}@code{errno}. One side effect of this mapping is that
12446: undefined OS errors produce a message with a strange number; e.g.,
12447: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
12448: @end table
12449:
12450: @c =====================================================================
12451: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
12452: @section The optional Facility word set
12453: @c =====================================================================
12454: @cindex system documentation, facility words
12455: @cindex facility words, system documentation
12456:
12457: @menu
12458: * facility-idef:: Implementation Defined Options
12459: * facility-ambcond:: Ambiguous Conditions
12460: @end menu
12461:
12462:
12463: @c ---------------------------------------------------------------------
12464: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
12465: @subsection Implementation Defined Options
12466: @c ---------------------------------------------------------------------
12467: @cindex implementation-defined options, facility words
12468: @cindex facility words, implementation-defined options
12469:
12470: @table @i
12471: @item encoding of keyboard events (@code{EKEY}):
12472: @cindex keyboard events, encoding in @code{EKEY}
12473: @cindex @code{EKEY}, encoding of keyboard events
12474: Keys corresponding to ASCII characters are encoded as ASCII characters.
12475: Other keys are encoded with the constants @code{k-left}, @code{k-right},
12476: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
12477: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
12478: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
12479:
12480:
12481: @item duration of a system clock tick:
12482: @cindex duration of a system clock tick
12483: @cindex clock tick duration
12484: System dependent. With respect to @code{MS}, the time is specified in
12485: microseconds. How well the OS and the hardware implement this, is
12486: another question.
12487:
12488: @item repeatability to be expected from the execution of @code{MS}:
12489: @cindex repeatability to be expected from the execution of @code{MS}
12490: @cindex @code{MS}, repeatability to be expected
12491: System dependent. On Unix, a lot depends on load. If the system is
12492: lightly loaded, and the delay is short enough that Gforth does not get
12493: swapped out, the performance should be acceptable. Under MS-DOS and
12494: other single-tasking systems, it should be good.
12495:
12496: @end table
12497:
12498:
12499: @c ---------------------------------------------------------------------
12500: @node facility-ambcond, , facility-idef, The optional Facility word set
12501: @subsection Ambiguous conditions
12502: @c ---------------------------------------------------------------------
12503: @cindex facility words, ambiguous conditions
12504: @cindex ambiguous conditions, facility words
12505:
12506: @table @i
12507: @item @code{AT-XY} can't be performed on user output device:
12508: @cindex @code{AT-XY} can't be performed on user output device
12509: Largely terminal dependent. No range checks are done on the arguments.
12510: No errors are reported. You may see some garbage appearing, you may see
12511: simply nothing happen.
12512:
12513: @end table
12514:
12515:
12516: @c =====================================================================
12517: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
12518: @section The optional File-Access word set
12519: @c =====================================================================
12520: @cindex system documentation, file words
12521: @cindex file words, system documentation
12522:
12523: @menu
12524: * file-idef:: Implementation Defined Options
12525: * file-ambcond:: Ambiguous Conditions
12526: @end menu
12527:
12528: @c ---------------------------------------------------------------------
12529: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
12530: @subsection Implementation Defined Options
12531: @c ---------------------------------------------------------------------
12532: @cindex implementation-defined options, file words
12533: @cindex file words, implementation-defined options
12534:
12535: @table @i
12536: @item file access methods used:
12537: @cindex file access methods used
12538: @code{R/O}, @code{R/W} and @code{BIN} work as you would
12539: expect. @code{W/O} translates into the C file opening mode @code{w} (or
12540: @code{wb}): The file is cleared, if it exists, and created, if it does
12541: not (with both @code{open-file} and @code{create-file}). Under Unix
12542: @code{create-file} creates a file with 666 permissions modified by your
12543: umask.
12544:
12545: @item file exceptions:
12546: @cindex file exceptions
12547: The file words do not raise exceptions (except, perhaps, memory access
12548: faults when you pass illegal addresses or file-ids).
12549:
12550: @item file line terminator:
12551: @cindex file line terminator
12552: System-dependent. Gforth uses C's newline character as line
12553: terminator. What the actual character code(s) of this are is
12554: system-dependent.
12555:
12556: @item file name format:
12557: @cindex file name format
12558: System dependent. Gforth just uses the file name format of your OS.
12559:
12560: @item information returned by @code{FILE-STATUS}:
12561: @cindex @code{FILE-STATUS}, returned information
12562: @code{FILE-STATUS} returns the most powerful file access mode allowed
12563: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
12564: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
12565: along with the returned mode.
12566:
12567: @item input file state after an exception when including source:
12568: @cindex exception when including source
12569: All files that are left via the exception are closed.
12570:
12571: @item @i{ior} values and meaning:
12572: @cindex @i{ior} values and meaning
12573: The @i{ior}s returned by the file and memory allocation words are
12574: intended as throw codes. They typically are in the range
12575: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
12576: @i{ior}s is -512@minus{}@i{errno}.
12577:
12578: @item maximum depth of file input nesting:
12579: @cindex maximum depth of file input nesting
12580: @cindex file input nesting, maximum depth
12581: limited by the amount of return stack, locals/TIB stack, and the number
12582: of open files available. This should not give you troubles.
12583:
12584: @item maximum size of input line:
12585: @cindex maximum size of input line
12586: @cindex input line size, maximum
12587: @code{/line}. Currently 255.
12588:
12589: @item methods of mapping block ranges to files:
12590: @cindex mapping block ranges to files
12591: @cindex files containing blocks
12592: @cindex blocks in files
12593: By default, blocks are accessed in the file @file{blocks.fb} in the
12594: current working directory. The file can be switched with @code{USE}.
12595:
12596: @item number of string buffers provided by @code{S"}:
12597: @cindex @code{S"}, number of string buffers
12598: 1
12599:
12600: @item size of string buffer used by @code{S"}:
12601: @cindex @code{S"}, size of string buffer
12602: @code{/line}. currently 255.
12603:
12604: @end table
12605:
12606: @c ---------------------------------------------------------------------
12607: @node file-ambcond, , file-idef, The optional File-Access word set
12608: @subsection Ambiguous conditions
12609: @c ---------------------------------------------------------------------
12610: @cindex file words, ambiguous conditions
12611: @cindex ambiguous conditions, file words
12612:
12613: @table @i
12614: @item attempting to position a file outside its boundaries:
12615: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
12616: @code{REPOSITION-FILE} is performed as usual: Afterwards,
12617: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
12618:
12619: @item attempting to read from file positions not yet written:
12620: @cindex reading from file positions not yet written
12621: End-of-file, i.e., zero characters are read and no error is reported.
12622:
12623: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
12624: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
12625: An appropriate exception may be thrown, but a memory fault or other
12626: problem is more probable.
12627:
12628: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
12629: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
12630: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
12631: The @i{ior} produced by the operation, that discovered the problem, is
12632: thrown.
12633:
12634: @item named file cannot be opened (@code{INCLUDED}):
12635: @cindex @code{INCLUDED}, named file cannot be opened
12636: The @i{ior} produced by @code{open-file} is thrown.
12637:
12638: @item requesting an unmapped block number:
12639: @cindex unmapped block numbers
12640: There are no unmapped legal block numbers. On some operating systems,
12641: writing a block with a large number may overflow the file system and
12642: have an error message as consequence.
12643:
12644: @item using @code{source-id} when @code{blk} is non-zero:
12645: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
12646: @code{source-id} performs its function. Typically it will give the id of
12647: the source which loaded the block. (Better ideas?)
12648:
12649: @end table
12650:
12651:
12652: @c =====================================================================
12653: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
12654: @section The optional Floating-Point word set
12655: @c =====================================================================
12656: @cindex system documentation, floating-point words
12657: @cindex floating-point words, system documentation
12658:
12659: @menu
12660: * floating-idef:: Implementation Defined Options
12661: * floating-ambcond:: Ambiguous Conditions
12662: @end menu
12663:
12664:
12665: @c ---------------------------------------------------------------------
12666: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
12667: @subsection Implementation Defined Options
12668: @c ---------------------------------------------------------------------
12669: @cindex implementation-defined options, floating-point words
12670: @cindex floating-point words, implementation-defined options
12671:
12672: @table @i
12673: @item format and range of floating point numbers:
12674: @cindex format and range of floating point numbers
12675: @cindex floating point numbers, format and range
12676: System-dependent; the @code{double} type of C.
12677:
12678: @item results of @code{REPRESENT} when @i{float} is out of range:
12679: @cindex @code{REPRESENT}, results when @i{float} is out of range
12680: System dependent; @code{REPRESENT} is implemented using the C library
12681: function @code{ecvt()} and inherits its behaviour in this respect.
12682:
12683: @item rounding or truncation of floating-point numbers:
12684: @cindex rounding of floating-point numbers
12685: @cindex truncation of floating-point numbers
12686: @cindex floating-point numbers, rounding or truncation
12687: System dependent; the rounding behaviour is inherited from the hosting C
12688: compiler. IEEE-FP-based (i.e., most) systems by default round to
12689: nearest, and break ties by rounding to even (i.e., such that the last
12690: bit of the mantissa is 0).
12691:
12692: @item size of floating-point stack:
12693: @cindex floating-point stack size
12694: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
12695: the floating-point stack (in floats). You can specify this on startup
12696: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
12697:
12698: @item width of floating-point stack:
12699: @cindex floating-point stack width
12700: @code{1 floats}.
12701:
12702: @end table
12703:
12704:
12705: @c ---------------------------------------------------------------------
12706: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
12707: @subsection Ambiguous conditions
12708: @c ---------------------------------------------------------------------
12709: @cindex floating-point words, ambiguous conditions
12710: @cindex ambiguous conditions, floating-point words
12711:
12712: @table @i
12713: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
12714: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
12715: System-dependent. Typically results in a @code{-23 THROW} like other
12716: alignment violations.
12717:
12718: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
12719: @cindex @code{f@@} used with an address that is not float aligned
12720: @cindex @code{f!} used with an address that is not float aligned
12721: System-dependent. Typically results in a @code{-23 THROW} like other
12722: alignment violations.
12723:
12724: @item floating-point result out of range:
12725: @cindex floating-point result out of range
12726: System-dependent. Can result in a @code{-55 THROW} (Floating-point
12727: unidentified fault), or can produce a special value representing, e.g.,
12728: Infinity.
12729:
12730: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
12731: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
12732: System-dependent. Typically results in an alignment fault like other
12733: alignment violations.
12734:
12735: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
12736: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
12737: The floating-point number is converted into decimal nonetheless.
12738:
12739: @item Both arguments are equal to zero (@code{FATAN2}):
12740: @cindex @code{FATAN2}, both arguments are equal to zero
12741: System-dependent. @code{FATAN2} is implemented using the C library
12742: function @code{atan2()}.
12743:
12744: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
12745: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
12746: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
12747: because of small errors and the tan will be a very large (or very small)
12748: but finite number.
12749:
12750: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
12751: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
12752: The result is rounded to the nearest float.
12753:
12754: @item dividing by zero:
12755: @cindex dividing by zero, floating-point
12756: @cindex floating-point dividing by zero
12757: @cindex floating-point unidentified fault, FP divide-by-zero
12758: @code{-55 throw} (Floating-point unidentified fault)
12759:
12760: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
12761: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
12762: System dependent. On IEEE-FP based systems the number is converted into
12763: an infinity.
12764:
12765: @item @i{float}<1 (@code{FACOSH}):
12766: @cindex @code{FACOSH}, @i{float}<1
12767: @cindex floating-point unidentified fault, @code{FACOSH}
12768: @code{-55 throw} (Floating-point unidentified fault)
12769:
12770: @item @i{float}=<-1 (@code{FLNP1}):
12771: @cindex @code{FLNP1}, @i{float}=<-1
12772: @cindex floating-point unidentified fault, @code{FLNP1}
12773: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
12774: negative infinity is typically produced for @i{float}=-1.
12775:
12776: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
12777: @cindex @code{FLN}, @i{float}=<0
12778: @cindex @code{FLOG}, @i{float}=<0
12779: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
12780: @code{-55 throw} (Floating-point unidentified fault). On IEEE-FP systems
12781: negative infinity is typically produced for @i{float}=0.
12782:
12783: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
12784: @cindex @code{FASINH}, @i{float}<0
12785: @cindex @code{FSQRT}, @i{float}<0
12786: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
12787: @code{-55 throw} (Floating-point unidentified fault). @code{fasinh}
12788: produces values for these inputs on my Linux box (Bug in the C library?)
12789:
12790: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
12791: @cindex @code{FACOS}, |@i{float}|>1
12792: @cindex @code{FASIN}, |@i{float}|>1
12793: @cindex @code{FATANH}, |@i{float}|>1
12794: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
12795: @code{-55 throw} (Floating-point unidentified fault).
12796:
12797: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
12798: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
12799: @cindex floating-point unidentified fault, @code{F>D}
12800: @code{-55 throw} (Floating-point unidentified fault).
12801:
12802: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
12803: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
12804: This does not happen.
12805: @end table
12806:
12807: @c =====================================================================
12808: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
12809: @section The optional Locals word set
12810: @c =====================================================================
12811: @cindex system documentation, locals words
12812: @cindex locals words, system documentation
12813:
12814: @menu
12815: * locals-idef:: Implementation Defined Options
12816: * locals-ambcond:: Ambiguous Conditions
12817: @end menu
12818:
12819:
12820: @c ---------------------------------------------------------------------
12821: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
12822: @subsection Implementation Defined Options
12823: @c ---------------------------------------------------------------------
12824: @cindex implementation-defined options, locals words
12825: @cindex locals words, implementation-defined options
12826:
12827: @table @i
12828: @item maximum number of locals in a definition:
12829: @cindex maximum number of locals in a definition
12830: @cindex locals, maximum number in a definition
12831: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
12832: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
12833: characters. The number of locals in a definition is bounded by the size
12834: of locals-buffer, which contains the names of the locals.
12835:
12836: @end table
12837:
12838:
12839: @c ---------------------------------------------------------------------
12840: @node locals-ambcond, , locals-idef, The optional Locals word set
12841: @subsection Ambiguous conditions
12842: @c ---------------------------------------------------------------------
12843: @cindex locals words, ambiguous conditions
12844: @cindex ambiguous conditions, locals words
12845:
12846: @table @i
12847: @item executing a named local in interpretation state:
12848: @cindex local in interpretation state
12849: @cindex Interpreting a compile-only word, for a local
12850: Locals have no interpretation semantics. If you try to perform the
12851: interpretation semantics, you will get a @code{-14 throw} somewhere
12852: (Interpreting a compile-only word). If you perform the compilation
12853: semantics, the locals access will be compiled (irrespective of state).
12854:
12855: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
12856: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
12857: @cindex @code{TO} on non-@code{VALUE}s and non-locals
12858: @cindex Invalid name argument, @code{TO}
12859: @code{-32 throw} (Invalid name argument)
12860:
12861: @end table
12862:
12863:
12864: @c =====================================================================
12865: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
12866: @section The optional Memory-Allocation word set
12867: @c =====================================================================
12868: @cindex system documentation, memory-allocation words
12869: @cindex memory-allocation words, system documentation
12870:
12871: @menu
12872: * memory-idef:: Implementation Defined Options
12873: @end menu
12874:
12875:
12876: @c ---------------------------------------------------------------------
12877: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
12878: @subsection Implementation Defined Options
12879: @c ---------------------------------------------------------------------
12880: @cindex implementation-defined options, memory-allocation words
12881: @cindex memory-allocation words, implementation-defined options
12882:
12883: @table @i
12884: @item values and meaning of @i{ior}:
12885: @cindex @i{ior} values and meaning
12886: The @i{ior}s returned by the file and memory allocation words are
12887: intended as throw codes. They typically are in the range
12888: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
12889: @i{ior}s is -512@minus{}@i{errno}.
12890:
12891: @end table
12892:
12893: @c =====================================================================
12894: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
12895: @section The optional Programming-Tools word set
12896: @c =====================================================================
12897: @cindex system documentation, programming-tools words
12898: @cindex programming-tools words, system documentation
12899:
12900: @menu
12901: * programming-idef:: Implementation Defined Options
12902: * programming-ambcond:: Ambiguous Conditions
12903: @end menu
12904:
12905:
12906: @c ---------------------------------------------------------------------
12907: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
12908: @subsection Implementation Defined Options
12909: @c ---------------------------------------------------------------------
12910: @cindex implementation-defined options, programming-tools words
12911: @cindex programming-tools words, implementation-defined options
12912:
12913: @table @i
12914: @item ending sequence for input following @code{;CODE} and @code{CODE}:
12915: @cindex @code{;CODE} ending sequence
12916: @cindex @code{CODE} ending sequence
12917: @code{END-CODE}
12918:
12919: @item manner of processing input following @code{;CODE} and @code{CODE}:
12920: @cindex @code{;CODE}, processing input
12921: @cindex @code{CODE}, processing input
12922: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
12923: the input is processed by the text interpreter, (starting) in interpret
12924: state.
12925:
12926: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
12927: @cindex @code{ASSEMBLER}, search order capability
12928: The ANS Forth search order word set.
12929:
12930: @item source and format of display by @code{SEE}:
12931: @cindex @code{SEE}, source and format of output
12932: The source for @code{see} is the intermediate code used by the inner
12933: interpreter. The current @code{see} tries to output Forth source code
12934: as well as possible.
12935:
12936: @end table
12937:
12938: @c ---------------------------------------------------------------------
12939: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
12940: @subsection Ambiguous conditions
12941: @c ---------------------------------------------------------------------
12942: @cindex programming-tools words, ambiguous conditions
12943: @cindex ambiguous conditions, programming-tools words
12944:
12945: @table @i
12946:
12947: @item deleting the compilation word list (@code{FORGET}):
12948: @cindex @code{FORGET}, deleting the compilation word list
12949: Not implemented (yet).
12950:
12951: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
12952: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
12953: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
12954: @cindex control-flow stack underflow
12955: This typically results in an @code{abort"} with a descriptive error
12956: message (may change into a @code{-22 throw} (Control structure mismatch)
12957: in the future). You may also get a memory access error. If you are
12958: unlucky, this ambiguous condition is not caught.
12959:
12960: @item @i{name} can't be found (@code{FORGET}):
12961: @cindex @code{FORGET}, @i{name} can't be found
12962: Not implemented (yet).
12963:
12964: @item @i{name} not defined via @code{CREATE}:
12965: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
12966: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
12967: the execution semantics of the last defined word no matter how it was
12968: defined.
12969:
12970: @item @code{POSTPONE} applied to @code{[IF]}:
12971: @cindex @code{POSTPONE} applied to @code{[IF]}
12972: @cindex @code{[IF]} and @code{POSTPONE}
12973: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
12974: equivalent to @code{[IF]}.
12975:
12976: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
12977: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
12978: Continue in the same state of conditional compilation in the next outer
12979: input source. Currently there is no warning to the user about this.
12980:
12981: @item removing a needed definition (@code{FORGET}):
12982: @cindex @code{FORGET}, removing a needed definition
12983: Not implemented (yet).
12984:
12985: @end table
12986:
12987:
12988: @c =====================================================================
12989: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
12990: @section The optional Search-Order word set
12991: @c =====================================================================
12992: @cindex system documentation, search-order words
12993: @cindex search-order words, system documentation
12994:
12995: @menu
12996: * search-idef:: Implementation Defined Options
12997: * search-ambcond:: Ambiguous Conditions
12998: @end menu
12999:
13000:
13001: @c ---------------------------------------------------------------------
13002: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
13003: @subsection Implementation Defined Options
13004: @c ---------------------------------------------------------------------
13005: @cindex implementation-defined options, search-order words
13006: @cindex search-order words, implementation-defined options
13007:
13008: @table @i
13009: @item maximum number of word lists in search order:
13010: @cindex maximum number of word lists in search order
13011: @cindex search order, maximum depth
13012: @code{s" wordlists" environment? drop .}. Currently 16.
13013:
13014: @item minimum search order:
13015: @cindex minimum search order
13016: @cindex search order, minimum
13017: @code{root root}.
13018:
13019: @end table
13020:
13021: @c ---------------------------------------------------------------------
13022: @node search-ambcond, , search-idef, The optional Search-Order word set
13023: @subsection Ambiguous conditions
13024: @c ---------------------------------------------------------------------
13025: @cindex search-order words, ambiguous conditions
13026: @cindex ambiguous conditions, search-order words
13027:
13028: @table @i
13029: @item changing the compilation word list (during compilation):
13030: @cindex changing the compilation word list (during compilation)
13031: @cindex compilation word list, change before definition ends
13032: The word is entered into the word list that was the compilation word list
13033: at the start of the definition. Any changes to the name field (e.g.,
13034: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
13035: are applied to the latest defined word (as reported by @code{last} or
13036: @code{lastxt}), if possible, irrespective of the compilation word list.
13037:
13038: @item search order empty (@code{previous}):
13039: @cindex @code{previous}, search order empty
13040: @cindex vocstack empty, @code{previous}
13041: @code{abort" Vocstack empty"}.
13042:
13043: @item too many word lists in search order (@code{also}):
13044: @cindex @code{also}, too many word lists in search order
13045: @cindex vocstack full, @code{also}
13046: @code{abort" Vocstack full"}.
13047:
13048: @end table
13049:
13050: @c ***************************************************************
13051: @node Standard vs Extensions, Model, ANS conformance, Top
13052: @chapter Should I use Gforth extensions?
13053: @cindex Gforth extensions
13054:
13055: As you read through the rest of this manual, you will see documentation
13056: for @i{Standard} words, and documentation for some appealing Gforth
13057: @i{extensions}. You might ask yourself the question: @i{``Should I
13058: restrict myself to the standard, or should I use the extensions?''}
13059:
13060: The answer depends on the goals you have for the program you are working
13061: on:
13062:
13063: @itemize @bullet
13064:
13065: @item Is it just for yourself or do you want to share it with others?
13066:
13067: @item
13068: If you want to share it, do the others all use Gforth?
13069:
13070: @item
13071: If it is just for yourself, do you want to restrict yourself to Gforth?
13072:
13073: @end itemize
13074:
13075: If restricting the program to Gforth is ok, then there is no reason not
13076: to use extensions. It is still a good idea to keep to the standard
13077: where it is easy, in case you want to reuse these parts in another
13078: program that you want to be portable.
13079:
13080: If you want to be able to port the program to other Forth systems, there
13081: are the following points to consider:
13082:
13083: @itemize @bullet
13084:
13085: @item
13086: Most Forth systems that are being maintained support the ANS Forth
13087: standard. So if your program complies with the standard, it will be
13088: portable among many systems.
13089:
13090: @item
13091: A number of the Gforth extensions can be implemented in ANS Forth using
13092: public-domain files provided in the @file{compat/} directory. These are
13093: mentioned in the text in passing. There is no reason not to use these
13094: extensions, your program will still be ANS Forth compliant; just include
13095: the appropriate compat files with your program.
13096:
13097: @item
13098: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
13099: analyse your program and determine what non-Standard words it relies
13100: upon. However, it does not check whether you use standard words in a
13101: non-standard way.
13102:
13103: @item
13104: Some techniques are not standardized by ANS Forth, and are hard or
13105: impossible to implement in a standard way, but can be implemented in
13106: most Forth systems easily, and usually in similar ways (e.g., accessing
13107: word headers). Forth has a rich historical precedent for programmers
13108: taking advantage of implementation-dependent features of their tools
13109: (for example, relying on a knowledge of the dictionary
13110: structure). Sometimes these techniques are necessary to extract every
13111: last bit of performance from the hardware, sometimes they are just a
13112: programming shorthand.
13113:
13114: @item
13115: Does using a Gforth extension save more work than the porting this part
13116: to other Forth systems (if any) will cost?
13117:
13118: @item
13119: Is the additional functionality worth the reduction in portability and
13120: the additional porting problems?
13121:
13122: @end itemize
13123:
13124: In order to perform these consideratios, you need to know what's
13125: standard and what's not. This manual generally states if something is
13126: non-standard, but the authoritative source is the standard document.
13127: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
13128: into the thought processes of the technical committee.
13129:
13130: Note also that portability between Forth systems is not the only
13131: portability issue; there is also the issue of portability between
13132: different platforms (processor/OS combinations).
13133:
13134: @c ***************************************************************
13135: @node Model, Integrating Gforth, Standard vs Extensions, Top
13136: @chapter Model
13137:
13138: This chapter has yet to be written. It will contain information, on
13139: which internal structures you can rely.
13140:
13141: @c ***************************************************************
13142: @node Integrating Gforth, Emacs and Gforth, Model, Top
13143: @chapter Integrating Gforth into C programs
13144:
13145: This is not yet implemented.
13146:
13147: Several people like to use Forth as scripting language for applications
13148: that are otherwise written in C, C++, or some other language.
13149:
13150: The Forth system ATLAST provides facilities for embedding it into
13151: applications; unfortunately it has several disadvantages: most
13152: importantly, it is not based on ANS Forth, and it is apparently dead
13153: (i.e., not developed further and not supported). The facilities
13154: provided by Gforth in this area are inspired by ATLAST's facilities, so
13155: making the switch should not be hard.
13156:
13157: We also tried to design the interface such that it can easily be
13158: implemented by other Forth systems, so that we may one day arrive at a
13159: standardized interface. Such a standard interface would allow you to
13160: replace the Forth system without having to rewrite C code.
13161:
13162: You embed the Gforth interpreter by linking with the library
13163: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
13164: global symbols in this library that belong to the interface, have the
13165: prefix @code{forth_}. (Global symbols that are used internally have the
13166: prefix @code{gforth_}).
13167:
13168: You can include the declarations of Forth types and the functions and
13169: variables of the interface with @code{#include <forth.h>}.
13170:
13171: Types.
13172:
13173: Variables.
13174:
13175: Data and FP Stack pointer. Area sizes.
13176:
13177: functions.
13178:
13179: forth_init(imagefile)
13180: forth_evaluate(string) exceptions?
13181: forth_goto(address) (or forth_execute(xt)?)
13182: forth_continue() (a corountining mechanism)
13183:
13184: Adding primitives.
13185:
13186: No checking.
13187:
13188: Signals?
13189:
13190: Accessing the Stacks
13191:
13192: @c ******************************************************************
13193: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
13194: @chapter Emacs and Gforth
13195: @cindex Emacs and Gforth
13196:
13197: @cindex @file{gforth.el}
13198: @cindex @file{forth.el}
13199: @cindex Rydqvist, Goran
13200: @cindex comment editing commands
13201: @cindex @code{\}, editing with Emacs
13202: @cindex debug tracer editing commands
13203: @cindex @code{~~}, removal with Emacs
13204: @cindex Forth mode in Emacs
13205: Gforth comes with @file{gforth.el}, an improved version of
13206: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
13207: improvements are:
13208:
13209: @itemize @bullet
13210: @item
13211: A better (but still not perfect) handling of indentation.
13212: @item
13213: Comment paragraph filling (@kbd{M-q})
13214: @item
13215: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
13216: @item
13217: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
13218: @item
13219: Support of the @code{info-lookup} feature for looking up the
13220: documentation of a word.
13221: @end itemize
13222:
13223: I left the stuff I do not use alone, even though some of it only makes
13224: sense for TILE. To get a description of these features, enter Forth mode
13225: and type @kbd{C-h m}.
13226:
13227: @cindex source location of error or debugging output in Emacs
13228: @cindex error output, finding the source location in Emacs
13229: @cindex debugging output, finding the source location in Emacs
13230: In addition, Gforth supports Emacs quite well: The source code locations
13231: given in error messages, debugging output (from @code{~~}) and failed
13232: assertion messages are in the right format for Emacs' compilation mode
13233: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
13234: Manual}) so the source location corresponding to an error or other
13235: message is only a few keystrokes away (@kbd{C-x `} for the next error,
13236: @kbd{C-c C-c} for the error under the cursor).
13237:
13238: @cindex @file{TAGS} file
13239: @cindex @file{etags.fs}
13240: @cindex viewing the source of a word in Emacs
13241: @cindex @code{require}, placement in files
13242: @cindex @code{include}, placement in files
13243: Also, if you @code{require} @file{etags.fs}, a new @file{TAGS} file will
13244: be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
13245: contains the definitions of all words defined afterwards. You can then
13246: find the source for a word using @kbd{M-.}. Note that emacs can use
13247: several tags files at the same time (e.g., one for the Gforth sources
13248: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
13249: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
13250: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
13251: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
13252: with @file{etags.fs}, you should avoid putting definitions both before
13253: and after @code{require} etc., otherwise you will see the same file
13254: visited several times by commands like @code{tags-search}.
13255:
13256: @cindex viewing the documentation of a word in Emacs
13257: @cindex context-sensitive help
13258: Moreover, for words documented in this manual, you can look up the
13259: glossary entry quickly by using @kbd{C-h TAB}
13260: (@code{info-lookup-symbol}, see @pxref{Documentation, ,Documentation
13261: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
13262: later and does not work for words containing @code{:}.
13263:
13264:
13265: @cindex @file{.emacs}
13266: To get all these benefits, add the following lines to your @file{.emacs}
13267: file:
13268:
13269: @example
13270: (autoload 'forth-mode "gforth.el")
13271: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
13272: @end example
13273:
13274: @c ******************************************************************
13275: @node Image Files, Engine, Emacs and Gforth, Top
13276: @chapter Image Files
13277: @cindex image file
13278: @cindex @file{.fi} files
13279: @cindex precompiled Forth code
13280: @cindex dictionary in persistent form
13281: @cindex persistent form of dictionary
13282:
13283: An image file is a file containing an image of the Forth dictionary,
13284: i.e., compiled Forth code and data residing in the dictionary. By
13285: convention, we use the extension @code{.fi} for image files.
13286:
13287: @menu
13288: * Image Licensing Issues:: Distribution terms for images.
13289: * Image File Background:: Why have image files?
13290: * Non-Relocatable Image Files:: don't always work.
13291: * Data-Relocatable Image Files:: are better.
13292: * Fully Relocatable Image Files:: better yet.
13293: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
13294: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
13295: * Modifying the Startup Sequence:: and turnkey applications.
13296: @end menu
13297:
13298: @node Image Licensing Issues, Image File Background, Image Files, Image Files
13299: @section Image Licensing Issues
13300: @cindex license for images
13301: @cindex image license
13302:
13303: An image created with @code{gforthmi} (@pxref{gforthmi}) or
13304: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
13305: original image; i.e., according to copyright law it is a derived work of
13306: the original image.
13307:
13308: Since Gforth is distributed under the GNU GPL, the newly created image
13309: falls under the GNU GPL, too. In particular, this means that if you
13310: distribute the image, you have to make all of the sources for the image
13311: available, including those you wrote. For details see @ref{License, ,
13312: GNU General Public License (Section 3)}.
13313:
13314: If you create an image with @code{cross} (@pxref{cross.fs}), the image
13315: contains only code compiled from the sources you gave it; if none of
13316: these sources is under the GPL, the terms discussed above do not apply
13317: to the image. However, if your image needs an engine (a gforth binary)
13318: that is under the GPL, you should make sure that you distribute both in
13319: a way that is at most a @emph{mere aggregation}, if you don't want the
13320: terms of the GPL to apply to the image.
13321:
13322: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
13323: @section Image File Background
13324: @cindex image file background
13325:
13326: Our Forth system consists not only of primitives, but also of
13327: definitions written in Forth. Since the Forth compiler itself belongs to
13328: those definitions, it is not possible to start the system with the
13329: primitives and the Forth source alone. Therefore we provide the Forth
13330: code as an image file in nearly executable form. When Gforth starts up,
13331: a C routine loads the image file into memory, optionally relocates the
13332: addresses, then sets up the memory (stacks etc.) according to
13333: information in the image file, and (finally) starts executing Forth
13334: code.
13335:
13336: The image file variants represent different compromises between the
13337: goals of making it easy to generate image files and making them
13338: portable.
13339:
13340: @cindex relocation at run-time
13341: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
13342: run-time. This avoids many of the complications discussed below (image
13343: files are data relocatable without further ado), but costs performance
13344: (one addition per memory access).
13345:
13346: @cindex relocation at load-time
13347: By contrast, the Gforth loader performs relocation at image load time. The
13348: loader also has to replace tokens that represent primitive calls with the
13349: appropriate code-field addresses (or code addresses in the case of
13350: direct threading).
13351:
13352: There are three kinds of image files, with different degrees of
13353: relocatability: non-relocatable, data-relocatable, and fully relocatable
13354: image files.
13355:
13356: @cindex image file loader
13357: @cindex relocating loader
13358: @cindex loader for image files
13359: These image file variants have several restrictions in common; they are
13360: caused by the design of the image file loader:
13361:
13362: @itemize @bullet
13363: @item
13364: There is only one segment; in particular, this means, that an image file
13365: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
13366: them). The contents of the stacks are not represented, either.
13367:
13368: @item
13369: The only kinds of relocation supported are: adding the same offset to
13370: all cells that represent data addresses; and replacing special tokens
13371: with code addresses or with pieces of machine code.
13372:
13373: If any complex computations involving addresses are performed, the
13374: results cannot be represented in the image file. Several applications that
13375: use such computations come to mind:
13376: @itemize @minus
13377: @item
13378: Hashing addresses (or data structures which contain addresses) for table
13379: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
13380: purpose, you will have no problem, because the hash tables are
13381: recomputed automatically when the system is started. If you use your own
13382: hash tables, you will have to do something similar.
13383:
13384: @item
13385: There's a cute implementation of doubly-linked lists that uses
13386: @code{XOR}ed addresses. You could represent such lists as singly-linked
13387: in the image file, and restore the doubly-linked representation on
13388: startup.@footnote{In my opinion, though, you should think thrice before
13389: using a doubly-linked list (whatever implementation).}
13390:
13391: @item
13392: The code addresses of run-time routines like @code{docol:} cannot be
13393: represented in the image file (because their tokens would be replaced by
13394: machine code in direct threaded implementations). As a workaround,
13395: compute these addresses at run-time with @code{>code-address} from the
13396: executions tokens of appropriate words (see the definitions of
13397: @code{docol:} and friends in @file{kernel.fs}).
13398:
13399: @item
13400: On many architectures addresses are represented in machine code in some
13401: shifted or mangled form. You cannot put @code{CODE} words that contain
13402: absolute addresses in this form in a relocatable image file. Workarounds
13403: are representing the address in some relative form (e.g., relative to
13404: the CFA, which is present in some register), or loading the address from
13405: a place where it is stored in a non-mangled form.
13406: @end itemize
13407: @end itemize
13408:
13409: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
13410: @section Non-Relocatable Image Files
13411: @cindex non-relocatable image files
13412: @cindex image file, non-relocatable
13413:
13414: These files are simple memory dumps of the dictionary. They are specific
13415: to the executable (i.e., @file{gforth} file) they were created
13416: with. What's worse, they are specific to the place on which the
13417: dictionary resided when the image was created. Now, there is no
13418: guarantee that the dictionary will reside at the same place the next
13419: time you start Gforth, so there's no guarantee that a non-relocatable
13420: image will work the next time (Gforth will complain instead of crashing,
13421: though).
13422:
13423: You can create a non-relocatable image file with
13424:
13425:
13426: doc-savesystem
13427:
13428:
13429: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
13430: @section Data-Relocatable Image Files
13431: @cindex data-relocatable image files
13432: @cindex image file, data-relocatable
13433:
13434: These files contain relocatable data addresses, but fixed code addresses
13435: (instead of tokens). They are specific to the executable (i.e.,
13436: @file{gforth} file) they were created with. For direct threading on some
13437: architectures (e.g., the i386), data-relocatable images do not work. You
13438: get a data-relocatable image, if you use @file{gforthmi} with a
13439: Gforth binary that is not doubly indirect threaded (@pxref{Fully
13440: Relocatable Image Files}).
13441:
13442: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
13443: @section Fully Relocatable Image Files
13444: @cindex fully relocatable image files
13445: @cindex image file, fully relocatable
13446:
13447: @cindex @file{kern*.fi}, relocatability
13448: @cindex @file{gforth.fi}, relocatability
13449: These image files have relocatable data addresses, and tokens for code
13450: addresses. They can be used with different binaries (e.g., with and
13451: without debugging) on the same machine, and even across machines with
13452: the same data formats (byte order, cell size, floating point
13453: format). However, they are usually specific to the version of Gforth
13454: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
13455: are fully relocatable.
13456:
13457: There are two ways to create a fully relocatable image file:
13458:
13459: @menu
13460: * gforthmi:: The normal way
13461: * cross.fs:: The hard way
13462: @end menu
13463:
13464: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
13465: @subsection @file{gforthmi}
13466: @cindex @file{comp-i.fs}
13467: @cindex @file{gforthmi}
13468:
13469: You will usually use @file{gforthmi}. If you want to create an
13470: image @i{file} that contains everything you would load by invoking
13471: Gforth with @code{gforth @i{options}}, you simply say:
13472: @example
13473: gforthmi @i{file} @i{options}
13474: @end example
13475:
13476: E.g., if you want to create an image @file{asm.fi} that has the file
13477: @file{asm.fs} loaded in addition to the usual stuff, you could do it
13478: like this:
13479:
13480: @example
13481: gforthmi asm.fi asm.fs
13482: @end example
13483:
13484: @file{gforthmi} is implemented as a sh script and works like this: It
13485: produces two non-relocatable images for different addresses and then
13486: compares them. Its output reflects this: first you see the output (if
13487: any) of the two Gforth invocations that produce the non-relocatable image
13488: files, then you see the output of the comparing program: It displays the
13489: offset used for data addresses and the offset used for code addresses;
13490: moreover, for each cell that cannot be represented correctly in the
13491: image files, it displays a line like this:
13492:
13493: @example
13494: 78DC BFFFFA50 BFFFFA40
13495: @end example
13496:
13497: This means that at offset $78dc from @code{forthstart}, one input image
13498: contains $bffffa50, and the other contains $bffffa40. Since these cells
13499: cannot be represented correctly in the output image, you should examine
13500: these places in the dictionary and verify that these cells are dead
13501: (i.e., not read before they are written).
13502:
13503: @cindex --application, @code{gforthmi} option
13504: If you insert the option @code{--application} in front of the image file
13505: name, you will get an image that uses the @code{--appl-image} option
13506: instead of the @code{--image-file} option (@pxref{Invoking
13507: Gforth}). When you execute such an image on Unix (by typing the image
13508: name as command), the Gforth engine will pass all options to the image
13509: instead of trying to interpret them as engine options.
13510:
13511: If you type @file{gforthmi} with no arguments, it prints some usage
13512: instructions.
13513:
13514: @cindex @code{savesystem} during @file{gforthmi}
13515: @cindex @code{bye} during @file{gforthmi}
13516: @cindex doubly indirect threaded code
13517: @cindex environment variables
13518: @cindex @code{GFORTHD} -- environment variable
13519: @cindex @code{GFORTH} -- environment variable
13520: @cindex @code{gforth-ditc}
13521: There are a few wrinkles: After processing the passed @i{options}, the
13522: words @code{savesystem} and @code{bye} must be visible. A special doubly
13523: indirect threaded version of the @file{gforth} executable is used for
13524: creating the non-relocatable images; you can pass the exact filename of
13525: this executable through the environment variable @code{GFORTHD}
13526: (default: @file{gforth-ditc}); if you pass a version that is not doubly
13527: indirect threaded, you will not get a fully relocatable image, but a
13528: data-relocatable image (because there is no code address offset). The
13529: normal @file{gforth} executable is used for creating the relocatable
13530: image; you can pass the exact filename of this executable through the
13531: environment variable @code{GFORTH}.
13532:
13533: @node cross.fs, , gforthmi, Fully Relocatable Image Files
13534: @subsection @file{cross.fs}
13535: @cindex @file{cross.fs}
13536: @cindex cross-compiler
13537: @cindex metacompiler
13538: @cindex target compiler
13539:
13540: You can also use @code{cross}, a batch compiler that accepts a Forth-like
13541: programming language (@pxref{Cross Compiler}).
13542:
13543: @code{cross} allows you to create image files for machines with
13544: different data sizes and data formats than the one used for generating
13545: the image file. You can also use it to create an application image that
13546: does not contain a Forth compiler. These features are bought with
13547: restrictions and inconveniences in programming. E.g., addresses have to
13548: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
13549: order to make the code relocatable.
13550:
13551:
13552: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
13553: @section Stack and Dictionary Sizes
13554: @cindex image file, stack and dictionary sizes
13555: @cindex dictionary size default
13556: @cindex stack size default
13557:
13558: If you invoke Gforth with a command line flag for the size
13559: (@pxref{Invoking Gforth}), the size you specify is stored in the
13560: dictionary. If you save the dictionary with @code{savesystem} or create
13561: an image with @file{gforthmi}, this size will become the default
13562: for the resulting image file. E.g., the following will create a
13563: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
13564:
13565: @example
13566: gforthmi gforth.fi -m 1M
13567: @end example
13568:
13569: In other words, if you want to set the default size for the dictionary
13570: and the stacks of an image, just invoke @file{gforthmi} with the
13571: appropriate options when creating the image.
13572:
13573: @cindex stack size, cache-friendly
13574: Note: For cache-friendly behaviour (i.e., good performance), you should
13575: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
13576: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
13577: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
13578:
13579: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
13580: @section Running Image Files
13581: @cindex running image files
13582: @cindex invoking image files
13583: @cindex image file invocation
13584:
13585: @cindex -i, invoke image file
13586: @cindex --image file, invoke image file
13587: You can invoke Gforth with an image file @i{image} instead of the
13588: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
13589: @example
13590: gforth -i @i{image}
13591: @end example
13592:
13593: @cindex executable image file
13594: @cindex image file, executable
13595: If your operating system supports starting scripts with a line of the
13596: form @code{#! ...}, you just have to type the image file name to start
13597: Gforth with this image file (note that the file extension @code{.fi} is
13598: just a convention). I.e., to run Gforth with the image file @i{image},
13599: you can just type @i{image} instead of @code{gforth -i @i{image}}.
13600: This works because every @code{.fi} file starts with a line of this
13601: format:
13602:
13603: @example
13604: #! /usr/local/bin/gforth-0.4.0 -i
13605: @end example
13606:
13607: The file and pathname for the Gforth engine specified on this line is
13608: the specific Gforth executable that it was built against; i.e. the value
13609: of the environment variable @code{GFORTH} at the time that
13610: @file{gforthmi} was executed.
13611:
13612: You can make use of the same shell capability to make a Forth source
13613: file into an executable. For example, if you place this text in a file:
13614:
13615: @example
13616: #! /usr/local/bin/gforth
13617:
13618: ." Hello, world" CR
13619: bye
13620: @end example
13621:
13622: @noindent
13623: and then make the file executable (chmod +x in Unix), you can run it
13624: directly from the command line. The sequence @code{#!} is used in two
13625: ways; firstly, it is recognised as a ``magic sequence'' by the operating
13626: system@footnote{The Unix kernel actually recognises two types of files:
13627: executable files and files of data, where the data is processed by an
13628: interpreter that is specified on the ``interpreter line'' -- the first
13629: line of the file, starting with the sequence #!. There may be a small
13630: limit (e.g., 32) on the number of characters that may be specified on
13631: the interpreter line.} secondly it is treated as a comment character by
13632: Gforth. Because of the second usage, a space is required between
13633: @code{#!} and the path to the executable.
13634:
13635: The disadvantage of this latter technique, compared with using
13636: @file{gforthmi}, is that it is slower; the Forth source code is compiled
13637: on-the-fly, each time the program is invoked.
13638:
13639:
13640: doc-#!
13641:
13642:
13643: @node Modifying the Startup Sequence, , Running Image Files, Image Files
13644: @section Modifying the Startup Sequence
13645: @cindex startup sequence for image file
13646: @cindex image file initialization sequence
13647: @cindex initialization sequence of image file
13648:
13649: You can add your own initialization to the startup sequence through the
13650: deferred word @code{'cold}. @code{'cold} is invoked just before the
13651: image-specific command line processing (by default, loading files and
13652: evaluating (@code{-e}) strings) starts.
13653:
13654: A sequence for adding your initialization usually looks like this:
13655:
13656: @example
13657: :noname
13658: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
13659: ... \ your stuff
13660: ; IS 'cold
13661: @end example
13662:
13663: @cindex turnkey image files
13664: @cindex image file, turnkey applications
13665: You can make a turnkey image by letting @code{'cold} execute a word
13666: (your turnkey application) that never returns; instead, it exits Gforth
13667: via @code{bye} or @code{throw}.
13668:
13669: @cindex command-line arguments, access
13670: @cindex arguments on the command line, access
13671: You can access the (image-specific) command-line arguments through the
13672: variables @code{argc} and @code{argv}. @code{arg} provides convenient
13673: access to @code{argv}.
13674:
13675: If @code{'cold} exits normally, Gforth processes the command-line
13676: arguments as files to be loaded and strings to be evaluated. Therefore,
13677: @code{'cold} should remove the arguments it has used in this case.
13678:
13679:
13680:
13681: doc-'cold
13682: doc-argc
13683: doc-argv
13684: doc-arg
13685:
13686:
13687:
13688: @c ******************************************************************
13689: @node Engine, Binding to System Library, Image Files, Top
13690: @chapter Engine
13691: @cindex engine
13692: @cindex virtual machine
13693:
13694: Reading this chapter is not necessary for programming with Gforth. It
13695: may be helpful for finding your way in the Gforth sources.
13696:
13697: The ideas in this section have also been published in Bernd Paysan,
13698: @cite{ANS fig/GNU/??? Forth} (in German), Forth-Tagung '93 and M. Anton
13699: Ertl, @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
13700: Portable Forth Engine}}, EuroForth '93.
13701:
13702: @menu
13703: * Portability::
13704: * Threading::
13705: * Primitives::
13706: * Performance::
13707: @end menu
13708:
13709: @node Portability, Threading, Engine, Engine
13710: @section Portability
13711: @cindex engine portability
13712:
13713: An important goal of the Gforth Project is availability across a wide
13714: range of personal machines. fig-Forth, and, to a lesser extent, F83,
13715: achieved this goal by manually coding the engine in assembly language
13716: for several then-popular processors. This approach is very
13717: labor-intensive and the results are short-lived due to progress in
13718: computer architecture.
13719:
13720: @cindex C, using C for the engine
13721: Others have avoided this problem by coding in C, e.g., Mitch Bradley
13722: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
13723: particularly popular for UNIX-based Forths due to the large variety of
13724: architectures of UNIX machines. Unfortunately an implementation in C
13725: does not mix well with the goals of efficiency and with using
13726: traditional techniques: Indirect or direct threading cannot be expressed
13727: in C, and switch threading, the fastest technique available in C, is
13728: significantly slower. Another problem with C is that it is very
13729: cumbersome to express double integer arithmetic.
13730:
13731: @cindex GNU C for the engine
13732: @cindex long long
13733: Fortunately, there is a portable language that does not have these
13734: limitations: GNU C, the version of C processed by the GNU C compiler
13735: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
13736: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
13737: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
13738: threading possible, its @code{long long} type (@pxref{Long Long, ,
13739: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
13740: double numbers@footnote{Unfortunately, long longs are not implemented
13741: properly on all machines (e.g., on alpha-osf1, long longs are only 64
13742: bits, the same size as longs (and pointers), but they should be twice as
13743: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
13744: C Manual}). So, we had to implement doubles in C after all. Still, on
13745: most machines we can use long longs and achieve better performance than
13746: with the emulation package.}. GNU C is available for free on all
13747: important (and many unimportant) UNIX machines, VMS, 80386s running
13748: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
13749: on all these machines.
13750:
13751: Writing in a portable language has the reputation of producing code that
13752: is slower than assembly. For our Forth engine we repeatedly looked at
13753: the code produced by the compiler and eliminated most compiler-induced
13754: inefficiencies by appropriate changes in the source code.
13755:
13756: @cindex explicit register declarations
13757: @cindex --enable-force-reg, configuration flag
13758: @cindex -DFORCE_REG
13759: However, register allocation cannot be portably influenced by the
13760: programmer, leading to some inefficiencies on register-starved
13761: machines. We use explicit register declarations (@pxref{Explicit Reg
13762: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
13763: improve the speed on some machines. They are turned on by using the
13764: configuration flag @code{--enable-force-reg} (@code{gcc} switch
13765: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
13766: machine, but also on the compiler version: On some machines some
13767: compiler versions produce incorrect code when certain explicit register
13768: declarations are used. So by default @code{-DFORCE_REG} is not used.
13769:
13770: @node Threading, Primitives, Portability, Engine
13771: @section Threading
13772: @cindex inner interpreter implementation
13773: @cindex threaded code implementation
13774:
13775: @cindex labels as values
13776: GNU C's labels as values extension (available since @code{gcc-2.0},
13777: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
13778: makes it possible to take the address of @i{label} by writing
13779: @code{&&@i{label}}. This address can then be used in a statement like
13780: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
13781: @code{goto x}.
13782:
13783: @cindex @code{NEXT}, indirect threaded
13784: @cindex indirect threaded inner interpreter
13785: @cindex inner interpreter, indirect threaded
13786: With this feature an indirect threaded @code{NEXT} looks like:
13787: @example
13788: cfa = *ip++;
13789: ca = *cfa;
13790: goto *ca;
13791: @end example
13792: @cindex instruction pointer
13793: For those unfamiliar with the names: @code{ip} is the Forth instruction
13794: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
13795: execution token and points to the code field of the next word to be
13796: executed; The @code{ca} (code address) fetched from there points to some
13797: executable code, e.g., a primitive or the colon definition handler
13798: @code{docol}.
13799:
13800: @cindex @code{NEXT}, direct threaded
13801: @cindex direct threaded inner interpreter
13802: @cindex inner interpreter, direct threaded
13803: Direct threading is even simpler:
13804: @example
13805: ca = *ip++;
13806: goto *ca;
13807: @end example
13808:
13809: Of course we have packaged the whole thing neatly in macros called
13810: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
13811:
13812: @menu
13813: * Scheduling::
13814: * Direct or Indirect Threaded?::
13815: * DOES>::
13816: @end menu
13817:
13818: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
13819: @subsection Scheduling
13820: @cindex inner interpreter optimization
13821:
13822: There is a little complication: Pipelined and superscalar processors,
13823: i.e., RISC and some modern CISC machines can process independent
13824: instructions while waiting for the results of an instruction. The
13825: compiler usually reorders (schedules) the instructions in a way that
13826: achieves good usage of these delay slots. However, on our first tries
13827: the compiler did not do well on scheduling primitives. E.g., for
13828: @code{+} implemented as
13829: @example
13830: n=sp[0]+sp[1];
13831: sp++;
13832: sp[0]=n;
13833: NEXT;
13834: @end example
13835: the @code{NEXT} comes strictly after the other code, i.e., there is nearly no
13836: scheduling. After a little thought the problem becomes clear: The
13837: compiler cannot know that @code{sp} and @code{ip} point to different
13838: addresses (and the version of @code{gcc} we used would not know it even
13839: if it was possible), so it could not move the load of the cfa above the
13840: store to the TOS. Indeed the pointers could be the same, if code on or
13841: very near the top of stack were executed. In the interest of speed we
13842: chose to forbid this probably unused ``feature'' and helped the compiler
13843: in scheduling: @code{NEXT} is divided into the loading part (@code{NEXT_P1})
13844: and the goto part (@code{NEXT_P2}). @code{+} now looks like:
13845: @example
13846: n=sp[0]+sp[1];
13847: sp++;
13848: NEXT_P1;
13849: sp[0]=n;
13850: NEXT_P2;
13851: @end example
13852: This can be scheduled optimally by the compiler.
13853:
13854: This division can be turned off with the switch @code{-DCISC_NEXT}. This
13855: switch is on by default on machines that do not profit from scheduling
13856: (e.g., the 80386), in order to preserve registers.
13857:
13858: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
13859: @subsection Direct or Indirect Threaded?
13860: @cindex threading, direct or indirect?
13861:
13862: @cindex -DDIRECT_THREADED
13863: Both! After packaging the nasty details in macro definitions we
13864: realized that we could switch between direct and indirect threading by
13865: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
13866: defining a few machine-specific macros for the direct-threading case.
13867: On the Forth level we also offer access words that hide the
13868: differences between the threading methods (@pxref{Threading Words}).
13869:
13870: Indirect threading is implemented completely machine-independently.
13871: Direct threading needs routines for creating jumps to the executable
13872: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
13873: machine-dependent, but they do not amount to many source lines. Therefore,
13874: even porting direct threading to a new machine requires little effort.
13875:
13876: @cindex --enable-indirect-threaded, configuration flag
13877: @cindex --enable-direct-threaded, configuration flag
13878: The default threading method is machine-dependent. You can enforce a
13879: specific threading method when building Gforth with the configuration
13880: flag @code{--enable-direct-threaded} or
13881: @code{--enable-indirect-threaded}. Note that direct threading is not
13882: supported on all machines.
13883:
13884: @node DOES>, , Direct or Indirect Threaded?, Threading
13885: @subsection DOES>
13886: @cindex @code{DOES>} implementation
13887:
13888: @cindex @code{dodoes} routine
13889: @cindex @code{DOES>}-code
13890: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
13891: the chunk of code executed by every word defined by a
13892: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
13893: the Forth code to be executed, i.e. the code after the
13894: @code{DOES>} (the @code{DOES>}-code)? There are two solutions:
13895:
13896: In fig-Forth the code field points directly to the @code{dodoes} and the
13897: @code{DOES>}-code address is stored in the cell after the code address (i.e. at
13898: @code{@i{CFA} cell+}). It may seem that this solution is illegal in
13899: the Forth-79 and all later standards, because in fig-Forth this address
13900: lies in the body (which is illegal in these standards). However, by
13901: making the code field larger for all words this solution becomes legal
13902: again. We use this approach for the indirect threaded version and for
13903: direct threading on some machines. Leaving a cell unused in most words
13904: is a bit wasteful, but on the machines we are targeting this is hardly a
13905: problem. The other reason for having a code field size of two cells is
13906: to avoid having different image files for direct and indirect threaded
13907: systems (direct threaded systems require two-cell code fields on many
13908: machines).
13909:
13910: @cindex @code{DOES>}-handler
13911: The other approach is that the code field points or jumps to the cell
13912: after @code{DOES>}. In this variant there is a jump to @code{dodoes} at
13913: this address (the @code{DOES>}-handler). @code{dodoes} can then get the
13914: @code{DOES>}-code address by computing the code address, i.e., the address of
13915: the jump to @code{dodoes}, and add the length of that jump field. A variant of
13916: this is to have a call to @code{dodoes} after the @code{DOES>}; then the
13917: return address (which can be found in the return register on RISCs) is
13918: the @code{DOES>}-code address. Since the two cells available in the code field
13919: are used up by the jump to the code address in direct threading on many
13920: architectures, we use this approach for direct threading on these
13921: architectures. We did not want to add another cell to the code field.
13922:
13923: @node Primitives, Performance, Threading, Engine
13924: @section Primitives
13925: @cindex primitives, implementation
13926: @cindex virtual machine instructions, implementation
13927:
13928: @menu
13929: * Automatic Generation::
13930: * TOS Optimization::
13931: * Produced code::
13932: @end menu
13933:
13934: @node Automatic Generation, TOS Optimization, Primitives, Primitives
13935: @subsection Automatic Generation
13936: @cindex primitives, automatic generation
13937:
13938: @cindex @file{prims2x.fs}
13939: Since the primitives are implemented in a portable language, there is no
13940: longer any need to minimize the number of primitives. On the contrary,
13941: having many primitives has an advantage: speed. In order to reduce the
13942: number of errors in primitives and to make programming them easier, we
13943: provide a tool, the primitive generator (@file{prims2x.fs}), that
13944: automatically generates most (and sometimes all) of the C code for a
13945: primitive from the stack effect notation. The source for a primitive
13946: has the following form:
13947:
13948: @cindex primitive source format
13949: @format
13950: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
13951: [@code{""}@i{glossary entry}@code{""}]
13952: @i{C code}
13953: [@code{:}
13954: @i{Forth code}]
13955: @end format
13956:
13957: The items in brackets are optional. The category and glossary fields
13958: are there for generating the documentation, the Forth code is there
13959: for manual implementations on machines without GNU C. E.g., the source
13960: for the primitive @code{+} is:
13961: @example
13962: + ( n1 n2 -- n ) core plus
13963: n = n1+n2;
13964: @end example
13965:
13966: This looks like a specification, but in fact @code{n = n1+n2} is C
13967: code. Our primitive generation tool extracts a lot of information from
13968: the stack effect notations@footnote{We use a one-stack notation, even
13969: though we have separate data and floating-point stacks; The separate
13970: notation can be generated easily from the unified notation.}: The number
13971: of items popped from and pushed on the stack, their type, and by what
13972: name they are referred to in the C code. It then generates a C code
13973: prelude and postlude for each primitive. The final C code for @code{+}
13974: looks like this:
13975:
13976: @example
13977: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
13978: /* */ /* documentation */
13979: @{
13980: DEF_CA /* definition of variable ca (indirect threading) */
13981: Cell n1; /* definitions of variables */
13982: Cell n2;
13983: Cell n;
13984: n1 = (Cell) sp[1]; /* input */
13985: n2 = (Cell) TOS;
13986: sp += 1; /* stack adjustment */
13987: NAME("+") /* debugging output (with -DDEBUG) */
13988: @{
13989: n = n1+n2; /* C code taken from the source */
13990: @}
13991: NEXT_P1; /* NEXT part 1 */
13992: TOS = (Cell)n; /* output */
13993: NEXT_P2; /* NEXT part 2 */
13994: @}
13995: @end example
13996:
13997: This looks long and inefficient, but the GNU C compiler optimizes quite
13998: well and produces optimal code for @code{+} on, e.g., the R3000 and the
13999: HP RISC machines: Defining the @code{n}s does not produce any code, and
14000: using them as intermediate storage also adds no cost.
14001:
14002: There are also other optimizations that are not illustrated by this
14003: example: assignments between simple variables are usually for free (copy
14004: propagation). If one of the stack items is not used by the primitive
14005: (e.g. in @code{drop}), the compiler eliminates the load from the stack
14006: (dead code elimination). On the other hand, there are some things that
14007: the compiler does not do, therefore they are performed by
14008: @file{prims2x.fs}: The compiler does not optimize code away that stores
14009: a stack item to the place where it just came from (e.g., @code{over}).
14010:
14011: While programming a primitive is usually easy, there are a few cases
14012: where the programmer has to take the actions of the generator into
14013: account, most notably @code{?dup}, but also words that do not (always)
14014: fall through to @code{NEXT}.
14015:
14016: @node TOS Optimization, Produced code, Automatic Generation, Primitives
14017: @subsection TOS Optimization
14018: @cindex TOS optimization for primitives
14019: @cindex primitives, keeping the TOS in a register
14020:
14021: An important optimization for stack machine emulators, e.g., Forth
14022: engines, is keeping one or more of the top stack items in
14023: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
14024: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
14025: @itemize @bullet
14026: @item
14027: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
14028: due to fewer loads from and stores to the stack.
14029: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
14030: @i{y<n}, due to additional moves between registers.
14031: @end itemize
14032:
14033: @cindex -DUSE_TOS
14034: @cindex -DUSE_NO_TOS
14035: In particular, keeping one item in a register is never a disadvantage,
14036: if there are enough registers. Keeping two items in registers is a
14037: disadvantage for frequent words like @code{?branch}, constants,
14038: variables, literals and @code{i}. Therefore our generator only produces
14039: code that keeps zero or one items in registers. The generated C code
14040: covers both cases; the selection between these alternatives is made at
14041: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
14042: code for @code{+} is just a simple variable name in the one-item case,
14043: otherwise it is a macro that expands into @code{sp[0]}. Note that the
14044: GNU C compiler tries to keep simple variables like @code{TOS} in
14045: registers, and it usually succeeds, if there are enough registers.
14046:
14047: @cindex -DUSE_FTOS
14048: @cindex -DUSE_NO_FTOS
14049: The primitive generator performs the TOS optimization for the
14050: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
14051: operations the benefit of this optimization is even larger:
14052: floating-point operations take quite long on most processors, but can be
14053: performed in parallel with other operations as long as their results are
14054: not used. If the FP-TOS is kept in a register, this works. If
14055: it is kept on the stack, i.e., in memory, the store into memory has to
14056: wait for the result of the floating-point operation, lengthening the
14057: execution time of the primitive considerably.
14058:
14059: The TOS optimization makes the automatic generation of primitives a
14060: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
14061: @code{TOS} is not sufficient. There are some special cases to
14062: consider:
14063: @itemize @bullet
14064: @item In the case of @code{dup ( w -- w w )} the generator must not
14065: eliminate the store to the original location of the item on the stack,
14066: if the TOS optimization is turned on.
14067: @item Primitives with stack effects of the form @code{--}
14068: @i{out1}...@i{outy} must store the TOS to the stack at the start.
14069: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
14070: must load the TOS from the stack at the end. But for the null stack
14071: effect @code{--} no stores or loads should be generated.
14072: @end itemize
14073:
14074: @node Produced code, , TOS Optimization, Primitives
14075: @subsection Produced code
14076: @cindex primitives, assembly code listing
14077:
14078: @cindex @file{engine.s}
14079: To see what assembly code is produced for the primitives on your machine
14080: with your compiler and your flag settings, type @code{make engine.s} and
14081: look at the resulting file @file{engine.s}.
14082:
14083: @node Performance, , Primitives, Engine
14084: @section Performance
14085: @cindex performance of some Forth interpreters
14086: @cindex engine performance
14087: @cindex benchmarking Forth systems
14088: @cindex Gforth performance
14089:
14090: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
14091: impossible to write a significantly faster engine.
14092:
14093: On register-starved machines like the 386 architecture processors
14094: improvements are possible, because @code{gcc} does not utilize the
14095: registers as well as a human, even with explicit register declarations;
14096: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
14097: and hand-tuned it for the 486; this system is 1.19 times faster on the
14098: Sieve benchmark on a 486DX2/66 than Gforth compiled with
14099: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
14100: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
14101: registers fit in real registers (and we can even afford to use the TOS
14102: optimization), resulting in a speedup of 1.14 on the sieve over the
14103: earlier results.
14104:
14105: @cindex Win32Forth performance
14106: @cindex NT Forth performance
14107: @cindex eforth performance
14108: @cindex ThisForth performance
14109: @cindex PFE performance
14110: @cindex TILE performance
14111: The potential advantage of assembly language implementations
14112: is not necessarily realized in complete Forth systems: We compared
14113: Gforth-0.4.9 (direct threaded, compiled with @code{gcc-2.95.1} and
14114: @code{-DFORCE_REG}) with Win32Forth 1.2093, LMI's NT Forth (Beta, May
14115: 1994) and Eforth (with and without peephole (aka pinhole) optimization
14116: of the threaded code); all these systems were written in assembly
14117: language. We also compared Gforth with three systems written in C:
14118: PFE-0.9.14 (compiled with @code{gcc-2.6.3} with the default
14119: configuration for Linux: @code{-O2 -fomit-frame-pointer -DUSE_REGS
14120: -DUNROLL_NEXT}), ThisForth Beta (compiled with @code{gcc-2.6.3 -O3
14121: -fomit-frame-pointer}; ThisForth employs peephole optimization of the
14122: threaded code) and TILE (compiled with @code{make opt}). We benchmarked
14123: Gforth, PFE, ThisForth and TILE on a 486DX2/66 under Linux. Kenneth
14124: O'Heskin kindly provided the results for Win32Forth and NT Forth on a
14125: 486DX2/66 with similar memory performance under Windows NT. Marcel
14126: Hendrix ported Eforth to Linux, then extended it to run the benchmarks,
14127: added the peephole optimizer, ran the benchmarks and reported the
14128: results.
14129:
14130: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
14131: matrix multiplication come from the Stanford integer benchmarks and have
14132: been translated into Forth by Martin Fraeman; we used the versions
14133: included in the TILE Forth package, but with bigger data set sizes; and
14134: a recursive Fibonacci number computation for benchmarking calling
14135: performance. The following table shows the time taken for the benchmarks
14136: scaled by the time taken by Gforth (in other words, it shows the speedup
14137: factor that Gforth achieved over the other systems).
14138:
14139: @example
14140: relative Win32- NT eforth This-
14141: time Gforth Forth Forth eforth +opt PFE Forth TILE
14142: sieve 1.00 1.58 1.30 1.58 0.97 1.80 3.63 9.79
14143: bubble 1.00 1.55 1.67 1.75 1.04 1.78 4.59
14144: matmul 1.00 1.67 1.53 1.66 0.84 1.79 4.63
14145: fib 1.00 1.75 1.53 1.40 0.99 1.99 3.43 4.93
14146: @end example
14147:
14148: You may be quite surprised by the good performance of Gforth when
14149: compared with systems written in assembly language. One important reason
14150: for the disappointing performance of these other systems is probably
14151: that they are not written optimally for the 486 (e.g., they use the
14152: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
14153: but costly method for relocating the Forth image: like @code{cforth}, it
14154: computes the actual addresses at run time, resulting in two address
14155: computations per @code{NEXT} (@pxref{Image File Background}).
14156:
14157: Only Eforth with the peephole optimizer performs comparable to
14158: Gforth. The speedups achieved with peephole optimization of threaded
14159: code are quite remarkable. Adding a peephole optimizer to Gforth should
14160: cause similar speedups.
14161:
14162: The speedup of Gforth over PFE, ThisForth and TILE can be easily
14163: explained with the self-imposed restriction of the latter systems to
14164: standard C, which makes efficient threading impossible (however, the
14165: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
14166: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
14167: Moreover, current C compilers have a hard time optimizing other aspects
14168: of the ThisForth and the TILE source.
14169:
14170: The performance of Gforth on 386 architecture processors varies widely
14171: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
14172: allocate any of the virtual machine registers into real machine
14173: registers by itself and would not work correctly with explicit register
14174: declarations, giving a 1.5 times slower engine (on a 486DX2/66 running
14175: the Sieve) than the one measured above.
14176:
14177: Note that there have been several releases of Win32Forth since the
14178: release presented here, so the results presented above may have little
14179: predictive value for the performance of Win32Forth today (results for
14180: the current release on an i486DX2/66 are welcome).
14181:
14182: @cindex @file{Benchres}
14183: In
14184: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
14185: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
14186: Maierhofer (presented at EuroForth '95), an indirect threaded version of
14187: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
14188: several native code systems; that version of Gforth is slower on a 486
14189: than the direct threaded version used here. You can find a newer version
14190: of these measurements at
14191: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
14192: find numbers for Gforth on various machines in @file{Benchres}.
14193:
14194: @c ******************************************************************
14195: @node Binding to System Library, Cross Compiler, Engine, Top
14196: @chapter Binding to System Library
14197:
14198: @node Cross Compiler, Bugs, Binding to System Library, Top
14199: @chapter Cross Compiler
14200: @cindex @file{cross.fs}
14201: @cindex cross-compiler
14202: @cindex metacompiler
14203: @cindex target compiler
14204:
14205: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
14206: mostly written in Forth, including crucial parts like the outer
14207: interpreter and compiler, it needs compiled Forth code to get
14208: started. The cross compiler allows to create new images for other
14209: architectures, even running under another Forth system.
14210:
14211: @menu
14212: * Using the Cross Compiler::
14213: * How the Cross Compiler Works::
14214: @end menu
14215:
14216: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
14217: @section Using the Cross Compiler
14218:
14219: The cross compiler uses a language that resembles Forth, but isn't. The
14220: main difference is that you can execute Forth code after definition,
14221: while you usually can't execute the code compiled by cross, because the
14222: code you are compiling is typically for a different computer than the
14223: one you are compiling on.
14224:
14225: The Makefile is already set up to allow you to create kernels for new
14226: architectures with a simple make command. The generic kernels using the
14227: GCC compiled virtual machine are created in the normal build process
14228: with @code{make}. To create a embedded Gforth executable for e.g. the
14229: 8086 processor (running on a DOS machine), type
14230:
14231: @example
14232: make kernl-8086.fi
14233: @end example
14234:
14235: This will use the machine description from the @file{arch/8086}
14236: directory to create a new kernel. A machine file may look like that:
14237:
14238: @example
14239: \ Parameter for target systems 06oct92py
14240:
14241: 4 Constant cell \ cell size in bytes
14242: 2 Constant cell<< \ cell shift to bytes
14243: 5 Constant cell>bit \ cell shift to bits
14244: 8 Constant bits/char \ bits per character
14245: 8 Constant bits/byte \ bits per byte [default: 8]
14246: 8 Constant float \ bytes per float
14247: 8 Constant /maxalign \ maximum alignment in bytes
14248: false Constant bigendian \ byte order
14249: ( true=big, false=little )
14250:
14251: include machpc.fs \ feature list
14252: @end example
14253:
14254: This part is obligatory for the cross compiler itself, the feature list
14255: is used by the kernel to conditionally compile some features in and out,
14256: depending on whether the target supports these features.
14257:
14258: There are some optional features, if you define your own primitives,
14259: have an assembler, or need special, nonstandard preparation to make the
14260: boot process work. @code{asm-include} include an assembler,
14261: @code{prims-include} includes primitives, and @code{>boot} prepares for
14262: booting.
14263:
14264: @example
14265: : asm-include ." Include assembler" cr
14266: s" arch/8086/asm.fs" included ;
14267:
14268: : prims-include ." Include primitives" cr
14269: s" arch/8086/prim.fs" included ;
14270:
14271: : >boot ." Prepare booting" cr
14272: s" ' boot >body into-forth 1+ !" evaluate ;
14273: @end example
14274:
14275: These words are used as sort of macro during the cross compilation in
14276: the file @file{kernel/main.fs}. Instead of using this macros, it would
14277: be possible --- but more complicated --- to write a new kernel project
14278: file, too.
14279:
14280: @file{kernel/main.fs} expects the machine description file name on the
14281: stack; the cross compiler itself (@file{cross.fs}) assumes that either
14282: @code{mach-file} leaves a counted string on the stack, or
14283: @code{machine-file} leaves an address, count pair of the filename on the
14284: stack.
14285:
14286: The feature list is typically controlled using @code{SetValue}, generic
14287: files that are used by several projects can use @code{DefaultValue}
14288: instead. Both functions work like @code{Value}, when the value isn't
14289: defined, but @code{SetValue} works like @code{to} if the value is
14290: defined, and @code{DefaultValue} doesn't set anything, if the value is
14291: defined.
14292:
14293: @example
14294: \ generic mach file for pc gforth 03sep97jaw
14295:
14296: true DefaultValue NIL \ relocating
14297:
14298: >ENVIRON
14299:
14300: true DefaultValue file \ controls the presence of the
14301: \ file access wordset
14302: true DefaultValue OS \ flag to indicate a operating system
14303:
14304: true DefaultValue prims \ true: primitives are c-code
14305:
14306: true DefaultValue floating \ floating point wordset is present
14307:
14308: true DefaultValue glocals \ gforth locals are present
14309: \ will be loaded
14310: true DefaultValue dcomps \ double number comparisons
14311:
14312: true DefaultValue hash \ hashing primitives are loaded/present
14313:
14314: true DefaultValue xconds \ used together with glocals,
14315: \ special conditionals supporting gforths'
14316: \ local variables
14317: true DefaultValue header \ save a header information
14318:
14319: true DefaultValue backtrace \ enables backtrace code
14320:
14321: false DefaultValue ec
14322: false DefaultValue crlf
14323:
14324: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
14325:
14326: &16 KB DefaultValue stack-size
14327: &15 KB &512 + DefaultValue fstack-size
14328: &15 KB DefaultValue rstack-size
14329: &14 KB &512 + DefaultValue lstack-size
14330: @end example
14331:
14332: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
14333: @section How the Cross Compiler Works
14334:
14335: @node Bugs, Origin, Cross Compiler, Top
14336: @appendix Bugs
14337: @cindex bug reporting
14338:
14339: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
14340:
14341: If you find a bug, please send a bug report to
14342: @email{bug-gforth@@gnu.org}. A bug report should include this
14343: information:
14344:
14345: @itemize @bullet
14346: @item
14347: The Gforth version used (it is announced at the start of an
14348: interactive Gforth session).
14349: @item
14350: The machine and operating system (on Unix
14351: systems @code{uname -a} will report this information).
14352: @item
14353: The installation options (send the file @file{config.status}).
14354: @item
14355: A complete list of changes (if any) you (or your installer) have made to the
14356: Gforth sources.
14357: @item
14358: A program (or a sequence of keyboard commands) that reproduces the bug.
14359: @item
14360: A description of what you think constitutes the buggy behaviour.
14361: @end itemize
14362:
14363: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
14364: to Report Bugs, gcc.info, GNU C Manual}.
14365:
14366:
14367: @node Origin, Forth-related information, Bugs, Top
14368: @appendix Authors and Ancestors of Gforth
14369:
14370: @section Authors and Contributors
14371: @cindex authors of Gforth
14372: @cindex contributors to Gforth
14373:
14374: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
14375: Ertl. The third major author was Jens Wilke. Lennart Benschop (who was
14376: one of Gforth's first users, in mid-1993) and Stuart Ramsden inspired us
14377: with their continuous feedback. Lennart Benshop contributed
14378: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
14379: support for calling C libraries. Helpful comments also came from Paul
14380: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
14381: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, and
14382: Robert Epprecht. Since the release of Gforth-0.2.1 there were also
14383: helpful comments from many others; thank you all, sorry for not listing
14384: you here (but digging through my mailbox to extract your names is on my
14385: to-do list). Since the release of Gforth-0.4.0 Neal Crook worked on the
14386: manual.
14387:
14388: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
14389: and autoconf, among others), and to the creators of the Internet: Gforth
14390: was developed across the Internet, and its authors did not meet
14391: physically for the first 4 years of development.
14392:
14393: @section Pedigree
14394: @cindex pedigree of Gforth
14395:
14396: Gforth descends from bigFORTH (1993) and fig-Forth. Gforth and PFE (by
14397: Dirk Zoller) will cross-fertilize each other. Of course, a significant
14398: part of the design of Gforth was prescribed by ANS Forth.
14399:
14400: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
14401: 32 bit native code version of VolksForth for the Atari ST, written
14402: mostly by Dietrich Weineck.
14403:
14404: VolksForth descends from F83. It was written by Klaus Schleisiek, Bernd
14405: Pennemann, Georg Rehfeld and Dietrich Weineck for the C64 (called
14406: UltraForth there) in the mid-80s and ported to the Atari ST in 1986.
14407:
14408: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
14409: Forth-83 standard. !! Pedigree? When?
14410:
14411: A team led by Bill Ragsdale implemented fig-Forth on many processors in
14412: 1979. Robert Selzer and Bill Ragsdale developed the original
14413: implementation of fig-Forth for the 6502 based on microForth.
14414:
14415: The principal architect of microForth was Dean Sanderson. microForth was
14416: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
14417: the 1802, and subsequently implemented on the 8080, the 6800 and the
14418: Z80.
14419:
14420: All earlier Forth systems were custom-made, usually by Charles Moore,
14421: who discovered (as he puts it) Forth during the late 60s. The first full
14422: Forth existed in 1971.
14423:
14424: A part of the information in this section comes from @cite{The Evolution
14425: of Forth} by Elizabeth D. Rather, Donald R. Colburn and Charles
14426: H. Moore, presented at the HOPL-II conference and preprinted in SIGPLAN
14427: Notices 28(3), 1993. You can find more historical and genealogical
14428: information about Forth there.
14429:
14430: @node Forth-related information, Word Index, Origin, Top
14431: @appendix Other Forth-related information
14432: @cindex Forth-related information
14433:
14434: @menu
14435: * Internet resources::
14436: * Books::
14437: * The Forth Interest Group::
14438: * Conferences::
14439: @end menu
14440:
14441:
14442: @node Internet resources, Books, Forth-related information, Forth-related information
14443: @section Internet resources
14444: @cindex internet resources
14445:
14446: @cindex comp.lang.forth
14447: @cindex frequently asked questions
14448: There is an active news group (comp.lang.forth) discussing Forth and
14449: Forth-related issues. A frequently-asked-questions (FAQ) list
14450: is posted to the news group regularly, and archived at these sites:
14451:
14452: @itemize @bullet
14453: @item
14454: @uref{ftp://rtfm.mit.edu/pub/usenet-by-group/comp.lang.forth/}
14455: @item
14456: @uref{ftp://ftp.forth.org/pub/Forth/FAQ/}
14457: @end itemize
14458:
14459: The FAQ list should be considered mandatory reading before posting to
14460: the news group.
14461:
14462: Here are some other web sites holding Forth-related material:
14463:
14464: @itemize @bullet
14465: @item
14466: @uref{http://www.taygeta.com/forth.html} -- Skip Carter's Forth pages.
14467: @item
14468: @uref{http://www.jwdt.com/~paysan/gforth.html} -- the Gforth home page.
14469: @item
14470: @uref{http://www.minerva.com/uathena.htm} -- home of ANS Forth Standard.
14471: @item
14472: @uref{http://dec.bournemouth.ac.uk/forth/index.html} -- the Forth
14473: Research page, including links to the Journal of Forth Application and
14474: Research (JFAR) and a searchable Forth bibliography.
14475: @end itemize
14476:
14477:
14478: @node Books, The Forth Interest Group, Internet resources, Forth-related information
14479: @section Books
14480: @cindex books on Forth
14481:
14482: As the Standard is relatively new, there are not many books out yet. It
14483: is not recommended to learn Forth by using Gforth and a book that is not
14484: written for ANS Forth, as you will not know your mistakes from the
14485: deviations of the book. However, books based on the Forth-83 standard
14486: should be ok, because ANS Forth is primarily an extension of Forth-83.
14487: Refer to the Forth FAQ for details of Forth-related books.
14488:
14489: @cindex standard document for ANS Forth
14490: @cindex ANS Forth document
14491: The definite reference if you want to write ANS Forth programs is, of
14492: course, the ANS Forth document. It is available in printed form from the
14493: National Standards Institute Sales Department (Tel.: USA (212) 642-4900;
14494: Fax.: USA (212) 302-1286) as document @cite{X3.215-1994} for about
14495: $200. You can also get it from Global Engineering Documents (Tel.: USA
14496: (800) 854-7179; Fax.: (303) 843-9880) for about $300.
14497:
14498: @cite{dpANS6}, the last draft of the standard, which was then submitted
14499: to ANSI for publication is available electronically and for free in some
14500: MS Word format, and it has been converted to HTML
14501: (@uref{http://www.taygeta.com/forth/dpans.html}; this HTML version also
14502: includes the answers to Requests for Interpretation (RFIs). Some
14503: pointers to these versions can be found through
14504: @*@uref{http://www.complang.tuwien.ac.at/projects/forth.html}.
14505:
14506:
14507: @node The Forth Interest Group, Conferences, Books, Forth-related information
14508: @section The Forth Interest Group
14509: @cindex Forth interest group (FIG)
14510:
14511: The Forth Interest Group (FIG) is a world-wide, non-profit,
14512: member-supported organisation. It publishes a regular magazine,
14513: @var{FORTH Dimensions}, and offers other benefits of membership. You can
14514: contact the FIG through their office email address:
14515: @email{office@@forth.org} or by visiting their web site at
14516: @uref{http://www.forth.org/}. This web site also includes links to FIG
14517: chapters in other countries and American cities
14518: (@uref{http://www.forth.org/chapters.html}).
14519:
14520: @node Conferences, , The Forth Interest Group, Forth-related information
14521: @section Conferences
14522: @cindex Conferences
14523:
14524: There are several regular conferences related to Forth. They are all
14525: well-publicised in @var{FORTH Dimensions} and on the comp.lang.forth
14526: news group:
14527:
14528: @itemize @bullet
14529: @item
14530: FORML -- the Forth modification laboratory convenes every year near
14531: Monterey, California.
14532: @item
14533: The Rochester Forth Conference -- an annual conference traditionally
14534: held in Rochester, New York.
14535: @item
14536: EuroForth -- this European conference takes place annually.
14537: @end itemize
14538:
14539:
14540: @node Word Index, Name Index, Forth-related information, Top
14541: @unnumbered Word Index
14542:
14543: This index is a list of Forth words that have ``glossary'' entries
14544: within this manual. Each word is listed with its stack effect and
14545: wordset.
14546:
14547: @printindex fn
14548:
14549: @node Name Index, Concept Index, Word Index, Top
14550: @unnumbered Name Index
14551:
14552: This index is a list of Forth words that have ``glossary'' entries
14553: within this manual.
14554:
14555: @printindex ky
14556:
14557: @node Concept Index, , Name Index, Top
14558: @unnumbered Concept and Word Index
14559:
14560: Not all entries listed in this index are present verbatim in the
14561: text. This index also duplicates, in abbreviated form, all of the words
14562: listed in the Word Index (only the names are listed for the words here).
14563:
14564: @printindex cp
14565:
14566: @contents
14567: @bye
14568:
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