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: * Concept Index:: A menu covering many topics
171:
172: @detailmenu
173: --- 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: * Files Tutorial::
212: * Interpretation and Compilation Semantics and Immediacy Tutorial::
213: * Execution Tokens Tutorial::
214: * Exceptions Tutorial::
215: * Defining Words Tutorial::
216: * Arrays and Records Tutorial::
217: * POSTPONE Tutorial::
218: * Literal Tutorial::
219: * Advanced macros Tutorial::
220: * Compilation Tokens Tutorial::
221: * Wordlists and Search Order Tutorial::
222:
223: An Introduction to ANS Forth
224:
225: * Introducing the Text Interpreter::
226: * Stacks and Postfix notation::
227: * Your first definition::
228: * How does that work?::
229: * Forth is written in Forth::
230: * Review - elements of a Forth system::
231: * Where to go next::
232: * Exercises::
233:
234: Forth Words
235:
236: * Notation::
237: * Case insensitivity::
238: * Comments::
239: * Boolean Flags::
240: * Arithmetic::
241: * Stack Manipulation::
242: * Memory::
243: * Control Structures::
244: * Defining Words::
245: * Interpretation and Compilation Semantics::
246: * Tokens for Words::
247: * Compiling words::
248: * The Text Interpreter::
249: * Word Lists::
250: * Environmental Queries::
251: * Files::
252: * Blocks::
253: * Other I/O::
254: * Locals::
255: * Structures::
256: * Object-oriented Forth::
257: * Programming Tools::
258: * Assembler and Code Words::
259: * Threading Words::
260: * Passing Commands to the OS::
261: * Keeping track of Time::
262: * Miscellaneous Words::
263:
264: Arithmetic
265:
266: * Single precision::
267: * Double precision:: Double-cell integer arithmetic
268: * Bitwise operations::
269: * Numeric comparison::
270: * Mixed precision:: Operations with single and double-cell integers
271: * Floating Point::
272:
273: Stack Manipulation
274:
275: * Data stack::
276: * Floating point stack::
277: * Return stack::
278: * Locals stack::
279: * Stack pointer manipulation::
280:
281: Memory
282:
283: * Memory model::
284: * Dictionary allocation::
285: * Heap Allocation::
286: * Memory Access::
287: * Address arithmetic::
288: * Memory Blocks::
289:
290: Control Structures
291:
292: * Selection:: IF ... ELSE ... ENDIF
293: * Simple Loops:: BEGIN ...
294: * Counted Loops:: DO
295: * Arbitrary control structures::
296: * Calls and returns::
297: * Exception Handling::
298:
299: Defining Words
300:
301: * CREATE::
302: * Variables:: Variables and user variables
303: * Constants::
304: * Values:: Initialised variables
305: * Colon Definitions::
306: * Anonymous Definitions:: Definitions without names
307: * Supplying names:: Passing definition names as strings
308: * User-defined Defining Words::
309: * Deferred words:: Allow forward references
310: * Aliases::
311:
312: User-defined Defining Words
313:
314: * CREATE..DOES> applications::
315: * CREATE..DOES> details::
316: * Advanced does> usage example::
317: * @code{Const-does>}::
318:
319: Interpretation and Compilation Semantics
320:
321: * Combined words::
322:
323: Tokens for Words
324:
325: * Execution token:: represents execution/interpretation semantics
326: * Compilation token:: represents compilation semantics
327: * Name token:: represents named words
328:
329: Compiling words
330:
331: * Literals:: Compiling data values
332: * Macros:: Compiling words
333:
334: The Text Interpreter
335:
336: * Input Sources::
337: * Number Conversion::
338: * Interpret/Compile states::
339: * Interpreter Directives::
340:
341: Word Lists
342:
343: * Vocabularies::
344: * Why use word lists?::
345: * Word list example::
346:
347: Files
348:
349: * Forth source files::
350: * General files::
351: * Search Paths::
352:
353: Search Paths
354:
355: * Source Search Paths::
356: * General Search Paths::
357:
358: Other I/O
359:
360: * Simple numeric output:: Predefined formats
361: * Formatted numeric output:: Formatted (pictured) output
362: * String Formats:: How Forth stores strings in memory
363: * Displaying characters and strings:: Other stuff
364: * Input:: Input
365:
366: Locals
367:
368: * Gforth locals::
369: * ANS Forth locals::
370:
371: Gforth locals
372:
373: * Where are locals visible by name?::
374: * How long do locals live?::
375: * Locals programming style::
376: * Locals implementation::
377:
378: Structures
379:
380: * Why explicit structure support?::
381: * Structure Usage::
382: * Structure Naming Convention::
383: * Structure Implementation::
384: * Structure Glossary::
385:
386: Object-oriented Forth
387:
388: * Why object-oriented programming?::
389: * Object-Oriented Terminology::
390: * Objects::
391: * OOF::
392: * Mini-OOF::
393: * Comparison with other object models::
394:
395: The @file{objects.fs} model
396:
397: * Properties of the Objects model::
398: * Basic Objects Usage::
399: * The Objects base class::
400: * Creating objects::
401: * Object-Oriented Programming Style::
402: * Class Binding::
403: * Method conveniences::
404: * Classes and Scoping::
405: * Dividing classes::
406: * Object Interfaces::
407: * Objects Implementation::
408: * Objects Glossary::
409:
410: The @file{oof.fs} model
411:
412: * Properties of the OOF model::
413: * Basic OOF Usage::
414: * The OOF base class::
415: * Class Declaration::
416: * Class Implementation::
417:
418: The @file{mini-oof.fs} model
419:
420: * Basic Mini-OOF Usage::
421: * Mini-OOF Example::
422: * Mini-OOF Implementation::
423:
424: Programming Tools
425:
426: * Examining::
427: * Forgetting words::
428: * Debugging:: Simple and quick.
429: * Assertions:: Making your programs self-checking.
430: * Singlestep Debugger:: Executing your program word by word.
431:
432: Assembler and Code Words
433:
434: * Code and ;code::
435: * Common Assembler:: Assembler Syntax
436: * Common Disassembler::
437: * 386 Assembler:: Deviations and special cases
438: * Alpha Assembler:: Deviations and special cases
439: * MIPS assembler:: Deviations and special cases
440: * Other assemblers:: How to write them
441:
442: Tools
443:
444: * ANS Report:: Report the words used, sorted by wordset.
445:
446: ANS conformance
447:
448: * The Core Words::
449: * The optional Block word set::
450: * The optional Double Number word set::
451: * The optional Exception word set::
452: * The optional Facility word set::
453: * The optional File-Access word set::
454: * The optional Floating-Point word set::
455: * The optional Locals word set::
456: * The optional Memory-Allocation word set::
457: * The optional Programming-Tools word set::
458: * The optional Search-Order word set::
459:
460: The Core Words
461:
462: * core-idef:: Implementation Defined Options
463: * core-ambcond:: Ambiguous Conditions
464: * core-other:: Other System Documentation
465:
466: The optional Block word set
467:
468: * block-idef:: Implementation Defined Options
469: * block-ambcond:: Ambiguous Conditions
470: * block-other:: Other System Documentation
471:
472: The optional Double Number word set
473:
474: * double-ambcond:: Ambiguous Conditions
475:
476: The optional Exception word set
477:
478: * exception-idef:: Implementation Defined Options
479:
480: The optional Facility word set
481:
482: * facility-idef:: Implementation Defined Options
483: * facility-ambcond:: Ambiguous Conditions
484:
485: The optional File-Access word set
486:
487: * file-idef:: Implementation Defined Options
488: * file-ambcond:: Ambiguous Conditions
489:
490: The optional Floating-Point word set
491:
492: * floating-idef:: Implementation Defined Options
493: * floating-ambcond:: Ambiguous Conditions
494:
495: The optional Locals word set
496:
497: * locals-idef:: Implementation Defined Options
498: * locals-ambcond:: Ambiguous Conditions
499:
500: The optional Memory-Allocation word set
501:
502: * memory-idef:: Implementation Defined Options
503:
504: The optional Programming-Tools word set
505:
506: * programming-idef:: Implementation Defined Options
507: * programming-ambcond:: Ambiguous Conditions
508:
509: The optional Search-Order word set
510:
511: * search-idef:: Implementation Defined Options
512: * search-ambcond:: Ambiguous Conditions
513:
514: Image Files
515:
516: * Image Licensing Issues:: Distribution terms for images.
517: * Image File Background:: Why have image files?
518: * Non-Relocatable Image Files:: don't always work.
519: * Data-Relocatable Image Files:: are better.
520: * Fully Relocatable Image Files:: better yet.
521: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
522: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
523: * Modifying the Startup Sequence:: and turnkey applications.
524:
525: Fully Relocatable Image Files
526:
527: * gforthmi:: The normal way
528: * cross.fs:: The hard way
529:
530: Engine
531:
532: * Portability::
533: * Threading::
534: * Primitives::
535: * Performance::
536:
537: Threading
538:
539: * Scheduling::
540: * Direct or Indirect Threaded?::
541: * DOES>::
542:
543: Primitives
544:
545: * Automatic Generation::
546: * TOS Optimization::
547: * Produced code::
548:
549: Cross Compiler
550:
551: * Using the Cross Compiler::
552: * How the Cross Compiler Works::
553:
554: @end detailmenu
555: @end menu
556:
557: @node License, Goals, Top, Top
558: @unnumbered GNU GENERAL PUBLIC LICENSE
559: @center Version 2, June 1991
560:
561: @display
562: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
563: 59 Temple Place, Suite 330, Boston, MA 02111, USA
564:
565: Everyone is permitted to copy and distribute verbatim copies
566: of this license document, but changing it is not allowed.
567: @end display
568:
569: @unnumberedsec Preamble
570:
571: The licenses for most software are designed to take away your
572: freedom to share and change it. By contrast, the GNU General Public
573: License is intended to guarantee your freedom to share and change free
574: software---to make sure the software is free for all its users. This
575: General Public License applies to most of the Free Software
576: Foundation's software and to any other program whose authors commit to
577: using it. (Some other Free Software Foundation software is covered by
578: the GNU Library General Public License instead.) You can apply it to
579: your programs, too.
580:
581: When we speak of free software, we are referring to freedom, not
582: price. Our General Public Licenses are designed to make sure that you
583: have the freedom to distribute copies of free software (and charge for
584: this service if you wish), that you receive source code or can get it
585: if you want it, that you can change the software or use pieces of it
586: in new free programs; and that you know you can do these things.
587:
588: To protect your rights, we need to make restrictions that forbid
589: anyone to deny you these rights or to ask you to surrender the rights.
590: These restrictions translate to certain responsibilities for you if you
591: distribute copies of the software, or if you modify it.
592:
593: For example, if you distribute copies of such a program, whether
594: gratis or for a fee, you must give the recipients all the rights that
595: you have. You must make sure that they, too, receive or can get the
596: source code. And you must show them these terms so they know their
597: rights.
598:
599: We protect your rights with two steps: (1) copyright the software, and
600: (2) offer you this license which gives you legal permission to copy,
601: distribute and/or modify the software.
602:
603: Also, for each author's protection and ours, we want to make certain
604: that everyone understands that there is no warranty for this free
605: software. If the software is modified by someone else and passed on, we
606: want its recipients to know that what they have is not the original, so
607: that any problems introduced by others will not reflect on the original
608: authors' reputations.
609:
610: Finally, any free program is threatened constantly by software
611: patents. We wish to avoid the danger that redistributors of a free
612: program will individually obtain patent licenses, in effect making the
613: program proprietary. To prevent this, we have made it clear that any
614: patent must be licensed for everyone's free use or not licensed at all.
615:
616: The precise terms and conditions for copying, distribution and
617: modification follow.
618:
619: @iftex
620: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
621: @end iftex
622: @ifnottex
623: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
624: @end ifnottex
625:
626: @enumerate 0
627: @item
628: This License applies to any program or other work which contains
629: a notice placed by the copyright holder saying it may be distributed
630: under the terms of this General Public License. The ``Program'', below,
631: refers to any such program or work, and a ``work based on the Program''
632: means either the Program or any derivative work under copyright law:
633: that is to say, a work containing the Program or a portion of it,
634: either verbatim or with modifications and/or translated into another
635: language. (Hereinafter, translation is included without limitation in
636: the term ``modification''.) Each licensee is addressed as ``you''.
637:
638: Activities other than copying, distribution and modification are not
639: covered by this License; they are outside its scope. The act of
640: running the Program is not restricted, and the output from the Program
641: is covered only if its contents constitute a work based on the
642: Program (independent of having been made by running the Program).
643: Whether that is true depends on what the Program does.
644:
645: @item
646: You may copy and distribute verbatim copies of the Program's
647: source code as you receive it, in any medium, provided that you
648: conspicuously and appropriately publish on each copy an appropriate
649: copyright notice and disclaimer of warranty; keep intact all the
650: notices that refer to this License and to the absence of any warranty;
651: and give any other recipients of the Program a copy of this License
652: along with the Program.
653:
654: You may charge a fee for the physical act of transferring a copy, and
655: you may at your option offer warranty protection in exchange for a fee.
656:
657: @item
658: You may modify your copy or copies of the Program or any portion
659: of it, thus forming a work based on the Program, and copy and
660: distribute such modifications or work under the terms of Section 1
661: above, provided that you also meet all of these conditions:
662:
663: @enumerate a
664: @item
665: You must cause the modified files to carry prominent notices
666: stating that you changed the files and the date of any change.
667:
668: @item
669: You must cause any work that you distribute or publish, that in
670: whole or in part contains or is derived from the Program or any
671: part thereof, to be licensed as a whole at no charge to all third
672: parties under the terms of this License.
673:
674: @item
675: If the modified program normally reads commands interactively
676: when run, you must cause it, when started running for such
677: interactive use in the most ordinary way, to print or display an
678: announcement including an appropriate copyright notice and a
679: notice that there is no warranty (or else, saying that you provide
680: a warranty) and that users may redistribute the program under
681: these conditions, and telling the user how to view a copy of this
682: License. (Exception: if the Program itself is interactive but
683: does not normally print such an announcement, your work based on
684: the Program is not required to print an announcement.)
685: @end enumerate
686:
687: These requirements apply to the modified work as a whole. If
688: identifiable sections of that work are not derived from the Program,
689: and can be reasonably considered independent and separate works in
690: themselves, then this License, and its terms, do not apply to those
691: sections when you distribute them as separate works. But when you
692: distribute the same sections as part of a whole which is a work based
693: on the Program, the distribution of the whole must be on the terms of
694: this License, whose permissions for other licensees extend to the
695: entire whole, and thus to each and every part regardless of who wrote it.
696:
697: Thus, it is not the intent of this section to claim rights or contest
698: your rights to work written entirely by you; rather, the intent is to
699: exercise the right to control the distribution of derivative or
700: collective works based on the Program.
701:
702: In addition, mere aggregation of another work not based on the Program
703: with the Program (or with a work based on the Program) on a volume of
704: a storage or distribution medium does not bring the other work under
705: the scope of this License.
706:
707: @item
708: You may copy and distribute the Program (or a work based on it,
709: under Section 2) in object code or executable form under the terms of
710: Sections 1 and 2 above provided that you also do one of the following:
711:
712: @enumerate a
713: @item
714: Accompany it with the complete corresponding machine-readable
715: source code, which must be distributed under the terms of Sections
716: 1 and 2 above on a medium customarily used for software interchange; or,
717:
718: @item
719: Accompany it with a written offer, valid for at least three
720: years, to give any third party, for a charge no more than your
721: cost of physically performing source distribution, a complete
722: machine-readable copy of the corresponding source code, to be
723: distributed under the terms of Sections 1 and 2 above on a medium
724: customarily used for software interchange; or,
725:
726: @item
727: Accompany it with the information you received as to the offer
728: to distribute corresponding source code. (This alternative is
729: allowed only for noncommercial distribution and only if you
730: received the program in object code or executable form with such
731: an offer, in accord with Subsection b above.)
732: @end enumerate
733:
734: The source code for a work means the preferred form of the work for
735: making modifications to it. For an executable work, complete source
736: code means all the source code for all modules it contains, plus any
737: associated interface definition files, plus the scripts used to
738: control compilation and installation of the executable. However, as a
739: special exception, the source code distributed need not include
740: anything that is normally distributed (in either source or binary
741: form) with the major components (compiler, kernel, and so on) of the
742: operating system on which the executable runs, unless that component
743: itself accompanies the executable.
744:
745: If distribution of executable or object code is made by offering
746: access to copy from a designated place, then offering equivalent
747: access to copy the source code from the same place counts as
748: distribution of the source code, even though third parties are not
749: compelled to copy the source along with the object code.
750:
751: @item
752: You may not copy, modify, sublicense, or distribute the Program
753: except as expressly provided under this License. Any attempt
754: otherwise to copy, modify, sublicense or distribute the Program is
755: void, and will automatically terminate your rights under this License.
756: However, parties who have received copies, or rights, from you under
757: this License will not have their licenses terminated so long as such
758: parties remain in full compliance.
759:
760: @item
761: You are not required to accept this License, since you have not
762: signed it. However, nothing else grants you permission to modify or
763: distribute the Program or its derivative works. These actions are
764: prohibited by law if you do not accept this License. Therefore, by
765: modifying or distributing the Program (or any work based on the
766: Program), you indicate your acceptance of this License to do so, and
767: all its terms and conditions for copying, distributing or modifying
768: the Program or works based on it.
769:
770: @item
771: Each time you redistribute the Program (or any work based on the
772: Program), the recipient automatically receives a license from the
773: original licensor to copy, distribute or modify the Program subject to
774: these terms and conditions. You may not impose any further
775: restrictions on the recipients' exercise of the rights granted herein.
776: You are not responsible for enforcing compliance by third parties to
777: this License.
778:
779: @item
780: If, as a consequence of a court judgment or allegation of patent
781: infringement or for any other reason (not limited to patent issues),
782: conditions are imposed on you (whether by court order, agreement or
783: otherwise) that contradict the conditions of this License, they do not
784: excuse you from the conditions of this License. If you cannot
785: distribute so as to satisfy simultaneously your obligations under this
786: License and any other pertinent obligations, then as a consequence you
787: may not distribute the Program at all. For example, if a patent
788: license would not permit royalty-free redistribution of the Program by
789: all those who receive copies directly or indirectly through you, then
790: the only way you could satisfy both it and this License would be to
791: refrain entirely from distribution of the Program.
792:
793: If any portion of this section is held invalid or unenforceable under
794: any particular circumstance, the balance of the section is intended to
795: apply and the section as a whole is intended to apply in other
796: circumstances.
797:
798: It is not the purpose of this section to induce you to infringe any
799: patents or other property right claims or to contest validity of any
800: such claims; this section has the sole purpose of protecting the
801: integrity of the free software distribution system, which is
802: implemented by public license practices. Many people have made
803: generous contributions to the wide range of software distributed
804: through that system in reliance on consistent application of that
805: system; it is up to the author/donor to decide if he or she is willing
806: to distribute software through any other system and a licensee cannot
807: impose that choice.
808:
809: This section is intended to make thoroughly clear what is believed to
810: be a consequence of the rest of this License.
811:
812: @item
813: If the distribution and/or use of the Program is restricted in
814: certain countries either by patents or by copyrighted interfaces, the
815: original copyright holder who places the Program under this License
816: may add an explicit geographical distribution limitation excluding
817: those countries, so that distribution is permitted only in or among
818: countries not thus excluded. In such case, this License incorporates
819: the limitation as if written in the body of this License.
820:
821: @item
822: The Free Software Foundation may publish revised and/or new versions
823: of the General Public License from time to time. Such new versions will
824: be similar in spirit to the present version, but may differ in detail to
825: address new problems or concerns.
826:
827: Each version is given a distinguishing version number. If the Program
828: specifies a version number of this License which applies to it and ``any
829: later version'', you have the option of following the terms and conditions
830: either of that version or of any later version published by the Free
831: Software Foundation. If the Program does not specify a version number of
832: this License, you may choose any version ever published by the Free Software
833: Foundation.
834:
835: @item
836: If you wish to incorporate parts of the Program into other free
837: programs whose distribution conditions are different, write to the author
838: to ask for permission. For software which is copyrighted by the Free
839: Software Foundation, write to the Free Software Foundation; we sometimes
840: make exceptions for this. Our decision will be guided by the two goals
841: of preserving the free status of all derivatives of our free software and
842: of promoting the sharing and reuse of software generally.
843:
844: @iftex
845: @heading NO WARRANTY
846: @end iftex
847: @ifnottex
848: @center NO WARRANTY
849: @end ifnottex
850:
851: @item
852: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
853: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
854: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
855: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
856: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
857: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
858: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
859: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
860: REPAIR OR CORRECTION.
861:
862: @item
863: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
864: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
865: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
866: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
867: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
868: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
869: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
870: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
871: POSSIBILITY OF SUCH DAMAGES.
872: @end enumerate
873:
874: @iftex
875: @heading END OF TERMS AND CONDITIONS
876: @end iftex
877: @ifnottex
878: @center END OF TERMS AND CONDITIONS
879: @end ifnottex
880:
881: @page
882: @unnumberedsec How to Apply These Terms to Your New Programs
883:
884: If you develop a new program, and you want it to be of the greatest
885: possible use to the public, the best way to achieve this is to make it
886: free software which everyone can redistribute and change under these terms.
887:
888: To do so, attach the following notices to the program. It is safest
889: to attach them to the start of each source file to most effectively
890: convey the exclusion of warranty; and each file should have at least
891: the ``copyright'' line and a pointer to where the full notice is found.
892:
893: @smallexample
894: @var{one line to give the program's name and a brief idea of what it does.}
895: Copyright (C) 19@var{yy} @var{name of author}
896:
897: This program is free software; you can redistribute it and/or modify
898: it under the terms of the GNU General Public License as published by
899: the Free Software Foundation; either version 2 of the License, or
900: (at your option) any later version.
901:
902: This program is distributed in the hope that it will be useful,
903: but WITHOUT ANY WARRANTY; without even the implied warranty of
904: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
905: GNU General Public License for more details.
906:
907: You should have received a copy of the GNU General Public License
908: along with this program; if not, write to the Free Software
909: Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
910: @end smallexample
911:
912: Also add information on how to contact you by electronic and paper mail.
913:
914: If the program is interactive, make it output a short notice like this
915: when it starts in an interactive mode:
916:
917: @smallexample
918: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
919: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
920: type `show w'.
921: This is free software, and you are welcome to redistribute it
922: under certain conditions; type `show c' for details.
923: @end smallexample
924:
925: The hypothetical commands @samp{show w} and @samp{show c} should show
926: the appropriate parts of the General Public License. Of course, the
927: commands you use may be called something other than @samp{show w} and
928: @samp{show c}; they could even be mouse-clicks or menu items---whatever
929: suits your program.
930:
931: You should also get your employer (if you work as a programmer) or your
932: school, if any, to sign a ``copyright disclaimer'' for the program, if
933: necessary. Here is a sample; alter the names:
934:
935: @smallexample
936: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
937: `Gnomovision' (which makes passes at compilers) written by James Hacker.
938:
939: @var{signature of Ty Coon}, 1 April 1989
940: Ty Coon, President of Vice
941: @end smallexample
942:
943: This General Public License does not permit incorporating your program into
944: proprietary programs. If your program is a subroutine library, you may
945: consider it more useful to permit linking proprietary applications with the
946: library. If this is what you want to do, use the GNU Library General
947: Public License instead of this License.
948:
949: @iftex
950: @unnumbered Preface
951: @cindex Preface
952: This manual documents Gforth. Some introductory material is provided for
953: readers who are unfamiliar with Forth or who are migrating to Gforth
954: from other Forth compilers. However, this manual is primarily a
955: reference manual.
956: @end iftex
957:
958: @comment TODO much more blurb here.
959:
960: @c ******************************************************************
961: @node Goals, Gforth Environment, License, Top
962: @comment node-name, next, previous, up
963: @chapter Goals of Gforth
964: @cindex goals of the Gforth project
965: The goal of the Gforth Project is to develop a standard model for
966: ANS Forth. This can be split into several subgoals:
967:
968: @itemize @bullet
969: @item
970: Gforth should conform to the ANS Forth Standard.
971: @item
972: It should be a model, i.e. it should define all the
973: implementation-dependent things.
974: @item
975: It should become standard, i.e. widely accepted and used. This goal
976: is the most difficult one.
977: @end itemize
978:
979: To achieve these goals Gforth should be
980: @itemize @bullet
981: @item
982: Similar to previous models (fig-Forth, F83)
983: @item
984: Powerful. It should provide for all the things that are considered
985: necessary today and even some that are not yet considered necessary.
986: @item
987: Efficient. It should not get the reputation of being exceptionally
988: slow.
989: @item
990: Free.
991: @item
992: Available on many machines/easy to port.
993: @end itemize
994:
995: Have we achieved these goals? Gforth conforms to the ANS Forth
996: standard. It may be considered a model, but we have not yet documented
997: which parts of the model are stable and which parts we are likely to
998: change. It certainly has not yet become a de facto standard, but it
999: appears to be quite popular. It has some similarities to and some
1000: differences from previous models. It has some powerful features, but not
1001: yet everything that we envisioned. We certainly have achieved our
1002: execution speed goals (@pxref{Performance})@footnote{However, in 1998
1003: the bar was raised when the major commercial Forth vendors switched to
1004: native code compilers.}. It is free and available on many machines.
1005:
1006: @c ******************************************************************
1007: @node Gforth Environment, Tutorial, Goals, Top
1008: @chapter Gforth Environment
1009: @cindex Gforth environment
1010:
1011: Note: ultimately, the Gforth man page will be auto-generated from the
1012: material in this chapter.
1013:
1014: @menu
1015: * Invoking Gforth:: Getting in
1016: * Leaving Gforth:: Getting out
1017: * Command-line editing::
1018: * Environment variables:: that affect how Gforth starts up
1019: * Gforth Files:: What gets installed and where
1020: * Startup speed:: When 35ms is not fast enough ...
1021: @end menu
1022:
1023: For related information about the creation of images see @ref{Image Files}.
1024:
1025: @comment ----------------------------------------------
1026: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1027: @section Invoking Gforth
1028: @cindex invoking Gforth
1029: @cindex running Gforth
1030: @cindex command-line options
1031: @cindex options on the command line
1032: @cindex flags on the command line
1033:
1034: Gforth is made up of two parts; an executable ``engine'' (named
1035: @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
1036: will usually just say @code{gforth} -- this automatically loads the
1037: default image file @file{gforth.fi}. In many other cases the default
1038: Gforth image will be invoked like this:
1039: @example
1040: gforth [file | -e forth-code] ...
1041: @end example
1042: @noindent
1043: This interprets the contents of the files and the Forth code in the order they
1044: are given.
1045:
1046: In addition to the @file{gforth} engine, there is also an engine called
1047: @file{gforth-fast}, which is faster, but gives less informative error
1048: messages (@pxref{Error messages}) and may catch fewer stack underflows.
1049: You should use it for debugged, performance-critical programs.
1050:
1051: In general, the command line looks like this:
1052:
1053: @example
1054: gforth[-fast] [engine options] [image options]
1055: @end example
1056:
1057: The engine options must come before the rest of the command
1058: line. They are:
1059:
1060: @table @code
1061: @cindex -i, command-line option
1062: @cindex --image-file, command-line option
1063: @item --image-file @i{file}
1064: @itemx -i @i{file}
1065: Loads the Forth image @i{file} instead of the default
1066: @file{gforth.fi} (@pxref{Image Files}).
1067:
1068: @cindex --appl-image, command-line option
1069: @item --appl-image @i{file}
1070: Loads the image @i{file} and leaves all further command-line arguments
1071: to the image (instead of processing them as engine options). This is
1072: useful for building executable application images on Unix, built with
1073: @code{gforthmi --application ...}.
1074:
1075: @cindex --path, command-line option
1076: @cindex -p, command-line option
1077: @item --path @i{path}
1078: @itemx -p @i{path}
1079: Uses @i{path} for searching the image file and Forth source code files
1080: instead of the default in the environment variable @code{GFORTHPATH} or
1081: the path specified at installation time (e.g.,
1082: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1083: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1084:
1085: @cindex --dictionary-size, command-line option
1086: @cindex -m, command-line option
1087: @cindex @i{size} parameters for command-line options
1088: @cindex size of the dictionary and the stacks
1089: @item --dictionary-size @i{size}
1090: @itemx -m @i{size}
1091: Allocate @i{size} space for the Forth dictionary space instead of
1092: using the default specified in the image (typically 256K). The
1093: @i{size} specification for this and subsequent options consists of
1094: an integer and a unit (e.g.,
1095: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1096: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1097: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1098: @code{e} is used.
1099:
1100: @cindex --data-stack-size, command-line option
1101: @cindex -d, command-line option
1102: @item --data-stack-size @i{size}
1103: @itemx -d @i{size}
1104: Allocate @i{size} space for the data stack instead of using the
1105: default specified in the image (typically 16K).
1106:
1107: @cindex --return-stack-size, command-line option
1108: @cindex -r, command-line option
1109: @item --return-stack-size @i{size}
1110: @itemx -r @i{size}
1111: Allocate @i{size} space for the return stack instead of using the
1112: default specified in the image (typically 15K).
1113:
1114: @cindex --fp-stack-size, command-line option
1115: @cindex -f, command-line option
1116: @item --fp-stack-size @i{size}
1117: @itemx -f @i{size}
1118: Allocate @i{size} space for the floating point stack instead of
1119: using the default specified in the image (typically 15.5K). In this case
1120: the unit specifier @code{e} refers to floating point numbers.
1121:
1122: @cindex --locals-stack-size, command-line option
1123: @cindex -l, command-line option
1124: @item --locals-stack-size @i{size}
1125: @itemx -l @i{size}
1126: Allocate @i{size} space for the locals stack instead of using the
1127: default specified in the image (typically 14.5K).
1128:
1129: @cindex -h, command-line option
1130: @cindex --help, command-line option
1131: @item --help
1132: @itemx -h
1133: Print a message about the command-line options
1134:
1135: @cindex -v, command-line option
1136: @cindex --version, command-line option
1137: @item --version
1138: @itemx -v
1139: Print version and exit
1140:
1141: @cindex --debug, command-line option
1142: @item --debug
1143: Print some information useful for debugging on startup.
1144:
1145: @cindex --offset-image, command-line option
1146: @item --offset-image
1147: Start the dictionary at a slightly different position than would be used
1148: otherwise (useful for creating data-relocatable images,
1149: @pxref{Data-Relocatable Image Files}).
1150:
1151: @cindex --no-offset-im, command-line option
1152: @item --no-offset-im
1153: Start the dictionary at the normal position.
1154:
1155: @cindex --clear-dictionary, command-line option
1156: @item --clear-dictionary
1157: Initialize all bytes in the dictionary to 0 before loading the image
1158: (@pxref{Data-Relocatable Image Files}).
1159:
1160: @cindex --die-on-signal, command-line-option
1161: @item --die-on-signal
1162: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1163: or the segmentation violation SIGSEGV) by translating it into a Forth
1164: @code{THROW}. With this option, Gforth exits if it receives such a
1165: signal. This option is useful when the engine and/or the image might be
1166: severely broken (such that it causes another signal before recovering
1167: from the first); this option avoids endless loops in such cases.
1168: @end table
1169:
1170: @cindex loading files at startup
1171: @cindex executing code on startup
1172: @cindex batch processing with Gforth
1173: As explained above, the image-specific command-line arguments for the
1174: default image @file{gforth.fi} consist of a sequence of filenames and
1175: @code{-e @var{forth-code}} options that are interpreted in the sequence
1176: in which they are given. The @code{-e @var{forth-code}} or
1177: @code{--evaluate @var{forth-code}} option evaluates the Forth
1178: code. This option takes only one argument; if you want to evaluate more
1179: Forth words, you have to quote them or use @code{-e} several times. To exit
1180: after processing the command line (instead of entering interactive mode)
1181: append @code{-e bye} to the command line.
1182:
1183: @cindex versions, invoking other versions of Gforth
1184: If you have several versions of Gforth installed, @code{gforth} will
1185: invoke the version that was installed last. @code{gforth-@i{version}}
1186: invokes a specific version. If your environment contains the variable
1187: @code{GFORTHPATH}, you may want to override it by using the
1188: @code{--path} option.
1189:
1190: Not yet implemented:
1191: On startup the system first executes the system initialization file
1192: (unless the option @code{--no-init-file} is given; note that the system
1193: resulting from using this option may not be ANS Forth conformant). Then
1194: the user initialization file @file{.gforth.fs} is executed, unless the
1195: option @code{--no-rc} is given; this file is searched for in @file{.},
1196: then in @file{~}, then in the normal path (see above).
1197:
1198:
1199:
1200: @comment ----------------------------------------------
1201: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1202: @section Leaving Gforth
1203: @cindex Gforth - leaving
1204: @cindex leaving Gforth
1205:
1206: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1207: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1208: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1209: data are discarded. For ways of saving the state of the system before
1210: leaving Gforth see @ref{Image Files}.
1211:
1212: doc-bye
1213:
1214:
1215: @comment ----------------------------------------------
1216: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1217: @section Command-line editing
1218: @cindex command-line editing
1219:
1220: Gforth maintains a history file that records every line that you type to
1221: the text interpreter. This file is preserved between sessions, and is
1222: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1223: repeatedly you can recall successively older commands from this (or
1224: previous) session(s). The full list of command-line editing facilities is:
1225:
1226: @itemize @bullet
1227: @item
1228: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1229: commands from the history buffer.
1230: @item
1231: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1232: from the history buffer.
1233: @item
1234: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1235: @item
1236: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1237: @item
1238: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1239: closing up the line.
1240: @item
1241: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1242: @item
1243: @kbd{Ctrl-a} to move the cursor to the start of the line.
1244: @item
1245: @kbd{Ctrl-e} to move the cursor to the end of the line.
1246: @item
1247: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1248: line.
1249: @item
1250: @key{TAB} to step through all possible full-word completions of the word
1251: currently being typed.
1252: @item
1253: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
1254: using @code{bye}).
1255: @item
1256: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
1257: character under the cursor.
1258: @end itemize
1259:
1260: When editing, displayable characters are inserted to the left of the
1261: cursor position; the line is always in ``insert'' (as opposed to
1262: ``overstrike'') mode.
1263:
1264: @cindex history file
1265: @cindex @file{.gforth-history}
1266: On Unix systems, the history file is @file{~/.gforth-history} by
1267: default@footnote{i.e. it is stored in the user's home directory.}. You
1268: can find out the name and location of your history file using:
1269:
1270: @example
1271: history-file type \ Unix-class systems
1272:
1273: history-file type \ Other systems
1274: history-dir type
1275: @end example
1276:
1277: If you enter long definitions by hand, you can use a text editor to
1278: paste them out of the history file into a Forth source file for reuse at
1279: a later time.
1280:
1281: Gforth never trims the size of the history file, so you should do this
1282: periodically, if necessary.
1283:
1284: @comment this is all defined in history.fs
1285: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1286: @comment chosen?
1287:
1288:
1289: @comment ----------------------------------------------
1290: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1291: @section Environment variables
1292: @cindex environment variables
1293:
1294: Gforth uses these environment variables:
1295:
1296: @itemize @bullet
1297: @item
1298: @cindex @code{GFORTHHIST} -- environment variable
1299: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1300: open/create the history file, @file{.gforth-history}. Default:
1301: @code{$HOME}.
1302:
1303: @item
1304: @cindex @code{GFORTHPATH} -- environment variable
1305: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1306: for Forth source-code files.
1307:
1308: @item
1309: @cindex @code{GFORTH} -- environment variable
1310: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1311:
1312: @item
1313: @cindex @code{GFORTHD} -- environment variable
1314: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1315:
1316: @item
1317: @cindex @code{TMP}, @code{TEMP} - environment variable
1318: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1319: location for the history file.
1320: @end itemize
1321:
1322: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1323: @comment mentioning these.
1324:
1325: All the Gforth environment variables default to sensible values if they
1326: are not set.
1327:
1328:
1329: @comment ----------------------------------------------
1330: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1331: @section Gforth files
1332: @cindex Gforth files
1333:
1334: When you install Gforth on a Unix system, it installs files in these
1335: locations by default:
1336:
1337: @itemize @bullet
1338: @item
1339: @file{/usr/local/bin/gforth}
1340: @item
1341: @file{/usr/local/bin/gforthmi}
1342: @item
1343: @file{/usr/local/man/man1/gforth.1} - man page.
1344: @item
1345: @file{/usr/local/info} - the Info version of this manual.
1346: @item
1347: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1348: @item
1349: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1350: @item
1351: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1352: @item
1353: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1354: @end itemize
1355:
1356: You can select different places for installation by using
1357: @code{configure} options (listed with @code{configure --help}).
1358:
1359: @comment ----------------------------------------------
1360: @node Startup speed, , Gforth Files, Gforth Environment
1361: @section Startup speed
1362: @cindex Startup speed
1363: @cindex speed, startup
1364:
1365: If Gforth is used for CGI scripts or in shell scripts, its startup
1366: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1367: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1368: system time.
1369:
1370: If startup speed is a problem, you may consider the following ways to
1371: improve it; or you may consider ways to reduce the number of startups
1372: (for example, by using Fast-CGI).
1373:
1374: The first step to improve startup speed is to statically link Gforth, by
1375: building it with @code{XLDFLAGS=-static}. This requires more memory for
1376: the code and will therefore slow down the first invocation, but
1377: subsequent invocations avoid the dynamic linking overhead. Another
1378: disadvantage is that Gforth won't profit from library upgrades. As a
1379: result, @code{gforth-static -e bye} takes about 17.1ms user and
1380: 8.2ms system time.
1381:
1382: The next step to improve startup speed is to use a non-relocatable image
1383: (@pxref{Non-Relocatable Image Files}). You can create this image with
1384: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1385: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1386: and a part of the copy-on-write overhead. The disadvantage is that the
1387: non-relocatable image does not work if the OS gives Gforth a different
1388: address for the dictionary, for whatever reason; so you better provide a
1389: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1390: bye} takes about 15.3ms user and 7.5ms system time.
1391:
1392: The final step is to disable dictionary hashing in Gforth. Gforth
1393: builds the hash table on startup, which takes much of the startup
1394: overhead. You can do this by commenting out the @code{include hash.fs}
1395: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1396: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1397: The disadvantages are that functionality like @code{table} and
1398: @code{ekey} is missing and that text interpretation (e.g., compiling)
1399: now takes much longer. So, you should only use this method if there is
1400: no significant text interpretation to perform (the script should be
1401: compiled into the image, amongst other things). @code{gforth-static -i
1402: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1403:
1404: @c ******************************************************************
1405: @node Tutorial, Introduction, Gforth Environment, Top
1406: @chapter Forth Tutorial
1407: @cindex Tutorial
1408: @cindex Forth Tutorial
1409:
1410: @c Topics from nac's Introduction that could be mentioned:
1411: @c press <ret> after each line
1412: @c Prompt
1413: @c numbers vs. words in dictionary on text interpretation
1414: @c what happens on redefinition
1415: @c parsing words (in particular, defining words)
1416:
1417: The difference of this chapter from the Introduction
1418: (@pxref{Introduction}) is that this tutorial is more fast-paced, should
1419: be used while sitting in front of a computer, and covers much more
1420: material, but does not explain how the Forth system works.
1421:
1422: This tutorial can be used with any ANS-compliant Forth; any
1423: Gforth-specific features are marked as such and you can skip them if you
1424: work with another Forth. This tutorial does not explain all features of
1425: Forth, just enough to get you started and give you some ideas about the
1426: facilities available in Forth. Read the rest of the manual and the
1427: standard when you are through this.
1428:
1429: The intended way to use this tutorial is that you work through it while
1430: sitting in front of the console, take a look at the examples and predict
1431: what they will do, then try them out; if the outcome is not as expected,
1432: find out why (e.g., by trying out variations of the example), so you
1433: understand what's going on. There are also some assignments that you
1434: should solve.
1435:
1436: This tutorial assumes that you have programmed before and know what,
1437: e.g., a loop is.
1438:
1439: @c !! explain compat library
1440:
1441: @menu
1442: * Starting Gforth Tutorial::
1443: * Syntax Tutorial::
1444: * Crash Course Tutorial::
1445: * Stack Tutorial::
1446: * Arithmetics Tutorial::
1447: * Stack Manipulation Tutorial::
1448: * Using files for Forth code Tutorial::
1449: * Comments Tutorial::
1450: * Colon Definitions Tutorial::
1451: * Decompilation Tutorial::
1452: * Stack-Effect Comments Tutorial::
1453: * Types Tutorial::
1454: * Factoring Tutorial::
1455: * Designing the stack effect Tutorial::
1456: * Local Variables Tutorial::
1457: * Conditional execution Tutorial::
1458: * Flags and Comparisons Tutorial::
1459: * General Loops Tutorial::
1460: * Counted loops Tutorial::
1461: * Recursion Tutorial::
1462: * Leaving definitions or loops Tutorial::
1463: * Return Stack Tutorial::
1464: * Memory Tutorial::
1465: * Characters and Strings Tutorial::
1466: * Alignment Tutorial::
1467: * Files Tutorial::
1468: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1469: * Execution Tokens Tutorial::
1470: * Exceptions Tutorial::
1471: * Defining Words Tutorial::
1472: * Arrays and Records Tutorial::
1473: * POSTPONE Tutorial::
1474: * Literal Tutorial::
1475: * Advanced macros Tutorial::
1476: * Compilation Tokens Tutorial::
1477: * Wordlists and Search Order Tutorial::
1478: @end menu
1479:
1480: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1481: @section Starting Gforth
1482: @cindex starting Gforth tutorial
1483: You can start Gforth by typing its name:
1484:
1485: @example
1486: gforth
1487: @end example
1488:
1489: That puts you into interactive mode; you can leave Gforth by typing
1490: @code{bye}. While in Gforth, you can edit the command line and access
1491: the command line history with cursor keys, similar to bash.
1492:
1493:
1494: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1495: @section Syntax
1496: @cindex syntax tutorial
1497:
1498: A @dfn{word} is a sequence of arbitrary characters (expcept white
1499: space). Words are separated by white space. E.g., each of the
1500: following lines contains exactly one word:
1501:
1502: @example
1503: word
1504: !@@#$%^&*()
1505: 1234567890
1506: 5!a
1507: @end example
1508:
1509: A frequent beginner's error is to leave away necessary white space,
1510: resulting in an error like @samp{Undefined word}; so if you see such an
1511: error, check if you have put spaces wherever necessary.
1512:
1513: @example
1514: ." hello, world" \ correct
1515: ."hello, world" \ gives an "Undefined word" error
1516: @end example
1517:
1518: Gforth and most other Forth systems ignore differences in case (they are
1519: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1520: your system is case-sensitive, you may have to type all the examples
1521: given here in upper case.
1522:
1523:
1524: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1525: @section Crash Course
1526:
1527: Type
1528:
1529: @example
1530: 0 0 !
1531: here execute
1532: ' catch >body 20 erase abort
1533: ' (quit) >body 20 erase
1534: @end example
1535:
1536: The last two examples are guaranteed to destroy parts of Gforth (and
1537: most other systems), so you better leave Gforth afterwards (if it has
1538: not finished by itself). On some systems you may have to kill gforth
1539: from outside (e.g., in Unix with @code{kill}).
1540:
1541: Now that you know how to produce crashes (and that there's not much to
1542: them), let's learn how to produce meaningful programs.
1543:
1544:
1545: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1546: @section Stack
1547: @cindex stack tutorial
1548:
1549: The most obvious feature of Forth is the stack. When you type in a
1550: number, it is pushed on the stack. You can display the content of the
1551: stack with @code{.s}.
1552:
1553: @example
1554: 1 2 .s
1555: 3 .s
1556: @end example
1557:
1558: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1559: appear in @code{.s} output as they appeared in the input.
1560:
1561: You can print the top of stack element with @code{.}.
1562:
1563: @example
1564: 1 2 3 . . .
1565: @end example
1566:
1567: In general, words consume their stack arguments (@code{.s} is an
1568: exception).
1569:
1570: @assignment
1571: What does the stack contain after @code{5 6 7 .}?
1572: @endassignment
1573:
1574:
1575: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1576: @section Arithmetics
1577: @cindex arithmetics tutorial
1578:
1579: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1580: operate on the top two stack items:
1581:
1582: @example
1583: 2 2 .s
1584: + .s
1585: .
1586: 2 1 - .
1587: 7 3 mod .
1588: @end example
1589:
1590: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1591: as in the corresponding infix expression (this is generally the case in
1592: Forth).
1593:
1594: Parentheses are superfluous (and not available), because the order of
1595: the words unambiguously determines the order of evaluation and the
1596: operands:
1597:
1598: @example
1599: 3 4 + 5 * .
1600: 3 4 5 * + .
1601: @end example
1602:
1603: @assignment
1604: What are the infix expressions corresponding to the Forth code above?
1605: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1606: known as Postfix or RPN (Reverse Polish Notation).}.
1607: @endassignment
1608:
1609: To change the sign, use @code{negate}:
1610:
1611: @example
1612: 2 negate .
1613: @end example
1614:
1615: @assignment
1616: Convert -(-3)*4-5 to Forth.
1617: @endassignment
1618:
1619: @code{/mod} performs both @code{/} and @code{mod}.
1620:
1621: @example
1622: 7 3 /mod . .
1623: @end example
1624:
1625: Reference: @ref{Arithmetic}.
1626:
1627:
1628: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1629: @section Stack Manipulation
1630: @cindex stack manipulation tutorial
1631:
1632: Stack manipulation words rearrange the data on the stack.
1633:
1634: @example
1635: 1 .s drop .s
1636: 1 .s dup .s drop drop .s
1637: 1 2 .s over .s drop drop drop
1638: 1 2 .s swap .s drop drop
1639: 1 2 3 .s rot .s drop drop drop
1640: @end example
1641:
1642: These are the most important stack manipulation words. There are also
1643: variants that manipulate twice as many stack items:
1644:
1645: @example
1646: 1 2 3 4 .s 2swap .s 2drop 2drop
1647: @end example
1648:
1649: Two more stack manipulation words are:
1650:
1651: @example
1652: 1 2 .s nip .s drop
1653: 1 2 .s tuck .s 2drop drop
1654: @end example
1655:
1656: @assignment
1657: Replace @code{nip} and @code{tuck} with combinations of other stack
1658: manipulation words.
1659:
1660: @example
1661: Given: How do you get:
1662: 1 2 3 3 2 1
1663: 1 2 3 1 2 3 2
1664: 1 2 3 1 2 3 3
1665: 1 2 3 1 3 3
1666: 1 2 3 2 1 3
1667: 1 2 3 4 4 3 2 1
1668: 1 2 3 1 2 3 1 2 3
1669: 1 2 3 4 1 2 3 4 1 2
1670: 1 2 3
1671: 1 2 3 1 2 3 4
1672: 1 2 3 1 3
1673: @end example
1674: @endassignment
1675:
1676: @example
1677: 5 dup * .
1678: @end example
1679:
1680: @assignment
1681: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1682: Write a piece of Forth code that expects two numbers on the stack
1683: (@var{a} and @var{b}, with @var{b} on top) and computes
1684: @code{(a-b)(a+1)}.
1685: @endassignment
1686:
1687: Reference: @ref{Stack Manipulation}.
1688:
1689:
1690: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1691: @section Using files for Forth code
1692: @cindex loading Forth code, tutorial
1693: @cindex files containing Forth code, tutorial
1694:
1695: While working at the Forth command line is convenient for one-line
1696: examples and short one-off code, you probably want to store your source
1697: code in files for convenient editing and persistence. You can use your
1698: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1699: Gforth}) to create @var{file.fs} and use
1700:
1701: @example
1702: s" @var{file.fs}" included
1703: @end example
1704:
1705: to load it into your Forth system. The file name extension I use for
1706: Forth files is @samp{.fs}.
1707:
1708: You can easily start Gforth with some files loaded like this:
1709:
1710: @example
1711: gforth @var{file1.fs} @var{file2.fs}
1712: @end example
1713:
1714: If an error occurs during loading these files, Gforth terminates,
1715: whereas an error during @code{INCLUDED} within Gforth usually gives you
1716: a Gforth command line. Starting the Forth system every time gives you a
1717: clean start every time, without interference from the results of earlier
1718: tries.
1719:
1720: I often put all the tests in a file, then load the code and run the
1721: tests with
1722:
1723: @example
1724: gforth @var{code.fs} @var{tests.fs} -e bye
1725: @end example
1726:
1727: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1728: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1729: restart this command without ado.
1730:
1731: The advantage of this approach is that the tests can be repeated easily
1732: every time the program ist changed, making it easy to catch bugs
1733: introduced by the change.
1734:
1735: Reference: @ref{Forth source files}.
1736:
1737:
1738: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1739: @section Comments
1740: @cindex comments tutorial
1741:
1742: @example
1743: \ That's a comment; it ends at the end of the line
1744: ( Another comment; it ends here: ) .s
1745: @end example
1746:
1747: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1748: separated with white space from the following text.
1749:
1750: @example
1751: \This gives an "Undefined word" error
1752: @end example
1753:
1754: The first @code{)} ends a comment started with @code{(}, so you cannot
1755: nest @code{(}-comments; and you cannot comment out text containing a
1756: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1757: avoid @code{)} in word names.}.
1758:
1759: I use @code{\}-comments for descriptive text and for commenting out code
1760: of one or more line; I use @code{(}-comments for describing the stack
1761: effect, the stack contents, or for commenting out sub-line pieces of
1762: code.
1763:
1764: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1765: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1766: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1767: with @kbd{M-q}.
1768:
1769: Reference: @ref{Comments}.
1770:
1771:
1772: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1773: @section Colon Definitions
1774: @cindex colon definitions, tutorial
1775: @cindex definitions, tutorial
1776: @cindex procedures, tutorial
1777: @cindex functions, tutorial
1778:
1779: are similar to procedures and functions in other programming languages.
1780:
1781: @example
1782: : squared ( n -- n^2 )
1783: dup * ;
1784: 5 squared .
1785: 7 squared .
1786: @end example
1787:
1788: @code{:} starts the colon definition; its name is @code{squared}. The
1789: following comment describes its stack effect. The words @code{dup *}
1790: are not executed, but compiled into the definition. @code{;} ends the
1791: colon definition.
1792:
1793: The newly-defined word can be used like any other word, including using
1794: it in other definitions:
1795:
1796: @example
1797: : cubed ( n -- n^3 )
1798: dup squared * ;
1799: -5 cubed .
1800: : fourth-power ( n -- n^4 )
1801: squared squared ;
1802: 3 fourth-power .
1803: @end example
1804:
1805: @assignment
1806: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1807: @code{/mod} in terms of other Forth words, and check if they work (hint:
1808: test your tests on the originals first). Don't let the
1809: @samp{redefined}-Messages spook you, they are just warnings.
1810: @endassignment
1811:
1812: Reference: @ref{Colon Definitions}.
1813:
1814:
1815: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1816: @section Decompilation
1817: @cindex decompilation tutorial
1818: @cindex see tutorial
1819:
1820: You can decompile colon definitions with @code{see}:
1821:
1822: @example
1823: see squared
1824: see cubed
1825: @end example
1826:
1827: In Gforth @code{see} shows you a reconstruction of the source code from
1828: the executable code. Informations that were present in the source, but
1829: not in the executable code, are lost (e.g., comments).
1830:
1831: You can also decompile the predefined words:
1832:
1833: @example
1834: see .
1835: see +
1836: @end example
1837:
1838:
1839: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1840: @section Stack-Effect Comments
1841: @cindex stack-effect comments, tutorial
1842: @cindex --, tutorial
1843: By convention the comment after the name of a definition describes the
1844: stack effect: The part in from of the @samp{--} describes the state of
1845: the stack before the execution of the definition, i.e., the parameters
1846: that are passed into the colon definition; the part behind the @samp{--}
1847: is the state of the stack after the execution of the definition, i.e.,
1848: the results of the definition. The stack comment only shows the top
1849: stack items that the definition accesses and/or changes.
1850:
1851: You should put a correct stack effect on every definition, even if it is
1852: just @code{( -- )}. You should also add some descriptive comment to
1853: more complicated words (I usually do this in the lines following
1854: @code{:}). If you don't do this, your code becomes unreadable (because
1855: you have to work through every definition before you can undertsand
1856: any).
1857:
1858: @assignment
1859: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1860: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1861: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1862: are done, you can compare your stack effects to those in this manual
1863: (@pxref{Word Index}).
1864: @endassignment
1865:
1866: Sometimes programmers put comments at various places in colon
1867: definitions that describe the contents of the stack at that place (stack
1868: comments); i.e., they are like the first part of a stack-effect
1869: comment. E.g.,
1870:
1871: @example
1872: : cubed ( n -- n^3 )
1873: dup squared ( n n^2 ) * ;
1874: @end example
1875:
1876: In this case the stack comment is pretty superfluous, because the word
1877: is simple enough. If you think it would be a good idea to add such a
1878: comment to increase readability, you should also consider factoring the
1879: word into several simpler words (@pxref{Factoring Tutorial,,
1880: Factoring}), which typically eliminates the need for the stack comment;
1881: however, if you decide not to refactor it, then having such a comment is
1882: better than not having it.
1883:
1884: The names of the stack items in stack-effect and stack comments in the
1885: standard, in this manual, and in many programs specify the type through
1886: a type prefix, similar to Fortran and Hungarian notation. The most
1887: frequent prefixes are:
1888:
1889: @table @code
1890: @item n
1891: signed integer
1892: @item u
1893: unsigned integer
1894: @item c
1895: character
1896: @item f
1897: Boolean flags, i.e. @code{false} or @code{true}.
1898: @item a-addr,a-
1899: Cell-aligned address
1900: @item c-addr,c-
1901: Char-aligned address (note that a Char may have two bytes in Windows NT)
1902: @item xt
1903: Execution token, same size as Cell
1904: @item w,x
1905: Cell, can contain an integer or an address. It usually takes 32, 64 or
1906: 16 bits (depending on your platform and Forth system). A cell is more
1907: commonly known as machine word, but the term @emph{word} already means
1908: something different in Forth.
1909: @item d
1910: signed double-cell integer
1911: @item ud
1912: unsigned double-cell integer
1913: @item r
1914: Float (on the FP stack)
1915: @end table
1916:
1917: You can find a more complete list in @ref{Notation}.
1918:
1919: @assignment
1920: Write stack-effect comments for all definitions you have written up to
1921: now.
1922: @endassignment
1923:
1924:
1925: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1926: @section Types
1927: @cindex types tutorial
1928:
1929: In Forth the names of the operations are not overloaded; so similar
1930: operations on different types need different names; e.g., @code{+} adds
1931: integers, and you have to use @code{f+} to add floating-point numbers.
1932: The following prefixes are often used for related operations on
1933: different types:
1934:
1935: @table @code
1936: @item (none)
1937: signed integer
1938: @item u
1939: unsigned integer
1940: @item c
1941: character
1942: @item d
1943: signed double-cell integer
1944: @item ud, du
1945: unsigned double-cell integer
1946: @item 2
1947: two cells (not-necessarily double-cell numbers)
1948: @item m, um
1949: mixed single-cell and double-cell operations
1950: @item f
1951: floating-point (note that in stack comments @samp{f} represents flags,
1952: and @samp{r} represents FP numbers).
1953: @end table
1954:
1955: If there are no differences between the signed and the unsigned variant
1956: (e.g., for @code{+}), there is only the prefix-less variant.
1957:
1958: Forth does not perform type checking, neither at compile time, nor at
1959: run time. If you use the wrong oeration, the data are interpreted
1960: incorrectly:
1961:
1962: @example
1963: -1 u.
1964: @end example
1965:
1966: If you have only experience with type-checked languages until now, and
1967: have heard how important type-checking is, don't panic! In my
1968: experience (and that of other Forthers), type errors in Forth code are
1969: usually easy to find (once you get used to it), the increased vigilance
1970: of the programmer tends to catch some harder errors in addition to most
1971: type errors, and you never have to work around the type system, so in
1972: most situations the lack of type-checking seems to be a win (projects to
1973: add type checking to Forth have not caught on).
1974:
1975:
1976: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1977: @section Factoring
1978: @cindex factoring tutorial
1979:
1980: If you try to write longer definitions, you will soon find it hard to
1981: keep track of the stack contents. Therefore, good Forth programmers
1982: tend to write only short definitions (e.g., three lines). The art of
1983: finding meaningful short definitions is known as factoring (as in
1984: factoring polynomials).
1985:
1986: Well-factored programs offer additional advantages: smaller, more
1987: general words, are easier to test and debug and can be reused more and
1988: better than larger, specialized words.
1989:
1990: So, if you run into difficulties with stack management, when writing
1991: code, try to define meaningful factors for the word, and define the word
1992: in terms of those. Even if a factor contains only two words, it is
1993: often helpful.
1994:
1995: Good factoring is not easy, and it takes some practice to get the knack
1996: for it; but even experienced Forth programmers often don't find the
1997: right solution right away, but only when rewriting the program. So, if
1998: you don't come up with a good solution immediately, keep trying, don't
1999: despair.
2000:
2001: @c example !!
2002:
2003:
2004: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
2005: @section Designing the stack effect
2006: @cindex Stack effect design, tutorial
2007: @cindex design of stack effects, tutorial
2008:
2009: In other languages you can use an arbitrary order of parameters for a
2010: function; and since there is only one result, you don't have to deal with
2011: the order of results, either.
2012:
2013: In Forth (and other stack-based languages, e.g., Postscript) the
2014: parameter and result order of a definition is important and should be
2015: designed well. The general guideline is to design the stack effect such
2016: that the word is simple to use in most cases, even if that complicates
2017: the implementation of the word. Some concrete rules are:
2018:
2019: @itemize @bullet
2020:
2021: @item
2022: Words consume all of their parameters (e.g., @code{.}).
2023:
2024: @item
2025: If there is a convention on the order of parameters (e.g., from
2026: mathematics or another programming language), stick with it (e.g.,
2027: @code{-}).
2028:
2029: @item
2030: If one parameter usually requires only a short computation (e.g., it is
2031: a constant), pass it on the top of the stack. Conversely, parameters
2032: that usually require a long sequence of code to compute should be passed
2033: as the bottom (i.e., first) parameter. This makes the code easier to
2034: read, because reader does not need to keep track of the bottom item
2035: through a long sequence of code (or, alternatively, through stack
2036: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
2037: address on top of the stack because it is usually simpler to compute
2038: than the stored value (often the address is just a variable).
2039:
2040: @item
2041: Similarly, results that are usually consumed quickly should be returned
2042: on the top of stack, whereas a result that is often used in long
2043: computations should be passed as bottom result. E.g., the file words
2044: like @code{open-file} return the error code on the top of stack, because
2045: it is usually consumed quickly by @code{throw}; moreover, the error code
2046: has to be checked before doing anything with the other results.
2047:
2048: @end itemize
2049:
2050: These rules are just general guidelines, don't lose sight of the overall
2051: goal to make the words easy to use. E.g., if the convention rule
2052: conflicts with the computation-length rule, you might decide in favour
2053: of the convention if the word will be used rarely, and in favour of the
2054: computation-length rule if the word will be used frequently (because
2055: with frequent use the cost of breaking the computation-length rule would
2056: be quite high, and frequent use makes it easier to remember an
2057: unconventional order).
2058:
2059: @c example !! structure package
2060:
2061:
2062: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2063: @section Local Variables
2064: @cindex local variables, tutorial
2065:
2066: You can define local variables (@emph{locals}) in a colon definition:
2067:
2068: @example
2069: : swap @{ a b -- b a @}
2070: b a ;
2071: 1 2 swap .s 2drop
2072: @end example
2073:
2074: (If your Forth system does not support this syntax, include
2075: @file{compat/anslocals.fs} first).
2076:
2077: In this example @code{@{ a b -- b a @}} is the locals definition; it
2078: takes two cells from the stack, puts the top of stack in @code{b} and
2079: the next stack element in @code{a}. @code{--} starts a comment ending
2080: with @code{@}}. After the locals definition, using the name of the
2081: local will push its value on the stack. You can leave the comment
2082: part (@code{-- b a}) away:
2083:
2084: @example
2085: : swap ( x1 x2 -- x2 x1 )
2086: @{ a b @} b a ;
2087: @end example
2088:
2089: In Gforth you can have several locals definitions, anywhere in a colon
2090: definition; in contrast, in a standard program you can have only one
2091: locals definition per colon definition, and that locals definition must
2092: be outside any controll structure.
2093:
2094: With locals you can write slightly longer definitions without running
2095: into stack trouble. However, I recommend trying to write colon
2096: definitions without locals for exercise purposes to help you gain the
2097: essential factoring skills.
2098:
2099: @assignment
2100: Rewrite your definitions until now with locals
2101: @endassignment
2102:
2103: Reference: @ref{Locals}.
2104:
2105:
2106: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2107: @section Conditional execution
2108: @cindex conditionals, tutorial
2109: @cindex if, tutorial
2110:
2111: In Forth you can use control structures only inside colon definitions.
2112: An @code{if}-structure looks like this:
2113:
2114: @example
2115: : abs ( n1 -- +n2 )
2116: dup 0 < if
2117: negate
2118: endif ;
2119: 5 abs .
2120: -5 abs .
2121: @end example
2122:
2123: @code{if} takes a flag from the stack. If the flag is non-zero (true),
2124: the following code is performed, otherwise execution continues after the
2125: @code{endif} (or @code{else}). @code{<} compares the top two stack
2126: elements and prioduces a flag:
2127:
2128: @example
2129: 1 2 < .
2130: 2 1 < .
2131: 1 1 < .
2132: @end example
2133:
2134: Actually the standard name for @code{endif} is @code{then}. This
2135: tutorial presents the examples using @code{endif}, because this is often
2136: less confusing for people familiar with other programming languages
2137: where @code{then} has a different meaning. If your system does not have
2138: @code{endif}, define it with
2139:
2140: @example
2141: : endif postpone then ; immediate
2142: @end example
2143:
2144: You can optionally use an @code{else}-part:
2145:
2146: @example
2147: : min ( n1 n2 -- n )
2148: 2dup < if
2149: drop
2150: else
2151: nip
2152: endif ;
2153: 2 3 min .
2154: 3 2 min .
2155: @end example
2156:
2157: @assignment
2158: Write @code{min} without @code{else}-part (hint: what's the definition
2159: of @code{nip}?).
2160: @endassignment
2161:
2162: Reference: @ref{Selection}.
2163:
2164:
2165: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2166: @section Flags and Comparisons
2167: @cindex flags tutorial
2168: @cindex comparison tutorial
2169:
2170: In a false-flag all bits are clear (0 when interpreted as integer). In
2171: a canonical true-flag all bits are set (-1 as a twos-complement signed
2172: integer); in many contexts (e.g., @code{if}) any non-zero value is
2173: treated as true flag.
2174:
2175: @example
2176: false .
2177: true .
2178: true hex u. decimal
2179: @end example
2180:
2181: Comparison words produce canonical flags:
2182:
2183: @example
2184: 1 1 = .
2185: 1 0= .
2186: 0 1 < .
2187: 0 0 < .
2188: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2189: -1 1 < .
2190: @end example
2191:
2192: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
2193: (or none) and the comparisons @code{= <> < > <= >=}. Only a part of
2194: these combinations are standard (for details see the standard,
2195: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
2196:
2197: You can use @code{and or xor invert} can be used as operations on
2198: canonical flags. Actually they are bitwise operations:
2199:
2200: @example
2201: 1 2 and .
2202: 1 2 or .
2203: 1 3 xor .
2204: 1 invert .
2205: @end example
2206:
2207: You can convert a zero/non-zero flag into a canonical flag with
2208: @code{0<>} (and complement it on the way with @code{0=}).
2209:
2210: @example
2211: 1 0= .
2212: 1 0<> .
2213: @end example
2214:
2215: You can use the all-bits-set feature of canonical flags and the bitwise
2216: operation of the Boolean operations to avoid @code{if}s:
2217:
2218: @example
2219: : foo ( n1 -- n2 )
2220: 0= if
2221: 14
2222: else
2223: 0
2224: endif ;
2225: 0 foo .
2226: 1 foo .
2227:
2228: : foo ( n1 -- n2 )
2229: 0= 14 and ;
2230: 0 foo .
2231: 1 foo .
2232: @end example
2233:
2234: @assignment
2235: Write @code{min} without @code{if}.
2236: @endassignment
2237:
2238: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
2239: @ref{Bitwise operations}.
2240:
2241:
2242: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2243: @section General Loops
2244: @cindex loops, indefinite, tutorial
2245:
2246: The endless loop is the most simple one:
2247:
2248: @example
2249: : endless ( -- )
2250: 0 begin
2251: dup . 1+
2252: again ;
2253: endless
2254: @end example
2255:
2256: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2257: does nothing at run-time, @code{again} jumps back to @code{begin}.
2258:
2259: A loop with one exit at any place looks like this:
2260:
2261: @example
2262: : log2 ( +n1 -- n2 )
2263: \ logarithmus dualis of n1>0, rounded down to the next integer
2264: assert( dup 0> )
2265: 2/ 0 begin
2266: over 0> while
2267: 1+ swap 2/ swap
2268: repeat
2269: nip ;
2270: 7 log2 .
2271: 8 log2 .
2272: @end example
2273:
2274: At run-time @code{while} consumes a flag; if it is 0, execution
2275: continues behind the @code{repeat}; if the flag is non-zero, execution
2276: continues behind the @code{while}. @code{Repeat} jumps back to
2277: @code{begin}, just like @code{again}.
2278:
2279: In Forth there are many combinations/abbreviations, like @code{1+}.
2280: However, @code{2/} is not one of them; it shifts its argument right by
2281: one bit (arithmetic shift right):
2282:
2283: @example
2284: -5 2 / .
2285: -5 2/ .
2286: @end example
2287:
2288: @code{assert(} is no standard word, but you can get it on systems other
2289: then Gforth by including @file{compat/assert.fs}. You can see what it
2290: does by trying
2291:
2292: @example
2293: 0 log2 .
2294: @end example
2295:
2296: Here's a loop with an exit at the end:
2297:
2298: @example
2299: : log2 ( +n1 -- n2 )
2300: \ logarithmus dualis of n1>0, rounded down to the next integer
2301: assert( dup 0 > )
2302: -1 begin
2303: 1+ swap 2/ swap
2304: over 0 <=
2305: until
2306: nip ;
2307: @end example
2308:
2309: @code{Until} consumes a flag; if it is non-zero, execution continues at
2310: the @code{begin}, otherwise after the @code{until}.
2311:
2312: @assignment
2313: Write a definition for computing the greatest common divisor.
2314: @endassignment
2315:
2316: Reference: @ref{Simple Loops}.
2317:
2318:
2319: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2320: @section Counted loops
2321: @cindex loops, counted, tutorial
2322:
2323: @example
2324: : ^ ( n1 u -- n )
2325: \ n = the uth power of u1
2326: 1 swap 0 u+do
2327: over *
2328: loop
2329: nip ;
2330: 3 2 ^ .
2331: 4 3 ^ .
2332: @end example
2333:
2334: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2335: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2336: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2337: times (or not at all, if @code{u3-u4<0}).
2338:
2339: You can see the stack effect design rules at work in the stack effect of
2340: the loop start words: Since the start value of the loop is more
2341: frequently constant than the end value, the start value is passed on
2342: the top-of-stack.
2343:
2344: You can access the counter of a counted loop with @code{i}:
2345:
2346: @example
2347: : fac ( u -- u! )
2348: 1 swap 1+ 1 u+do
2349: i *
2350: loop ;
2351: 5 fac .
2352: 7 fac .
2353: @end example
2354:
2355: There is also @code{+do}, which expects signed numbers (important for
2356: deciding whether to enter the loop).
2357:
2358: @assignment
2359: Write a definition for computing the nth Fibonacci number.
2360: @endassignment
2361:
2362: You can also use increments other than 1:
2363:
2364: @example
2365: : up2 ( n1 n2 -- )
2366: +do
2367: i .
2368: 2 +loop ;
2369: 10 0 up2
2370:
2371: : down2 ( n1 n2 -- )
2372: -do
2373: i .
2374: 2 -loop ;
2375: 0 10 down2
2376: @end example
2377:
2378: Reference: @ref{Counted Loops}.
2379:
2380:
2381: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2382: @section Recursion
2383: @cindex recursion tutorial
2384:
2385: Usually the name of a definition is not visible in the definition; but
2386: earlier definitions are usually visible:
2387:
2388: @example
2389: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2390: : / ( n1 n2 -- n )
2391: dup 0= if
2392: -10 throw \ report division by zero
2393: endif
2394: / \ old version
2395: ;
2396: 1 0 /
2397: @end example
2398:
2399: For recursive definitions you can use @code{recursive} (non-standard) or
2400: @code{recurse}:
2401:
2402: @example
2403: : fac1 ( n -- n! ) recursive
2404: dup 0> if
2405: dup 1- fac1 *
2406: else
2407: drop 1
2408: endif ;
2409: 7 fac1 .
2410:
2411: : fac2 ( n -- n! )
2412: dup 0> if
2413: dup 1- recurse *
2414: else
2415: drop 1
2416: endif ;
2417: 8 fac2 .
2418: @end example
2419:
2420: @assignment
2421: Write a recursive definition for computing the nth Fibonacci number.
2422: @endassignment
2423:
2424: Reference (including indirect recursion): @xref{Calls and returns}.
2425:
2426:
2427: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2428: @section Leaving definitions or loops
2429: @cindex leaving definitions, tutorial
2430: @cindex leaving loops, tutorial
2431:
2432: @code{EXIT} exits the current definition right away. For every counted
2433: loop that is left in this way, an @code{UNLOOP} has to be performed
2434: before the @code{EXIT}:
2435:
2436: @c !! real examples
2437: @example
2438: : ...
2439: ... u+do
2440: ... if
2441: ... unloop exit
2442: endif
2443: ...
2444: loop
2445: ... ;
2446: @end example
2447:
2448: @code{LEAVE} leaves the innermost counted loop right away:
2449:
2450: @example
2451: : ...
2452: ... u+do
2453: ... if
2454: ... leave
2455: endif
2456: ...
2457: loop
2458: ... ;
2459: @end example
2460:
2461: @c !! example
2462:
2463: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
2464:
2465:
2466: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2467: @section Return Stack
2468: @cindex return stack tutorial
2469:
2470: In addition to the data stack Forth also has a second stack, the return
2471: stack; most Forth systems store the return addresses of procedure calls
2472: there (thus its name). Programmers can also use this stack:
2473:
2474: @example
2475: : foo ( n1 n2 -- )
2476: .s
2477: >r .s
2478: r@@ .
2479: >r .s
2480: r@@ .
2481: r> .
2482: r@@ .
2483: r> . ;
2484: 1 2 foo
2485: @end example
2486:
2487: @code{>r} takes an element from the data stack and pushes it onto the
2488: return stack; conversely, @code{r>} moves an elementm from the return to
2489: the data stack; @code{r@@} pushes a copy of the top of the return stack
2490: on the return stack.
2491:
2492: Forth programmers usually use the return stack for storing data
2493: temporarily, if using the data stack alone would be too complex, and
2494: factoring and locals are not an option:
2495:
2496: @example
2497: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2498: rot >r rot r> ;
2499: @end example
2500:
2501: The return address of the definition and the loop control parameters of
2502: counted loops usually reside on the return stack, so you have to take
2503: all items, that you have pushed on the return stack in a colon
2504: definition or counted loop, from the return stack before the definition
2505: or loop ends. You cannot access items that you pushed on the return
2506: stack outside some definition or loop within the definition of loop.
2507:
2508: If you miscount the return stack items, this usually ends in a crash:
2509:
2510: @example
2511: : crash ( n -- )
2512: >r ;
2513: 5 crash
2514: @end example
2515:
2516: You cannot mix using locals and using the return stack (according to the
2517: standard; Gforth has no problem). However, they solve the same
2518: problems, so this shouldn't be an issue.
2519:
2520: @assignment
2521: Can you rewrite any of the definitions you wrote until now in a better
2522: way using the return stack?
2523: @endassignment
2524:
2525: Reference: @ref{Return stack}.
2526:
2527:
2528: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2529: @section Memory
2530: @cindex memory access/allocation tutorial
2531:
2532: You can create a global variable @code{v} with
2533:
2534: @example
2535: variable v ( -- addr )
2536: @end example
2537:
2538: @code{v} pushes the address of a cell in memory on the stack. This cell
2539: was reserved by @code{variable}. You can use @code{!} (store) to store
2540: values into this cell and @code{@@} (fetch) to load the value from the
2541: stack into memory:
2542:
2543: @example
2544: v .
2545: 5 v ! .s
2546: v @@ .
2547: @end example
2548:
2549: You can see a raw dump of memory with @code{dump}:
2550:
2551: @example
2552: v 1 cells .s dump
2553: @end example
2554:
2555: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2556: generally, address units (aus)) that @code{n1 cells} occupy. You can
2557: also reserve more memory:
2558:
2559: @example
2560: create v2 20 cells allot
2561: v2 20 cells dump
2562: @end example
2563:
2564: creates a word @code{v2} and reserves 20 uninitialized cells; the
2565: address pushed by @code{v2} points to the start of these 20 cells. You
2566: can use address arithmetic to access these cells:
2567:
2568: @example
2569: 3 v2 5 cells + !
2570: v2 20 cells dump
2571: @end example
2572:
2573: You can reserve and initialize memory with @code{,}:
2574:
2575: @example
2576: create v3
2577: 5 , 4 , 3 , 2 , 1 ,
2578: v3 @@ .
2579: v3 cell+ @@ .
2580: v3 2 cells + @@ .
2581: v3 5 cells dump
2582: @end example
2583:
2584: @assignment
2585: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2586: @code{u} cells, with the first of these cells at @code{addr}, the next
2587: one at @code{addr cell+} etc.
2588: @endassignment
2589:
2590: You can also reserve memory without creating a new word:
2591:
2592: @example
2593: here 10 cells allot .
2594: here .
2595: @end example
2596:
2597: @code{Here} pushes the start address of the memory area. You should
2598: store it somewhere, or you will have a hard time finding the memory area
2599: again.
2600:
2601: @code{Allot} manages dictionary memory. The dictionary memory contains
2602: the system's data structures for words etc. on Gforth and most other
2603: Forth systems. It is managed like a stack: You can free the memory that
2604: you have just @code{allot}ed with
2605:
2606: @example
2607: -10 cells allot
2608: here .
2609: @end example
2610:
2611: Note that you cannot do this if you have created a new word in the
2612: meantime (because then your @code{allot}ed memory is no longer on the
2613: top of the dictionary ``stack'').
2614:
2615: Alternatively, you can use @code{allocate} and @code{free} which allow
2616: freeing memory in any order:
2617:
2618: @example
2619: 10 cells allocate throw .s
2620: 20 cells allocate throw .s
2621: swap
2622: free throw
2623: free throw
2624: @end example
2625:
2626: The @code{throw}s deal with errors (e.g., out of memory).
2627:
2628: And there is also a
2629: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2630: garbage collector}, which eliminates the need to @code{free} memory
2631: explicitly.
2632:
2633: Reference: @ref{Memory}.
2634:
2635:
2636: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2637: @section Characters and Strings
2638: @cindex strings tutorial
2639: @cindex characters tutorial
2640:
2641: On the stack characters take up a cell, like numbers. In memory they
2642: have their own size (one 8-bit byte on most systems), and therefore
2643: require their own words for memory access:
2644:
2645: @example
2646: create v4
2647: 104 c, 97 c, 108 c, 108 c, 111 c,
2648: v4 4 chars + c@@ .
2649: v4 5 chars dump
2650: @end example
2651:
2652: The preferred representation of strings on the stack is @code{addr
2653: u-count}, where @code{addr} is the address of the first character and
2654: @code{u-count} is the number of characters in the string.
2655:
2656: @example
2657: v4 5 type
2658: @end example
2659:
2660: You get a string constant with
2661:
2662: @example
2663: s" hello, world" .s
2664: type
2665: @end example
2666:
2667: Make sure you have a space between @code{s"} and the string; @code{s"}
2668: is a normal Forth word and must be delimited with white space (try what
2669: happens when you remove the space).
2670:
2671: However, this interpretive use of @code{s"} is quite restricted: the
2672: string exists only until the next call of @code{s"} (some Forth systems
2673: keep more than one of these strings, but usually they still have a
2674: limited lifetime).
2675:
2676: @example
2677: s" hello," s" world" .s
2678: type
2679: type
2680: @end example
2681:
2682: You can also use @code{s"} in a definition, and the resulting
2683: strings then live forever (well, for as long as the definition):
2684:
2685: @example
2686: : foo s" hello," s" world" ;
2687: foo .s
2688: type
2689: type
2690: @end example
2691:
2692: @assignment
2693: @code{Emit ( c -- )} types @code{c} as character (not a number).
2694: Implement @code{type ( addr u -- )}.
2695: @endassignment
2696:
2697: Reference: @ref{Memory Blocks}.
2698:
2699:
2700: @node Alignment Tutorial, Files Tutorial, Characters and Strings Tutorial, Tutorial
2701: @section Alignment
2702: @cindex alignment tutorial
2703: @cindex memory alignment tutorial
2704:
2705: On many processors cells have to be aligned in memory, if you want to
2706: access them with @code{@@} and @code{!} (and even if the processor does
2707: not require alignment, access to aligned cells is faster).
2708:
2709: @code{Create} aligns @code{here} (i.e., the place where the next
2710: allocation will occur, and that the @code{create}d word points to).
2711: Likewise, the memory produced by @code{allocate} starts at an aligned
2712: address. Adding a number of @code{cells} to an aligned address produces
2713: another aligned address.
2714:
2715: However, address arithmetic involving @code{char+} and @code{chars} can
2716: create an address that is not cell-aligned. @code{Aligned ( addr --
2717: a-addr )} produces the next aligned address:
2718:
2719: @example
2720: v3 char+ aligned .s @@ .
2721: v3 char+ .s @@ .
2722: @end example
2723:
2724: Similarly, @code{align} advances @code{here} to the next aligned
2725: address:
2726:
2727: @example
2728: create v5 97 c,
2729: here .
2730: align here .
2731: 1000 ,
2732: @end example
2733:
2734: Note that you should use aligned addresses even if your processor does
2735: not require them, if you want your program to be portable.
2736:
2737: Reference: @ref{Address arithmetic}.
2738:
2739:
2740: @node Files Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Alignment Tutorial, Tutorial
2741: @section Files
2742: @cindex files tutorial
2743:
2744: This section gives a short introduction into how to use files inside
2745: Forth. It's broken up into five easy steps:
2746:
2747: @enumerate 1
2748: @item Opened an ASCII text file for input
2749: @item Opened a file for output
2750: @item Read input file until string matched (or some other condition matched)
2751: @item Wrote some lines from input ( modified or not) to output
2752: @item Closed the files.
2753: @end enumerate
2754:
2755: @subsection Open file for input
2756:
2757: @example
2758: s" foo.in" r/o open-file throw Value fd-in
2759: @end example
2760:
2761: @subsection Create file for output
2762:
2763: @example
2764: s" foo.out" w/o create-file throw Value fd-out
2765: @end example
2766:
2767: The available file modes are r/o for read-only access, r/w for
2768: read-write access, and w/o for write-only access. You could open both
2769: files with r/w, too, if you like. All file words return error codes; for
2770: most applications, it's best to pass there error codes with @code{throw}
2771: to the outer error handler.
2772:
2773: If you want words for opening and assigning, define them as follows:
2774:
2775: @example
2776: 0 Value fd-in
2777: 0 Value fd-out
2778: : open-input ( addr u -- ) r/o open-file throw to fd-in ;
2779: : open-output ( addr u -- ) w/o create-file throw to fd-out ;
2780: @end example
2781:
2782: Usage example:
2783:
2784: @example
2785: s" foo.in" open-input
2786: s" foo.out" open-output
2787: @end example
2788:
2789: @subsection Scan file for a particular line
2790:
2791: @example
2792: 256 Constant max-line
2793: Create line-buffer max-line 2 + allot
2794:
2795: : scan-file ( addr u -- )
2796: begin
2797: line-buffer max-line fd-in read-line throw
2798: while
2799: >r 2dup line-buffer r> compare 0=
2800: until
2801: else
2802: drop
2803: then
2804: 2drop ;
2805: @end example
2806:
2807: @code{read-line ( addr u1 fd -- u2 flag ior )} reads up to u1 bytes into
2808: the buffer at addr, and returns the number of bytes read, a flag that is
2809: false when the end of file is reached, and an error code.
2810:
2811: @code{compare ( addr1 u1 addr2 u2 -- n )} compares two strings and
2812: returns zero if both strings are equal. It returns a positive number if
2813: the first string is lexically greater, a negative if the second string
2814: is lexically greater.
2815:
2816: We haven't seen this loop here; it has two exits. Since the @code{while}
2817: exits with the number of bytes read on the stack, we have to clean up
2818: that separately; that's after the @code{else}.
2819:
2820: Usage example:
2821:
2822: @example
2823: s" The text I search is here" scan-file
2824: @end example
2825:
2826: @subsection Copy input to output
2827:
2828: @example
2829: : copy-file ( -- )
2830: begin
2831: line-buffer max-line fd-in read-line throw
2832: while
2833: line-buffer swap fd-out write-file throw
2834: repeat ;
2835: @end example
2836:
2837: @subsection Close files
2838:
2839: @example
2840: fd-in close-file throw
2841: fd-out close-file throw
2842: @end example
2843:
2844: Likewise, you can put that into definitions, too:
2845:
2846: @example
2847: : close-input ( -- ) fd-in close-file throw ;
2848: : close-output ( -- ) fd-out close-file throw ;
2849: @end example
2850:
2851: @assignment
2852: How could you modify @code{copy-file} so that it copies until a second line is
2853: matched? Can you write a program that extracts a section of a text file,
2854: given the line that starts and the line that terminates that section?
2855: @endassignment
2856:
2857: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Files Tutorial, Tutorial
2858: @section Interpretation and Compilation Semantics and Immediacy
2859: @cindex semantics tutorial
2860: @cindex interpretation semantics tutorial
2861: @cindex compilation semantics tutorial
2862: @cindex immediate, tutorial
2863:
2864: When a word is compiled, it behaves differently from being interpreted.
2865: E.g., consider @code{+}:
2866:
2867: @example
2868: 1 2 + .
2869: : foo + ;
2870: @end example
2871:
2872: These two behaviours are known as compilation and interpretation
2873: semantics. For normal words (e.g., @code{+}), the compilation semantics
2874: is to append the interpretation semantics to the currently defined word
2875: (@code{foo} in the example above). I.e., when @code{foo} is executed
2876: later, the interpretation semantics of @code{+} (i.e., adding two
2877: numbers) will be performed.
2878:
2879: However, there are words with non-default compilation semantics, e.g.,
2880: the control-flow words like @code{if}. You can use @code{immediate} to
2881: change the compilation semantics of the last defined word to be equal to
2882: the interpretation semantics:
2883:
2884: @example
2885: : [FOO] ( -- )
2886: 5 . ; immediate
2887:
2888: [FOO]
2889: : bar ( -- )
2890: [FOO] ;
2891: bar
2892: see bar
2893: @end example
2894:
2895: Two conventions to mark words with non-default compilation semnatics are
2896: names with brackets (more frequently used) and to write them all in
2897: upper case (less frequently used).
2898:
2899: In Gforth (and many other systems) you can also remove the
2900: interpretation semantics with @code{compile-only} (the compilation
2901: semantics is derived from the original interpretation semantics):
2902:
2903: @example
2904: : flip ( -- )
2905: 6 . ; compile-only \ but not immediate
2906: flip
2907:
2908: : flop ( -- )
2909: flip ;
2910: flop
2911: @end example
2912:
2913: In this example the interpretation semantics of @code{flop} is equal to
2914: the original interpretation semantics of @code{flip}.
2915:
2916: The text interpreter has two states: in interpret state, it performs the
2917: interpretation semantics of words it encounters; in compile state, it
2918: performs the compilation semantics of these words.
2919:
2920: Among other things, @code{:} switches into compile state, and @code{;}
2921: switches back to interpret state. They contain the factors @code{]}
2922: (switch to compile state) and @code{[} (switch to interpret state), that
2923: do nothing but switch the state.
2924:
2925: @example
2926: : xxx ( -- )
2927: [ 5 . ]
2928: ;
2929:
2930: xxx
2931: see xxx
2932: @end example
2933:
2934: These brackets are also the source of the naming convention mentioned
2935: above.
2936:
2937: Reference: @ref{Interpretation and Compilation Semantics}.
2938:
2939:
2940: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2941: @section Execution Tokens
2942: @cindex execution tokens tutorial
2943: @cindex XT tutorial
2944:
2945: @code{' word} gives you the execution token (XT) of a word. The XT is a
2946: cell representing the interpretation semantics of a word. You can
2947: execute this semantics with @code{execute}:
2948:
2949: @example
2950: ' + .s
2951: 1 2 rot execute .
2952: @end example
2953:
2954: The XT is similar to a function pointer in C. However, parameter
2955: passing through the stack makes it a little more flexible:
2956:
2957: @example
2958: : map-array ( ... addr u xt -- ... )
2959: \ executes xt ( ... x -- ... ) for every element of the array starting
2960: \ at addr and containing u elements
2961: @{ xt @}
2962: cells over + swap ?do
2963: i @@ xt execute
2964: 1 cells +loop ;
2965:
2966: create a 3 , 4 , 2 , -1 , 4 ,
2967: a 5 ' . map-array .s
2968: 0 a 5 ' + map-array .
2969: s" max-n" environment? drop .s
2970: a 5 ' min map-array .
2971: @end example
2972:
2973: You can use map-array with the XTs of words that consume one element
2974: more than they produce. In theory you can also use it with other XTs,
2975: but the stack effect then depends on the size of the array, which is
2976: hard to understand.
2977:
2978: Since XTs are cell-sized, you can store them in memory and manipulate
2979: them on the stack like other cells. You can also compile the XT into a
2980: word with @code{compile,}:
2981:
2982: @example
2983: : foo1 ( n1 n2 -- n )
2984: [ ' + compile, ] ;
2985: see foo
2986: @end example
2987:
2988: This is non-standard, because @code{compile,} has no compilation
2989: semantics in the standard, but it works in good Forth systems. For the
2990: broken ones, use
2991:
2992: @example
2993: : [compile,] compile, ; immediate
2994:
2995: : foo1 ( n1 n2 -- n )
2996: [ ' + ] [compile,] ;
2997: see foo
2998: @end example
2999:
3000: @code{'} is a word with default compilation semantics; it parses the
3001: next word when its interpretation semantics are executed, not during
3002: compilation:
3003:
3004: @example
3005: : foo ( -- xt )
3006: ' ;
3007: see foo
3008: : bar ( ... "word" -- ... )
3009: ' execute ;
3010: see bar
3011: 1 2 bar + .
3012: @end example
3013:
3014: You often want to parse a word during compilation and compile its XT so
3015: it will be pushed on the stack at run-time. @code{[']} does this:
3016:
3017: @example
3018: : xt-+ ( -- xt )
3019: ['] + ;
3020: see xt-+
3021: 1 2 xt-+ execute .
3022: @end example
3023:
3024: Many programmers tend to see @code{'} and the word it parses as one
3025: unit, and expect it to behave like @code{[']} when compiled, and are
3026: confused by the actual behaviour. If you are, just remember that the
3027: Forth system just takes @code{'} as one unit and has no idea that it is
3028: a parsing word (attempts to convenience programmers in this issue have
3029: usually resulted in even worse pitfalls, see
3030: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
3031: @code{State}-smartness---Why it is evil and How to Exorcise it}).
3032:
3033: Note that the state of the interpreter does not come into play when
3034: creating and executing XTs. I.e., even when you execute @code{'} in
3035: compile state, it still gives you the interpretation semantics. And
3036: whatever that state is, @code{execute} performs the semantics
3037: represented by the XT (i.e., for XTs produced with @code{'} the
3038: interpretation semantics).
3039:
3040: Reference: @ref{Tokens for Words}.
3041:
3042:
3043: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
3044: @section Exceptions
3045: @cindex exceptions tutorial
3046:
3047: @code{throw ( n -- )} causes an exception unless n is zero.
3048:
3049: @example
3050: 100 throw .s
3051: 0 throw .s
3052: @end example
3053:
3054: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
3055: it catches exceptions and pushes the number of the exception on the
3056: stack (or 0, if the xt executed without exception). If there was an
3057: exception, the stacks have the same depth as when entering @code{catch}:
3058:
3059: @example
3060: .s
3061: 3 0 ' / catch .s
3062: 3 2 ' / catch .s
3063: @end example
3064:
3065: @assignment
3066: Try the same with @code{execute} instead of @code{catch}.
3067: @endassignment
3068:
3069: @code{Throw} always jumps to the dynamically next enclosing
3070: @code{catch}, even if it has to leave several call levels to achieve
3071: this:
3072:
3073: @example
3074: : foo 100 throw ;
3075: : foo1 foo ." after foo" ;
3076: : bar ['] foo1 catch ;
3077: bar .
3078: @end example
3079:
3080: It is often important to restore a value upon leaving a definition, even
3081: if the definition is left through an exception. You can ensure this
3082: like this:
3083:
3084: @example
3085: : ...
3086: save-x
3087: ['] word-changing-x catch ( ... n )
3088: restore-x
3089: ( ... n ) throw ;
3090: @end example
3091:
3092: Gforth provides an alternative syntax in addition to @code{catch}:
3093: @code{try ... recover ... endtry}. If the code between @code{try} and
3094: @code{recover} has an exception, the stack depths are restored, the
3095: exception number is pushed on the stack, and the code between
3096: @code{recover} and @code{endtry} is performed. E.g., the definition for
3097: @code{catch} is
3098:
3099: @example
3100: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
3101: try
3102: execute 0
3103: recover
3104: nip
3105: endtry ;
3106: @end example
3107:
3108: The equivalent to the restoration code above is
3109:
3110: @example
3111: : ...
3112: save-x
3113: try
3114: word-changing-x 0
3115: recover endtry
3116: restore-x
3117: throw ;
3118: @end example
3119:
3120: This works if @code{word-changing-x} does not change the stack depth,
3121: otherwise you should add some code between @code{recover} and
3122: @code{endtry} to balance the stack.
3123:
3124: Reference: @ref{Exception Handling}.
3125:
3126:
3127: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
3128: @section Defining Words
3129: @cindex defining words tutorial
3130: @cindex does> tutorial
3131: @cindex create...does> tutorial
3132:
3133: @c before semantics?
3134:
3135: @code{:}, @code{create}, and @code{variable} are definition words: They
3136: define other words. @code{Constant} is another definition word:
3137:
3138: @example
3139: 5 constant foo
3140: foo .
3141: @end example
3142:
3143: You can also use the prefixes @code{2} (double-cell) and @code{f}
3144: (floating point) with @code{variable} and @code{constant}.
3145:
3146: You can also define your own defining words. E.g.:
3147:
3148: @example
3149: : variable ( "name" -- )
3150: create 0 , ;
3151: @end example
3152:
3153: You can also define defining words that create words that do something
3154: other than just producing their address:
3155:
3156: @example
3157: : constant ( n "name" -- )
3158: create ,
3159: does> ( -- n )
3160: ( addr ) @@ ;
3161:
3162: 5 constant foo
3163: foo .
3164: @end example
3165:
3166: The definition of @code{constant} above ends at the @code{does>}; i.e.,
3167: @code{does>} replaces @code{;}, but it also does something else: It
3168: changes the last defined word such that it pushes the address of the
3169: body of the word and then performs the code after the @code{does>}
3170: whenever it is called.
3171:
3172: In the example above, @code{constant} uses @code{,} to store 5 into the
3173: body of @code{foo}. When @code{foo} executes, it pushes the address of
3174: the body onto the stack, then (in the code after the @code{does>})
3175: fetches the 5 from there.
3176:
3177: The stack comment near the @code{does>} reflects the stack effect of the
3178: defined word, not the stack effect of the code after the @code{does>}
3179: (the difference is that the code expects the address of the body that
3180: the stack comment does not show).
3181:
3182: You can use these definition words to do factoring in cases that involve
3183: (other) definition words. E.g., a field offset is always added to an
3184: address. Instead of defining
3185:
3186: @example
3187: 2 cells constant offset-field1
3188: @end example
3189:
3190: and using this like
3191:
3192: @example
3193: ( addr ) offset-field1 +
3194: @end example
3195:
3196: you can define a definition word
3197:
3198: @example
3199: : simple-field ( n "name" -- )
3200: create ,
3201: does> ( n1 -- n1+n )
3202: ( addr ) @@ + ;
3203: @end example
3204:
3205: Definition and use of field offsets now look like this:
3206:
3207: @example
3208: 2 cells simple-field field1
3209: create mystruct 4 cells allot
3210: mystruct .s field1 .s drop
3211: @end example
3212:
3213: If you want to do something with the word without performing the code
3214: after the @code{does>}, you can access the body of a @code{create}d word
3215: with @code{>body ( xt -- addr )}:
3216:
3217: @example
3218: : value ( n "name" -- )
3219: create ,
3220: does> ( -- n1 )
3221: @@ ;
3222: : to ( n "name" -- )
3223: ' >body ! ;
3224:
3225: 5 value foo
3226: foo .
3227: 7 to foo
3228: foo .
3229: @end example
3230:
3231: @assignment
3232: Define @code{defer ( "name" -- )}, which creates a word that stores an
3233: XT (at the start the XT of @code{abort}), and upon execution
3234: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3235: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3236: recursion is one application of @code{defer}.
3237: @endassignment
3238:
3239: Reference: @ref{User-defined Defining Words}.
3240:
3241:
3242: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3243: @section Arrays and Records
3244: @cindex arrays tutorial
3245: @cindex records tutorial
3246: @cindex structs tutorial
3247:
3248: Forth has no standard words for defining data structures such as arrays
3249: and records (structs in C terminology), but you can build them yourself
3250: based on address arithmetic. You can also define words for defining
3251: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
3252:
3253: One of the first projects a Forth newcomer sets out upon when learning
3254: about defining words is an array defining word (possibly for
3255: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3256: learn something from it. However, don't be disappointed when you later
3257: learn that you have little use for these words (inappropriate use would
3258: be even worse). I have not yet found a set of useful array words yet;
3259: the needs are just too diverse, and named, global arrays (the result of
3260: naive use of defining words) are often not flexible enough (e.g.,
3261: consider how to pass them as parameters). Another such project is a set
3262: of words to help dealing with strings.
3263:
3264: On the other hand, there is a useful set of record words, and it has
3265: been defined in @file{compat/struct.fs}; these words are predefined in
3266: Gforth. They are explained in depth elsewhere in this manual (see
3267: @pxref{Structures}). The @code{simple-field} example above is
3268: simplified variant of fields in this package.
3269:
3270:
3271: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3272: @section @code{POSTPONE}
3273: @cindex postpone tutorial
3274:
3275: You can compile the compilation semantics (instead of compiling the
3276: interpretation semantics) of a word with @code{POSTPONE}:
3277:
3278: @example
3279: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3280: POSTPONE + ; immediate
3281: : foo ( n1 n2 -- n )
3282: MY-+ ;
3283: 1 2 foo .
3284: see foo
3285: @end example
3286:
3287: During the definition of @code{foo} the text interpreter performs the
3288: compilation semantics of @code{MY-+}, which performs the compilation
3289: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3290:
3291: This example also displays separate stack comments for the compilation
3292: semantics and for the stack effect of the compiled code. For words with
3293: default compilation semantics these stack effects are usually not
3294: displayed; the stack effect of the compilation semantics is always
3295: @code{( -- )} for these words, the stack effect for the compiled code is
3296: the stack effect of the interpretation semantics.
3297:
3298: Note that the state of the interpreter does not come into play when
3299: performing the compilation semantics in this way. You can also perform
3300: it interpretively, e.g.:
3301:
3302: @example
3303: : foo2 ( n1 n2 -- n )
3304: [ MY-+ ] ;
3305: 1 2 foo .
3306: see foo
3307: @end example
3308:
3309: However, there are some broken Forth systems where this does not always
3310: work, and therefore this practice was been declared non-standard in
3311: 1999.
3312: @c !! repair.fs
3313:
3314: Here is another example for using @code{POSTPONE}:
3315:
3316: @example
3317: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3318: POSTPONE negate POSTPONE + ; immediate compile-only
3319: : bar ( n1 n2 -- n )
3320: MY-- ;
3321: 2 1 bar .
3322: see bar
3323: @end example
3324:
3325: You can define @code{ENDIF} in this way:
3326:
3327: @example
3328: : ENDIF ( Compilation: orig -- )
3329: POSTPONE then ; immediate
3330: @end example
3331:
3332: @assignment
3333: Write @code{MY-2DUP} that has compilation semantics equivalent to
3334: @code{2dup}, but compiles @code{over over}.
3335: @endassignment
3336:
3337: @c !! @xref{Macros} for reference
3338:
3339:
3340: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3341: @section @code{Literal}
3342: @cindex literal tutorial
3343:
3344: You cannot @code{POSTPONE} numbers:
3345:
3346: @example
3347: : [FOO] POSTPONE 500 ; immediate
3348: @end example
3349:
3350: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
3351:
3352: @example
3353: : [FOO] ( compilation: --; run-time: -- n )
3354: 500 POSTPONE literal ; immediate
3355:
3356: : flip [FOO] ;
3357: flip .
3358: see flip
3359: @end example
3360:
3361: @code{LITERAL} consumes a number at compile-time (when it's compilation
3362: semantics are executed) and pushes it at run-time (when the code it
3363: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3364: number computed at compile time into the current word:
3365:
3366: @example
3367: : bar ( -- n )
3368: [ 2 2 + ] literal ;
3369: see bar
3370: @end example
3371:
3372: @assignment
3373: Write @code{]L} which allows writing the example above as @code{: bar (
3374: -- n ) [ 2 2 + ]L ;}
3375: @endassignment
3376:
3377: @c !! @xref{Macros} for reference
3378:
3379:
3380: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3381: @section Advanced macros
3382: @cindex macros, advanced tutorial
3383: @cindex run-time code generation, tutorial
3384:
3385: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
3386: Execution Tokens}. It frequently performs @code{execute}, a relatively
3387: expensive operation in some Forth implementations. You can use
3388: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3389: and produce a word that contains the word to be performed directly:
3390:
3391: @c use ]] ... [[
3392: @example
3393: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3394: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3395: \ array beginning at addr and containing u elements
3396: @{ xt @}
3397: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
3398: POSTPONE i POSTPONE @@ xt compile,
3399: 1 cells POSTPONE literal POSTPONE +loop ;
3400:
3401: : sum-array ( addr u -- n )
3402: 0 rot rot [ ' + compile-map-array ] ;
3403: see sum-array
3404: a 5 sum-array .
3405: @end example
3406:
3407: You can use the full power of Forth for generating the code; here's an
3408: example where the code is generated in a loop:
3409:
3410: @example
3411: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3412: \ n2=n1+(addr1)*n, addr2=addr1+cell
3413: POSTPONE tuck POSTPONE @@
3414: POSTPONE literal POSTPONE * POSTPONE +
3415: POSTPONE swap POSTPONE cell+ ;
3416:
3417: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
3418: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
3419: 0 postpone literal postpone swap
3420: [ ' compile-vmul-step compile-map-array ]
3421: postpone drop ;
3422: see compile-vmul
3423:
3424: : a-vmul ( addr -- n )
3425: \ n=a*v, where v is a vector that's as long as a and starts at addr
3426: [ a 5 compile-vmul ] ;
3427: see a-vmul
3428: a a-vmul .
3429: @end example
3430:
3431: This example uses @code{compile-map-array} to show off, but you could
3432: also use @code{map-array} instead (try it now!).
3433:
3434: You can use this technique for efficient multiplication of large
3435: matrices. In matrix multiplication, you multiply every line of one
3436: matrix with every column of the other matrix. You can generate the code
3437: for one line once, and use it for every column. The only downside of
3438: this technique is that it is cumbersome to recover the memory consumed
3439: by the generated code when you are done (and in more complicated cases
3440: it is not possible portably).
3441:
3442: @c !! @xref{Macros} for reference
3443:
3444:
3445: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3446: @section Compilation Tokens
3447: @cindex compilation tokens, tutorial
3448: @cindex CT, tutorial
3449:
3450: This section is Gforth-specific. You can skip it.
3451:
3452: @code{' word compile,} compiles the interpretation semantics. For words
3453: with default compilation semantics this is the same as performing the
3454: compilation semantics. To represent the compilation semantics of other
3455: words (e.g., words like @code{if} that have no interpretation
3456: semantics), Gforth has the concept of a compilation token (CT,
3457: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3458: You can perform the compilation semantics represented by a CT with
3459: @code{execute}:
3460:
3461: @example
3462: : foo2 ( n1 n2 -- n )
3463: [ comp' + execute ] ;
3464: see foo
3465: @end example
3466:
3467: You can compile the compilation semantics represented by a CT with
3468: @code{postpone,}:
3469:
3470: @example
3471: : foo3 ( -- )
3472: [ comp' + postpone, ] ;
3473: see foo3
3474: @end example
3475:
3476: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
3477: @code{comp'} is particularly useful for words that have no
3478: interpretation semantics:
3479:
3480: @example
3481: ' if
3482: comp' if .s 2drop
3483: @end example
3484:
3485: Reference: @ref{Tokens for Words}.
3486:
3487:
3488: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3489: @section Wordlists and Search Order
3490: @cindex wordlists tutorial
3491: @cindex search order, tutorial
3492:
3493: The dictionary is not just a memory area that allows you to allocate
3494: memory with @code{allot}, it also contains the Forth words, arranged in
3495: several wordlists. When searching for a word in a wordlist,
3496: conceptually you start searching at the youngest and proceed towards
3497: older words (in reality most systems nowadays use hash-tables); i.e., if
3498: you define a word with the same name as an older word, the new word
3499: shadows the older word.
3500:
3501: Which wordlists are searched in which order is determined by the search
3502: order. You can display the search order with @code{order}. It displays
3503: first the search order, starting with the wordlist searched first, then
3504: it displays the wordlist that will contain newly defined words.
3505:
3506: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
3507:
3508: @example
3509: wordlist constant mywords
3510: @end example
3511:
3512: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3513: defined words (the @emph{current} wordlist):
3514:
3515: @example
3516: mywords set-current
3517: order
3518: @end example
3519:
3520: Gforth does not display a name for the wordlist in @code{mywords}
3521: because this wordlist was created anonymously with @code{wordlist}.
3522:
3523: You can get the current wordlist with @code{get-current ( -- wid)}. If
3524: you want to put something into a specific wordlist without overall
3525: effect on the current wordlist, this typically looks like this:
3526:
3527: @example
3528: get-current mywords set-current ( wid )
3529: create someword
3530: ( wid ) set-current
3531: @end example
3532:
3533: You can write the search order with @code{set-order ( wid1 .. widn n --
3534: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3535: searched wordlist is topmost.
3536:
3537: @example
3538: get-order mywords swap 1+ set-order
3539: order
3540: @end example
3541:
3542: Yes, the order of wordlists in the output of @code{order} is reversed
3543: from stack comments and the output of @code{.s} and thus unintuitive.
3544:
3545: @assignment
3546: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3547: wordlist to the search order. Define @code{previous ( -- )}, which
3548: removes the first searched wordlist from the search order. Experiment
3549: with boundary conditions (you will see some crashes or situations that
3550: are hard or impossible to leave).
3551: @endassignment
3552:
3553: The search order is a powerful foundation for providing features similar
3554: to Modula-2 modules and C++ namespaces. However, trying to modularize
3555: programs in this way has disadvantages for debugging and reuse/factoring
3556: that overcome the advantages in my experience (I don't do huge projects,
3557: though). These disadvantages are not so clear in other
3558: languages/programming environments, because these languages are not so
3559: strong in debugging and reuse.
3560:
3561: @c !! example
3562:
3563: Reference: @ref{Word Lists}.
3564:
3565: @c ******************************************************************
3566: @node Introduction, Words, Tutorial, Top
3567: @comment node-name, next, previous, up
3568: @chapter An Introduction to ANS Forth
3569: @cindex Forth - an introduction
3570:
3571: The difference of this chapter from the Tutorial (@pxref{Tutorial}) is
3572: that it is slower-paced in its examples, but uses them to dive deep into
3573: explaining Forth internals (not covered by the Tutorial). Apart from
3574: that, this chapter covers far less material. It is suitable for reading
3575: without using a computer.
3576:
3577: The primary purpose of this manual is to document Gforth. However, since
3578: Forth is not a widely-known language and there is a lack of up-to-date
3579: teaching material, it seems worthwhile to provide some introductory
3580: material. For other sources of Forth-related
3581: information, see @ref{Forth-related information}.
3582:
3583: The examples in this section should work on any ANS Forth; the
3584: output shown was produced using Gforth. Each example attempts to
3585: reproduce the exact output that Gforth produces. If you try out the
3586: examples (and you should), what you should type is shown @kbd{like this}
3587: and Gforth's response is shown @code{like this}. The single exception is
3588: that, where the example shows @key{RET} it means that you should
3589: press the ``carriage return'' key. Unfortunately, some output formats for
3590: this manual cannot show the difference between @kbd{this} and
3591: @code{this} which will make trying out the examples harder (but not
3592: impossible).
3593:
3594: Forth is an unusual language. It provides an interactive development
3595: environment which includes both an interpreter and compiler. Forth
3596: programming style encourages you to break a problem down into many
3597: @cindex factoring
3598: small fragments (@dfn{factoring}), and then to develop and test each
3599: fragment interactively. Forth advocates assert that breaking the
3600: edit-compile-test cycle used by conventional programming languages can
3601: lead to great productivity improvements.
3602:
3603: @menu
3604: * Introducing the Text Interpreter::
3605: * Stacks and Postfix notation::
3606: * Your first definition::
3607: * How does that work?::
3608: * Forth is written in Forth::
3609: * Review - elements of a Forth system::
3610: * Where to go next::
3611: * Exercises::
3612: @end menu
3613:
3614: @comment ----------------------------------------------
3615: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3616: @section Introducing the Text Interpreter
3617: @cindex text interpreter
3618: @cindex outer interpreter
3619:
3620: @c IMO this is too detailed and the pace is too slow for
3621: @c an introduction. If you know German, take a look at
3622: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3623: @c to see how I do it - anton
3624:
3625: @c nac-> Where I have accepted your comments 100% and modified the text
3626: @c accordingly, I have deleted your comments. Elsewhere I have added a
3627: @c response like this to attempt to rationalise what I have done. Of
3628: @c course, this is a very clumsy mechanism for something that would be
3629: @c done far more efficiently over a beer. Please delete any dialogue
3630: @c you consider closed.
3631:
3632: When you invoke the Forth image, you will see a startup banner printed
3633: and nothing else (if you have Gforth installed on your system, try
3634: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
3635: its command line interpreter, which is called the @dfn{Text Interpreter}
3636: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
3637: about the text interpreter as you read through this chapter, for more
3638: detail @pxref{The Text Interpreter}).
3639:
3640: Although it's not obvious, Forth is actually waiting for your
3641: input. Type a number and press the @key{RET} key:
3642:
3643: @example
3644: @kbd{45@key{RET}} ok
3645: @end example
3646:
3647: Rather than give you a prompt to invite you to input something, the text
3648: interpreter prints a status message @i{after} it has processed a line
3649: of input. The status message in this case (``@code{ ok}'' followed by
3650: carriage-return) indicates that the text interpreter was able to process
3651: all of your input successfully. Now type something illegal:
3652:
3653: @example
3654: @kbd{qwer341@key{RET}}
3655: :1: Undefined word
3656: qwer341
3657: ^^^^^^^
3658: $400D2BA8 Bounce
3659: $400DBDA8 no.extensions
3660: @end example
3661:
3662: The exact text, other than the ``Undefined word'' may differ slightly on
3663: your system, but the effect is the same; when the text interpreter
3664: detects an error, it discards any remaining text on a line, resets
3665: certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
3666: messages}.
3667:
3668: The text interpreter waits for you to press carriage-return, and then
3669: processes your input line. Starting at the beginning of the line, it
3670: breaks the line into groups of characters separated by spaces. For each
3671: group of characters in turn, it makes two attempts to do something:
3672:
3673: @itemize @bullet
3674: @item
3675: @cindex name dictionary
3676: It tries to treat it as a command. It does this by searching a @dfn{name
3677: dictionary}. If the group of characters matches an entry in the name
3678: dictionary, the name dictionary provides the text interpreter with
3679: information that allows the text interpreter perform some actions. In
3680: Forth jargon, we say that the group
3681: @cindex word
3682: @cindex definition
3683: @cindex execution token
3684: @cindex xt
3685: of characters names a @dfn{word}, that the dictionary search returns an
3686: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3687: word, and that the text interpreter executes the xt. Often, the terms
3688: @dfn{word} and @dfn{definition} are used interchangeably.
3689: @item
3690: If the text interpreter fails to find a match in the name dictionary, it
3691: tries to treat the group of characters as a number in the current number
3692: base (when you start up Forth, the current number base is base 10). If
3693: the group of characters legitimately represents a number, the text
3694: interpreter pushes the number onto a stack (we'll learn more about that
3695: in the next section).
3696: @end itemize
3697:
3698: If the text interpreter is unable to do either of these things with any
3699: group of characters, it discards the group of characters and the rest of
3700: the line, then prints an error message. If the text interpreter reaches
3701: the end of the line without error, it prints the status message ``@code{ ok}''
3702: followed by carriage-return.
3703:
3704: This is the simplest command we can give to the text interpreter:
3705:
3706: @example
3707: @key{RET} ok
3708: @end example
3709:
3710: The text interpreter did everything we asked it to do (nothing) without
3711: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3712: command:
3713:
3714: @example
3715: @kbd{12 dup fred dup@key{RET}}
3716: :1: Undefined word
3717: 12 dup fred dup
3718: ^^^^
3719: $400D2BA8 Bounce
3720: $400DBDA8 no.extensions
3721: @end example
3722:
3723: When you press the carriage-return key, the text interpreter starts to
3724: work its way along the line:
3725:
3726: @itemize @bullet
3727: @item
3728: When it gets to the space after the @code{2}, it takes the group of
3729: characters @code{12} and looks them up in the name
3730: dictionary@footnote{We can't tell if it found them or not, but assume
3731: for now that it did not}. There is no match for this group of characters
3732: in the name dictionary, so it tries to treat them as a number. It is
3733: able to do this successfully, so it puts the number, 12, ``on the stack''
3734: (whatever that means).
3735: @item
3736: The text interpreter resumes scanning the line and gets the next group
3737: of characters, @code{dup}. It looks it up in the name dictionary and
3738: (you'll have to take my word for this) finds it, and executes the word
3739: @code{dup} (whatever that means).
3740: @item
3741: Once again, the text interpreter resumes scanning the line and gets the
3742: group of characters @code{fred}. It looks them up in the name
3743: dictionary, but can't find them. It tries to treat them as a number, but
3744: they don't represent any legal number.
3745: @end itemize
3746:
3747: At this point, the text interpreter gives up and prints an error
3748: message. The error message shows exactly how far the text interpreter
3749: got in processing the line. In particular, it shows that the text
3750: interpreter made no attempt to do anything with the final character
3751: group, @code{dup}, even though we have good reason to believe that the
3752: text interpreter would have no problem looking that word up and
3753: executing it a second time.
3754:
3755:
3756: @comment ----------------------------------------------
3757: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3758: @section Stacks, postfix notation and parameter passing
3759: @cindex text interpreter
3760: @cindex outer interpreter
3761:
3762: In procedural programming languages (like C and Pascal), the
3763: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3764: functions or procedures are called with @dfn{explicit parameters}. For
3765: example, in C we might write:
3766:
3767: @example
3768: total = total + new_volume(length,height,depth);
3769: @end example
3770:
3771: @noindent
3772: where new_volume is a function-call to another piece of code, and total,
3773: length, height and depth are all variables. length, height and depth are
3774: parameters to the function-call.
3775:
3776: In Forth, the equivalent of the function or procedure is the
3777: @dfn{definition} and parameters are implicitly passed between
3778: definitions using a shared stack that is visible to the
3779: programmer. Although Forth does support variables, the existence of the
3780: stack means that they are used far less often than in most other
3781: programming languages. When the text interpreter encounters a number, it
3782: will place (@dfn{push}) it on the stack. There are several stacks (the
3783: actual number is implementation-dependent ...) and the particular stack
3784: used for any operation is implied unambiguously by the operation being
3785: performed. The stack used for all integer operations is called the @dfn{data
3786: stack} and, since this is the stack used most commonly, references to
3787: ``the data stack'' are often abbreviated to ``the stack''.
3788:
3789: The stacks have a last-in, first-out (LIFO) organisation. If you type:
3790:
3791: @example
3792: @kbd{1 2 3@key{RET}} ok
3793: @end example
3794:
3795: Then this instructs the text interpreter to placed three numbers on the
3796: (data) stack. An analogy for the behaviour of the stack is to take a
3797: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3798: the table. The 3 was the last card onto the pile (``last-in'') and if
3799: you take a card off the pile then, unless you're prepared to fiddle a
3800: bit, the card that you take off will be the 3 (``first-out''). The
3801: number that will be first-out of the stack is called the @dfn{top of
3802: stack}, which
3803: @cindex TOS definition
3804: is often abbreviated to @dfn{TOS}.
3805:
3806: To understand how parameters are passed in Forth, consider the
3807: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3808: be surprised to learn that this definition performs addition. More
3809: precisely, it adds two number together and produces a result. Where does
3810: it get the two numbers from? It takes the top two numbers off the
3811: stack. Where does it place the result? On the stack. You can act-out the
3812: behaviour of @code{+} with your playing cards like this:
3813:
3814: @itemize @bullet
3815: @item
3816: Pick up two cards from the stack on the table
3817: @item
3818: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3819: numbers''
3820: @item
3821: Decide that the answer is 5
3822: @item
3823: Shuffle the two cards back into the pack and find a 5
3824: @item
3825: Put a 5 on the remaining ace that's on the table.
3826: @end itemize
3827:
3828: If you don't have a pack of cards handy but you do have Forth running,
3829: you can use the definition @code{.s} to show the current state of the stack,
3830: without affecting the stack. Type:
3831:
3832: @example
3833: @kbd{clearstack 1 2 3@key{RET}} ok
3834: @kbd{.s@key{RET}} <3> 1 2 3 ok
3835: @end example
3836:
3837: The text interpreter looks up the word @code{clearstack} and executes
3838: it; it tidies up the stack and removes any entries that may have been
3839: left on it by earlier examples. The text interpreter pushes each of the
3840: three numbers in turn onto the stack. Finally, the text interpreter
3841: looks up the word @code{.s} and executes it. The effect of executing
3842: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3843: followed by a list of all the items on the stack; the item on the far
3844: right-hand side is the TOS.
3845:
3846: You can now type:
3847:
3848: @example
3849: @kbd{+ .s@key{RET}} <2> 1 5 ok
3850: @end example
3851:
3852: @noindent
3853: which is correct; there are now 2 items on the stack and the result of
3854: the addition is 5.
3855:
3856: If you're playing with cards, try doing a second addition: pick up the
3857: two cards, work out that their sum is 6, shuffle them into the pack,
3858: look for a 6 and place that on the table. You now have just one item on
3859: the stack. What happens if you try to do a third addition? Pick up the
3860: first card, pick up the second card -- ah! There is no second card. This
3861: is called a @dfn{stack underflow} and consitutes an error. If you try to
3862: do the same thing with Forth it often reports an error (probably a Stack
3863: Underflow or an Invalid Memory Address error).
3864:
3865: The opposite situation to a stack underflow is a @dfn{stack overflow},
3866: which simply accepts that there is a finite amount of storage space
3867: reserved for the stack. To stretch the playing card analogy, if you had
3868: enough packs of cards and you piled the cards up on the table, you would
3869: eventually be unable to add another card; you'd hit the ceiling. Gforth
3870: allows you to set the maximum size of the stacks. In general, the only
3871: time that you will get a stack overflow is because a definition has a
3872: bug in it and is generating data on the stack uncontrollably.
3873:
3874: There's one final use for the playing card analogy. If you model your
3875: stack using a pack of playing cards, the maximum number of items on
3876: your stack will be 52 (I assume you didn't use the Joker). The maximum
3877: @i{value} of any item on the stack is 13 (the King). In fact, the only
3878: possible numbers are positive integer numbers 1 through 13; you can't
3879: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3880: think about some of the cards, you can accommodate different
3881: numbers. For example, you could think of the Jack as representing 0,
3882: the Queen as representing -1 and the King as representing -2. Your
3883: @i{range} remains unchanged (you can still only represent a total of 13
3884: numbers) but the numbers that you can represent are -2 through 10.
3885:
3886: In that analogy, the limit was the amount of information that a single
3887: stack entry could hold, and Forth has a similar limit. In Forth, the
3888: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3889: implementation dependent and affects the maximum value that a stack
3890: entry can hold. A Standard Forth provides a cell size of at least
3891: 16-bits, and most desktop systems use a cell size of 32-bits.
3892:
3893: Forth does not do any type checking for you, so you are free to
3894: manipulate and combine stack items in any way you wish. A convenient way
3895: of treating stack items is as 2's complement signed integers, and that
3896: is what Standard words like @code{+} do. Therefore you can type:
3897:
3898: @example
3899: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
3900: @end example
3901:
3902: If you use numbers and definitions like @code{+} in order to turn Forth
3903: into a great big pocket calculator, you will realise that it's rather
3904: different from a normal calculator. Rather than typing 2 + 3 = you had
3905: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3906: result). The terminology used to describe this difference is to say that
3907: your calculator uses @dfn{Infix Notation} (parameters and operators are
3908: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3909: operators are separate), also called @dfn{Reverse Polish Notation}.
3910:
3911: Whilst postfix notation might look confusing to begin with, it has
3912: several important advantages:
3913:
3914: @itemize @bullet
3915: @item
3916: it is unambiguous
3917: @item
3918: it is more concise
3919: @item
3920: it fits naturally with a stack-based system
3921: @end itemize
3922:
3923: To examine these claims in more detail, consider these sums:
3924:
3925: @example
3926: 6 + 5 * 4 =
3927: 4 * 5 + 6 =
3928: @end example
3929:
3930: If you're just learning maths or your maths is very rusty, you will
3931: probably come up with the answer 44 for the first and 26 for the
3932: second. If you are a bit of a whizz at maths you will remember the
3933: @i{convention} that multiplication takes precendence over addition, and
3934: you'd come up with the answer 26 both times. To explain the answer 26
3935: to someone who got the answer 44, you'd probably rewrite the first sum
3936: like this:
3937:
3938: @example
3939: 6 + (5 * 4) =
3940: @end example
3941:
3942: If what you really wanted was to perform the addition before the
3943: multiplication, you would have to use parentheses to force it.
3944:
3945: If you did the first two sums on a pocket calculator you would probably
3946: get the right answers, unless you were very cautious and entered them using
3947: these keystroke sequences:
3948:
3949: 6 + 5 = * 4 =
3950: 4 * 5 = + 6 =
3951:
3952: Postfix notation is unambiguous because the order that the operators
3953: are applied is always explicit; that also means that parentheses are
3954: never required. The operators are @i{active} (the act of quoting the
3955: operator makes the operation occur) which removes the need for ``=''.
3956:
3957: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3958: equivalent ways:
3959:
3960: @example
3961: 6 5 4 * + or:
3962: 5 4 * 6 +
3963: @end example
3964:
3965: An important thing that you should notice about this notation is that
3966: the @i{order} of the numbers does not change; if you want to subtract
3967: 2 from 10 you type @code{10 2 -}.
3968:
3969: The reason that Forth uses postfix notation is very simple to explain: it
3970: makes the implementation extremely simple, and it follows naturally from
3971: using the stack as a mechanism for passing parameters. Another way of
3972: thinking about this is to realise that all Forth definitions are
3973: @i{active}; they execute as they are encountered by the text
3974: interpreter. The result of this is that the syntax of Forth is trivially
3975: simple.
3976:
3977:
3978:
3979: @comment ----------------------------------------------
3980: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3981: @section Your first Forth definition
3982: @cindex first definition
3983:
3984: Until now, the examples we've seen have been trivial; we've just been
3985: using Forth as a bigger-than-pocket calculator. Also, each calculation
3986: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3987: again@footnote{That's not quite true. If you press the up-arrow key on
3988: your keyboard you should be able to scroll back to any earlier command,
3989: edit it and re-enter it.} In this section we'll see how to add new
3990: words to Forth's vocabulary.
3991:
3992: The easiest way to create a new word is to use a @dfn{colon
3993: definition}. We'll define a few and try them out before worrying too
3994: much about how they work. Try typing in these examples; be careful to
3995: copy the spaces accurately:
3996:
3997: @example
3998: : add-two 2 + . ;
3999: : greet ." Hello and welcome" ;
4000: : demo 5 add-two ;
4001: @end example
4002:
4003: @noindent
4004: Now try them out:
4005:
4006: @example
4007: @kbd{greet@key{RET}} Hello and welcome ok
4008: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
4009: @kbd{4 add-two@key{RET}} 6 ok
4010: @kbd{demo@key{RET}} 7 ok
4011: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
4012: @end example
4013:
4014: The first new thing that we've introduced here is the pair of words
4015: @code{:} and @code{;}. These are used to start and terminate a new
4016: definition, respectively. The first word after the @code{:} is the name
4017: for the new definition.
4018:
4019: As you can see from the examples, a definition is built up of words that
4020: have already been defined; Forth makes no distinction between
4021: definitions that existed when you started the system up, and those that
4022: you define yourself.
4023:
4024: The examples also introduce the words @code{.} (dot), @code{."}
4025: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
4026: the stack and displays it. It's like @code{.s} except that it only
4027: displays the top item of the stack and it is destructive; after it has
4028: executed, the number is no longer on the stack. There is always one
4029: space printed after the number, and no spaces before it. Dot-quote
4030: defines a string (a sequence of characters) that will be printed when
4031: the word is executed. The string can contain any printable characters
4032: except @code{"}. A @code{"} has a special function; it is not a Forth
4033: word but it acts as a delimiter (the way that delimiters work is
4034: described in the next section). Finally, @code{dup} duplicates the value
4035: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
4036:
4037: We already know that the text interpreter searches through the
4038: dictionary to locate names. If you've followed the examples earlier, you
4039: will already have a definition called @code{add-two}. Lets try modifying
4040: it by typing in a new definition:
4041:
4042: @example
4043: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
4044: @end example
4045:
4046: Forth recognised that we were defining a word that already exists, and
4047: printed a message to warn us of that fact. Let's try out the new
4048: definition:
4049:
4050: @example
4051: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
4052: @end example
4053:
4054: @noindent
4055: All that we've actually done here, though, is to create a new
4056: definition, with a particular name. The fact that there was already a
4057: definition with the same name did not make any difference to the way
4058: that the new definition was created (except that Forth printed a warning
4059: message). The old definition of add-two still exists (try @code{demo}
4060: again to see that this is true). Any new definition will use the new
4061: definition of @code{add-two}, but old definitions continue to use the
4062: version that already existed at the time that they were @code{compiled}.
4063:
4064: Before you go on to the next section, try defining and redefining some
4065: words of your own.
4066:
4067: @comment ----------------------------------------------
4068: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
4069: @section How does that work?
4070: @cindex parsing words
4071:
4072: @c That's pretty deep (IMO way too deep) for an introduction. - anton
4073:
4074: @c Is it a good idea to talk about the interpretation semantics of a
4075: @c number? We don't have an xt to go along with it. - anton
4076:
4077: @c Now that I have eliminated execution semantics, I wonder if it would not
4078: @c be better to keep them (or add run-time semantics), to make it easier to
4079: @c explain what compilation semantics usually does. - anton
4080:
4081: @c nac-> I removed the term ``default compilation sematics'' from the
4082: @c introductory chapter. Removing ``execution semantics'' was making
4083: @c everything simpler to explain, then I think the use of this term made
4084: @c everything more complex again. I replaced it with ``default
4085: @c semantics'' (which is used elsewhere in the manual) by which I mean
4086: @c ``a definition that has neither the immediate nor the compile-only
4087: @c flag set''.
4088:
4089: @c anton: I have eliminated default semantics (except in one place where it
4090: @c means "default interpretation and compilation semantics"), because it
4091: @c makes no sense in the presence of combined words. I reverted to
4092: @c "execution semantics" where necessary.
4093:
4094: @c nac-> I reworded big chunks of the ``how does that work''
4095: @c section (and, unusually for me, I think I even made it shorter!). See
4096: @c what you think -- I know I have not addressed your primary concern
4097: @c that it is too heavy-going for an introduction. From what I understood
4098: @c of your course notes it looks as though they might be a good framework.
4099: @c Things that I've tried to capture here are some things that came as a
4100: @c great revelation here when I first understood them. Also, I like the
4101: @c fact that a very simple code example shows up almost all of the issues
4102: @c that you need to understand to see how Forth works. That's unique and
4103: @c worthwhile to emphasise.
4104:
4105: @c anton: I think it's a good idea to present the details, especially those
4106: @c that you found to be a revelation, and probably the tutorial tries to be
4107: @c too superficial and does not get some of the things across that make
4108: @c Forth special. I do believe that most of the time these things should
4109: @c be discussed at the end of a section or in separate sections instead of
4110: @c in the middle of a section (e.g., the stuff you added in "User-defined
4111: @c defining words" leads in a completely different direction from the rest
4112: @c of the section).
4113:
4114: Now we're going to take another look at the definition of @code{add-two}
4115: from the previous section. From our knowledge of the way that the text
4116: interpreter works, we would have expected this result when we tried to
4117: define @code{add-two}:
4118:
4119: @example
4120: @kbd{: add-two 2 + . ;@key{RET}}
4121: ^^^^^^^
4122: Error: Undefined word
4123: @end example
4124:
4125: The reason that this didn't happen is bound up in the way that @code{:}
4126: works. The word @code{:} does two special things. The first special
4127: thing that it does prevents the text interpreter from ever seeing the
4128: characters @code{add-two}. The text interpreter uses a variable called
4129: @cindex modifying >IN
4130: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
4131: input line. When it encounters the word @code{:} it behaves in exactly
4132: the same way as it does for any other word; it looks it up in the name
4133: dictionary, finds its xt and executes it. When @code{:} executes, it
4134: looks at the input buffer, finds the word @code{add-two} and advances the
4135: value of @code{>IN} to point past it. It then does some other stuff
4136: associated with creating the new definition (including creating an entry
4137: for @code{add-two} in the name dictionary). When the execution of @code{:}
4138: completes, control returns to the text interpreter, which is oblivious
4139: to the fact that it has been tricked into ignoring part of the input
4140: line.
4141:
4142: @cindex parsing words
4143: Words like @code{:} -- words that advance the value of @code{>IN} and so
4144: prevent the text interpreter from acting on the whole of the input line
4145: -- are called @dfn{parsing words}.
4146:
4147: @cindex @code{state} - effect on the text interpreter
4148: @cindex text interpreter - effect of state
4149: The second special thing that @code{:} does is change the value of a
4150: variable called @code{state}, which affects the way that the text
4151: interpreter behaves. When Gforth starts up, @code{state} has the value
4152: 0, and the text interpreter is said to be @dfn{interpreting}. During a
4153: colon definition (started with @code{:}), @code{state} is set to -1 and
4154: the text interpreter is said to be @dfn{compiling}.
4155:
4156: In this example, the text interpreter is compiling when it processes the
4157: string ``@code{2 + . ;}''. It still breaks the string down into
4158: character sequences in the same way. However, instead of pushing the
4159: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
4160: into the definition of @code{add-two} that will make the number @code{2} get
4161: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
4162: the behaviours of @code{+} and @code{.} are also compiled into the
4163: definition.
4164:
4165: One category of words don't get compiled. These so-called @dfn{immediate
4166: words} get executed (performed @i{now}) regardless of whether the text
4167: interpreter is interpreting or compiling. The word @code{;} is an
4168: immediate word. Rather than being compiled into the definition, it
4169: executes. Its effect is to terminate the current definition, which
4170: includes changing the value of @code{state} back to 0.
4171:
4172: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
4173: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
4174: definition.
4175:
4176: In Forth, every word or number can be described in terms of two
4177: properties:
4178:
4179: @itemize @bullet
4180: @item
4181: @cindex interpretation semantics
4182: Its @dfn{interpretation semantics} describe how it will behave when the
4183: text interpreter encounters it in @dfn{interpret} state. The
4184: interpretation semantics of a word are represented by an @dfn{execution
4185: token}.
4186: @item
4187: @cindex compilation semantics
4188: Its @dfn{compilation semantics} describe how it will behave when the
4189: text interpreter encounters it in @dfn{compile} state. The compilation
4190: semantics of a word are represented in an implementation-dependent way;
4191: Gforth uses a @dfn{compilation token}.
4192: @end itemize
4193:
4194: @noindent
4195: Numbers are always treated in a fixed way:
4196:
4197: @itemize @bullet
4198: @item
4199: When the number is @dfn{interpreted}, its behaviour is to push the
4200: number onto the stack.
4201: @item
4202: When the number is @dfn{compiled}, a piece of code is appended to the
4203: current definition that pushes the number when it runs. (In other words,
4204: the compilation semantics of a number are to postpone its interpretation
4205: semantics until the run-time of the definition that it is being compiled
4206: into.)
4207: @end itemize
4208:
4209: Words don't behave in such a regular way, but most have @i{default
4210: semantics} which means that they behave like this:
4211:
4212: @itemize @bullet
4213: @item
4214: The @dfn{interpretation semantics} of the word are to do something useful.
4215: @item
4216: The @dfn{compilation semantics} of the word are to append its
4217: @dfn{interpretation semantics} to the current definition (so that its
4218: run-time behaviour is to do something useful).
4219: @end itemize
4220:
4221: @cindex immediate words
4222: The actual behaviour of any particular word can be controlled by using
4223: the words @code{immediate} and @code{compile-only} when the word is
4224: defined. These words set flags in the name dictionary entry of the most
4225: recently defined word, and these flags are retrieved by the text
4226: interpreter when it finds the word in the name dictionary.
4227:
4228: A word that is marked as @dfn{immediate} has compilation semantics that
4229: are identical to its interpretation semantics. In other words, it
4230: behaves like this:
4231:
4232: @itemize @bullet
4233: @item
4234: The @dfn{interpretation semantics} of the word are to do something useful.
4235: @item
4236: The @dfn{compilation semantics} of the word are to do something useful
4237: (and actually the same thing); i.e., it is executed during compilation.
4238: @end itemize
4239:
4240: Marking a word as @dfn{compile-only} prohibits the text interpreter from
4241: performing the interpretation semantics of the word directly; an attempt
4242: to do so will generate an error. It is never necessary to use
4243: @code{compile-only} (and it is not even part of ANS Forth, though it is
4244: provided by many implementations) but it is good etiquette to apply it
4245: to a word that will not behave correctly (and might have unexpected
4246: side-effects) in interpret state. For example, it is only legal to use
4247: the conditional word @code{IF} within a definition. If you forget this
4248: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4249: @code{compile-only} allows the text interpreter to generate a helpful
4250: error message rather than subjecting you to the consequences of your
4251: folly.
4252:
4253: This example shows the difference between an immediate and a
4254: non-immediate word:
4255:
4256: @example
4257: : show-state state @@ . ;
4258: : show-state-now show-state ; immediate
4259: : word1 show-state ;
4260: : word2 show-state-now ;
4261: @end example
4262:
4263: The word @code{immediate} after the definition of @code{show-state-now}
4264: makes that word an immediate word. These definitions introduce a new
4265: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4266: variable, and leaves it on the stack. Therefore, the behaviour of
4267: @code{show-state} is to print a number that represents the current value
4268: of @code{state}.
4269:
4270: When you execute @code{word1}, it prints the number 0, indicating that
4271: the system is interpreting. When the text interpreter compiled the
4272: definition of @code{word1}, it encountered @code{show-state} whose
4273: compilation semantics are to append its interpretation semantics to the
4274: current definition. When you execute @code{word1}, it performs the
4275: interpretation semantics of @code{show-state}. At the time that @code{word1}
4276: (and therefore @code{show-state}) are executed, the system is
4277: interpreting.
4278:
4279: When you pressed @key{RET} after entering the definition of @code{word2},
4280: you should have seen the number -1 printed, followed by ``@code{
4281: ok}''. When the text interpreter compiled the definition of
4282: @code{word2}, it encountered @code{show-state-now}, an immediate word,
4283: whose compilation semantics are therefore to perform its interpretation
4284: semantics. It is executed straight away (even before the text
4285: interpreter has moved on to process another group of characters; the
4286: @code{;} in this example). The effect of executing it are to display the
4287: value of @code{state} @i{at the time that the definition of}
4288: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4289: system is compiling at this time. If you execute @code{word2} it does
4290: nothing at all.
4291:
4292: @cindex @code{."}, how it works
4293: Before leaving the subject of immediate words, consider the behaviour of
4294: @code{."} in the definition of @code{greet}, in the previous
4295: section. This word is both a parsing word and an immediate word. Notice
4296: that there is a space between @code{."} and the start of the text
4297: @code{Hello and welcome}, but that there is no space between the last
4298: letter of @code{welcome} and the @code{"} character. The reason for this
4299: is that @code{."} is a Forth word; it must have a space after it so that
4300: the text interpreter can identify it. The @code{"} is not a Forth word;
4301: it is a @dfn{delimiter}. The examples earlier show that, when the string
4302: is displayed, there is neither a space before the @code{H} nor after the
4303: @code{e}. Since @code{."} is an immediate word, it executes at the time
4304: that @code{greet} is defined. When it executes, its behaviour is to
4305: search forward in the input line looking for the delimiter. When it
4306: finds the delimiter, it updates @code{>IN} to point past the
4307: delimiter. It also compiles some magic code into the definition of
4308: @code{greet}; the xt of a run-time routine that prints a text string. It
4309: compiles the string @code{Hello and welcome} into memory so that it is
4310: available to be printed later. When the text interpreter gains control,
4311: the next word it finds in the input stream is @code{;} and so it
4312: terminates the definition of @code{greet}.
4313:
4314:
4315: @comment ----------------------------------------------
4316: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4317: @section Forth is written in Forth
4318: @cindex structure of Forth programs
4319:
4320: When you start up a Forth compiler, a large number of definitions
4321: already exist. In Forth, you develop a new application using bottom-up
4322: programming techniques to create new definitions that are defined in
4323: terms of existing definitions. As you create each definition you can
4324: test and debug it interactively.
4325:
4326: If you have tried out the examples in this section, you will probably
4327: have typed them in by hand; when you leave Gforth, your definitions will
4328: be lost. You can avoid this by using a text editor to enter Forth source
4329: code into a file, and then loading code from the file using
4330: @code{include} (@pxref{Forth source files}). A Forth source file is
4331: processed by the text interpreter, just as though you had typed it in by
4332: hand@footnote{Actually, there are some subtle differences -- see
4333: @ref{The Text Interpreter}.}.
4334:
4335: Gforth also supports the traditional Forth alternative to using text
4336: files for program entry (@pxref{Blocks}).
4337:
4338: In common with many, if not most, Forth compilers, most of Gforth is
4339: actually written in Forth. All of the @file{.fs} files in the
4340: installation directory@footnote{For example,
4341: @file{/usr/local/share/gforth...}} are Forth source files, which you can
4342: study to see examples of Forth programming.
4343:
4344: Gforth maintains a history file that records every line that you type to
4345: the text interpreter. This file is preserved between sessions, and is
4346: used to provide a command-line recall facility. If you enter long
4347: definitions by hand, you can use a text editor to paste them out of the
4348: history file into a Forth source file for reuse at a later time
4349: (for more information @pxref{Command-line editing}).
4350:
4351:
4352: @comment ----------------------------------------------
4353: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4354: @section Review - elements of a Forth system
4355: @cindex elements of a Forth system
4356:
4357: To summarise this chapter:
4358:
4359: @itemize @bullet
4360: @item
4361: Forth programs use @dfn{factoring} to break a problem down into small
4362: fragments called @dfn{words} or @dfn{definitions}.
4363: @item
4364: Forth program development is an interactive process.
4365: @item
4366: The main command loop that accepts input, and controls both
4367: interpretation and compilation, is called the @dfn{text interpreter}
4368: (also known as the @dfn{outer interpreter}).
4369: @item
4370: Forth has a very simple syntax, consisting of words and numbers
4371: separated by spaces or carriage-return characters. Any additional syntax
4372: is imposed by @dfn{parsing words}.
4373: @item
4374: Forth uses a stack to pass parameters between words. As a result, it
4375: uses postfix notation.
4376: @item
4377: To use a word that has previously been defined, the text interpreter
4378: searches for the word in the @dfn{name dictionary}.
4379: @item
4380: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
4381: @item
4382: The text interpreter uses the value of @code{state} to select between
4383: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4384: semantics} of a word that it encounters.
4385: @item
4386: The relationship between the @dfn{interpretation semantics} and
4387: @dfn{compilation semantics} for a word
4388: depend upon the way in which the word was defined (for example, whether
4389: it is an @dfn{immediate} word).
4390: @item
4391: Forth definitions can be implemented in Forth (called @dfn{high-level
4392: definitions}) or in some other way (usually a lower-level language and
4393: as a result often called @dfn{low-level definitions}, @dfn{code
4394: definitions} or @dfn{primitives}).
4395: @item
4396: Many Forth systems are implemented mainly in Forth.
4397: @end itemize
4398:
4399:
4400: @comment ----------------------------------------------
4401: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
4402: @section Where To Go Next
4403: @cindex where to go next
4404:
4405: Amazing as it may seem, if you have read (and understood) this far, you
4406: know almost all the fundamentals about the inner workings of a Forth
4407: system. You certainly know enough to be able to read and understand the
4408: rest of this manual and the ANS Forth document, to learn more about the
4409: facilities that Forth in general and Gforth in particular provide. Even
4410: scarier, you know almost enough to implement your own Forth system.
4411: However, that's not a good idea just yet... better to try writing some
4412: programs in Gforth.
4413:
4414: Forth has such a rich vocabulary that it can be hard to know where to
4415: start in learning it. This section suggests a few sets of words that are
4416: enough to write small but useful programs. Use the word index in this
4417: document to learn more about each word, then try it out and try to write
4418: small definitions using it. Start by experimenting with these words:
4419:
4420: @itemize @bullet
4421: @item
4422: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4423: @item
4424: Comparison: @code{MIN MAX =}
4425: @item
4426: Logic: @code{AND OR XOR NOT}
4427: @item
4428: Stack manipulation: @code{DUP DROP SWAP OVER}
4429: @item
4430: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
4431: @item
4432: Input/Output: @code{. ." EMIT CR KEY}
4433: @item
4434: Defining words: @code{: ; CREATE}
4435: @item
4436: Memory allocation words: @code{ALLOT ,}
4437: @item
4438: Tools: @code{SEE WORDS .S MARKER}
4439: @end itemize
4440:
4441: When you have mastered those, go on to:
4442:
4443: @itemize @bullet
4444: @item
4445: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
4446: @item
4447: Memory access: @code{@@ !}
4448: @end itemize
4449:
4450: When you have mastered these, there's nothing for it but to read through
4451: the whole of this manual and find out what you've missed.
4452:
4453: @comment ----------------------------------------------
4454: @node Exercises, , Where to go next, Introduction
4455: @section Exercises
4456: @cindex exercises
4457:
4458: TODO: provide a set of programming excercises linked into the stuff done
4459: already and into other sections of the manual. Provide solutions to all
4460: the exercises in a .fs file in the distribution.
4461:
4462: @c Get some inspiration from Starting Forth and Kelly&Spies.
4463:
4464: @c excercises:
4465: @c 1. take inches and convert to feet and inches.
4466: @c 2. take temperature and convert from fahrenheight to celcius;
4467: @c may need to care about symmetric vs floored??
4468: @c 3. take input line and do character substitution
4469: @c to encipher or decipher
4470: @c 4. as above but work on a file for in and out
4471: @c 5. take input line and convert to pig-latin
4472: @c
4473: @c thing of sets of things to exercise then come up with
4474: @c problems that need those things.
4475:
4476:
4477: @c ******************************************************************
4478: @node Words, Error messages, Introduction, Top
4479: @chapter Forth Words
4480: @cindex words
4481:
4482: @menu
4483: * Notation::
4484: * Case insensitivity::
4485: * Comments::
4486: * Boolean Flags::
4487: * Arithmetic::
4488: * Stack Manipulation::
4489: * Memory::
4490: * Control Structures::
4491: * Defining Words::
4492: * Interpretation and Compilation Semantics::
4493: * Tokens for Words::
4494: * Compiling words::
4495: * The Text Interpreter::
4496: * Word Lists::
4497: * Environmental Queries::
4498: * Files::
4499: * Blocks::
4500: * Other I/O::
4501: * Locals::
4502: * Structures::
4503: * Object-oriented Forth::
4504: * Programming Tools::
4505: * Assembler and Code Words::
4506: * Threading Words::
4507: * Passing Commands to the OS::
4508: * Keeping track of Time::
4509: * Miscellaneous Words::
4510: @end menu
4511:
4512: @node Notation, Case insensitivity, Words, Words
4513: @section Notation
4514: @cindex notation of glossary entries
4515: @cindex format of glossary entries
4516: @cindex glossary notation format
4517: @cindex word glossary entry format
4518:
4519: The Forth words are described in this section in the glossary notation
4520: that has become a de-facto standard for Forth texts:
4521:
4522: @format
4523: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
4524: @end format
4525: @i{Description}
4526:
4527: @table @var
4528: @item word
4529: The name of the word.
4530:
4531: @item Stack effect
4532: @cindex stack effect
4533: The stack effect is written in the notation @code{@i{before} --
4534: @i{after}}, where @i{before} and @i{after} describe the top of
4535: stack entries before and after the execution of the word. The rest of
4536: the stack is not touched by the word. The top of stack is rightmost,
4537: i.e., a stack sequence is written as it is typed in. Note that Gforth
4538: uses a separate floating point stack, but a unified stack
4539: notation. Also, return stack effects are not shown in @i{stack
4540: effect}, but in @i{Description}. The name of a stack item describes
4541: the type and/or the function of the item. See below for a discussion of
4542: the types.
4543:
4544: All words have two stack effects: A compile-time stack effect and a
4545: run-time stack effect. The compile-time stack-effect of most words is
4546: @i{ -- }. If the compile-time stack-effect of a word deviates from
4547: this standard behaviour, or the word does other unusual things at
4548: compile time, both stack effects are shown; otherwise only the run-time
4549: stack effect is shown.
4550:
4551: @cindex pronounciation of words
4552: @item pronunciation
4553: How the word is pronounced.
4554:
4555: @cindex wordset
4556: @cindex environment wordset
4557: @item wordset
4558: The ANS Forth standard is divided into several word sets. A standard
4559: system need not support all of them. Therefore, in theory, the fewer
4560: word sets your program uses the more portable it will be. However, we
4561: suspect that most ANS Forth systems on personal machines will feature
4562: all word sets. Words that are not defined in ANS Forth have
4563: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
4564: describes words that will work in future releases of Gforth;
4565: @code{gforth-internal} words are more volatile. Environmental query
4566: strings are also displayed like words; you can recognize them by the
4567: @code{environment} in the word set field.
4568:
4569: @item Description
4570: A description of the behaviour of the word.
4571: @end table
4572:
4573: @cindex types of stack items
4574: @cindex stack item types
4575: The type of a stack item is specified by the character(s) the name
4576: starts with:
4577:
4578: @table @code
4579: @item f
4580: @cindex @code{f}, stack item type
4581: Boolean flags, i.e. @code{false} or @code{true}.
4582: @item c
4583: @cindex @code{c}, stack item type
4584: Char
4585: @item w
4586: @cindex @code{w}, stack item type
4587: Cell, can contain an integer or an address
4588: @item n
4589: @cindex @code{n}, stack item type
4590: signed integer
4591: @item u
4592: @cindex @code{u}, stack item type
4593: unsigned integer
4594: @item d
4595: @cindex @code{d}, stack item type
4596: double sized signed integer
4597: @item ud
4598: @cindex @code{ud}, stack item type
4599: double sized unsigned integer
4600: @item r
4601: @cindex @code{r}, stack item type
4602: Float (on the FP stack)
4603: @item a-
4604: @cindex @code{a_}, stack item type
4605: Cell-aligned address
4606: @item c-
4607: @cindex @code{c_}, stack item type
4608: Char-aligned address (note that a Char may have two bytes in Windows NT)
4609: @item f-
4610: @cindex @code{f_}, stack item type
4611: Float-aligned address
4612: @item df-
4613: @cindex @code{df_}, stack item type
4614: Address aligned for IEEE double precision float
4615: @item sf-
4616: @cindex @code{sf_}, stack item type
4617: Address aligned for IEEE single precision float
4618: @item xt
4619: @cindex @code{xt}, stack item type
4620: Execution token, same size as Cell
4621: @item wid
4622: @cindex @code{wid}, stack item type
4623: Word list ID, same size as Cell
4624: @item ior, wior
4625: @cindex ior type description
4626: @cindex wior type description
4627: I/O result code, cell-sized. In Gforth, you can @code{throw} iors.
4628: @item f83name
4629: @cindex @code{f83name}, stack item type
4630: Pointer to a name structure
4631: @item "
4632: @cindex @code{"}, stack item type
4633: string in the input stream (not on the stack). The terminating character
4634: is a blank by default. If it is not a blank, it is shown in @code{<>}
4635: quotes.
4636: @end table
4637:
4638: @comment ----------------------------------------------
4639: @node Case insensitivity, Comments, Notation, Words
4640: @section Case insensitivity
4641: @cindex case sensitivity
4642: @cindex upper and lower case
4643:
4644: Gforth is case-insensitive; you can enter definitions and invoke
4645: Standard words using upper, lower or mixed case (however,
4646: @pxref{core-idef, Implementation-defined options, Implementation-defined
4647: options}).
4648:
4649: ANS Forth only @i{requires} implementations to recognise Standard words
4650: when they are typed entirely in upper case. Therefore, a Standard
4651: program must use upper case for all Standard words. You can use whatever
4652: case you like for words that you define, but in a Standard program you
4653: have to use the words in the same case that you defined them.
4654:
4655: Gforth supports case sensitivity through @code{table}s (case-sensitive
4656: wordlists, @pxref{Word Lists}).
4657:
4658: Two people have asked how to convert Gforth to be case-sensitive; while
4659: we think this is a bad idea, you can change all wordlists into tables
4660: like this:
4661:
4662: @example
4663: ' table-find forth-wordlist wordlist-map @ !
4664: @end example
4665:
4666: Note that you now have to type the predefined words in the same case
4667: that we defined them, which are varying. You may want to convert them
4668: to your favourite case before doing this operation (I won't explain how,
4669: because if you are even contemplating doing this, you'd better have
4670: enough knowledge of Forth systems to know this already).
4671:
4672: @node Comments, Boolean Flags, Case insensitivity, Words
4673: @section Comments
4674: @cindex comments
4675:
4676: Forth supports two styles of comment; the traditional @i{in-line} comment,
4677: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
4678:
4679:
4680: doc-(
4681: doc-\
4682: doc-\G
4683:
4684:
4685: @node Boolean Flags, Arithmetic, Comments, Words
4686: @section Boolean Flags
4687: @cindex Boolean flags
4688:
4689: A Boolean flag is cell-sized. A cell with all bits clear represents the
4690: flag @code{false} and a flag with all bits set represents the flag
4691: @code{true}. Words that check a flag (for example, @code{IF}) will treat
4692: a cell that has @i{any} bit set as @code{true}.
4693: @c on and off to Memory?
4694: @c true and false to "Bitwise operations" or "Numeric comparison"?
4695:
4696: doc-true
4697: doc-false
4698: doc-on
4699: doc-off
4700:
4701:
4702: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
4703: @section Arithmetic
4704: @cindex arithmetic words
4705:
4706: @cindex division with potentially negative operands
4707: Forth arithmetic is not checked, i.e., you will not hear about integer
4708: overflow on addition or multiplication, you may hear about division by
4709: zero if you are lucky. The operator is written after the operands, but
4710: the operands are still in the original order. I.e., the infix @code{2-1}
4711: corresponds to @code{2 1 -}. Forth offers a variety of division
4712: operators. If you perform division with potentially negative operands,
4713: you do not want to use @code{/} or @code{/mod} with its undefined
4714: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4715: former, @pxref{Mixed precision}).
4716: @comment TODO discuss the different division forms and the std approach
4717:
4718: @menu
4719: * Single precision::
4720: * Double precision:: Double-cell integer arithmetic
4721: * Bitwise operations::
4722: * Numeric comparison::
4723: * Mixed precision:: Operations with single and double-cell integers
4724: * Floating Point::
4725: @end menu
4726:
4727: @node Single precision, Double precision, Arithmetic, Arithmetic
4728: @subsection Single precision
4729: @cindex single precision arithmetic words
4730:
4731: @c !! cell undefined
4732:
4733: By default, numbers in Forth are single-precision integers that are one
4734: cell in size. They can be signed or unsigned, depending upon how you
4735: treat them. For the rules used by the text interpreter for recognising
4736: single-precision integers see @ref{Number Conversion}.
4737:
4738: These words are all defined for signed operands, but some of them also
4739: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
4740: @code{*}.
4741:
4742: doc-+
4743: doc-1+
4744: doc--
4745: doc-1-
4746: doc-*
4747: doc-/
4748: doc-mod
4749: doc-/mod
4750: doc-negate
4751: doc-abs
4752: doc-min
4753: doc-max
4754: doc-floored
4755:
4756:
4757: @node Double precision, Bitwise operations, Single precision, Arithmetic
4758: @subsection Double precision
4759: @cindex double precision arithmetic words
4760:
4761: For the rules used by the text interpreter for
4762: recognising double-precision integers, see @ref{Number Conversion}.
4763:
4764: A double precision number is represented by a cell pair, with the most
4765: significant cell at the TOS. It is trivial to convert an unsigned single
4766: to a double: simply push a @code{0} onto the TOS. Since numbers are
4767: represented by Gforth using 2's complement arithmetic, converting a
4768: signed single to a (signed) double requires sign-extension across the
4769: most significant cell. This can be achieved using @code{s>d}. The moral
4770: of the story is that you cannot convert a number without knowing whether
4771: it represents an unsigned or a signed number.
4772:
4773: These words are all defined for signed operands, but some of them also
4774: work for unsigned numbers: @code{d+}, @code{d-}.
4775:
4776: doc-s>d
4777: doc-d>s
4778: doc-d+
4779: doc-d-
4780: doc-dnegate
4781: doc-dabs
4782: doc-dmin
4783: doc-dmax
4784:
4785:
4786: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
4787: @subsection Bitwise operations
4788: @cindex bitwise operation words
4789:
4790:
4791: doc-and
4792: doc-or
4793: doc-xor
4794: doc-invert
4795: doc-lshift
4796: doc-rshift
4797: doc-2*
4798: doc-d2*
4799: doc-2/
4800: doc-d2/
4801:
4802:
4803: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
4804: @subsection Numeric comparison
4805: @cindex numeric comparison words
4806:
4807: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
4808: d0= d0<>}) work for for both signed and unsigned numbers.
4809:
4810: doc-<
4811: doc-<=
4812: doc-<>
4813: doc-=
4814: doc->
4815: doc->=
4816:
4817: doc-0<
4818: doc-0<=
4819: doc-0<>
4820: doc-0=
4821: doc-0>
4822: doc-0>=
4823:
4824: doc-u<
4825: doc-u<=
4826: @c u<> and u= exist but are the same as <> and =
4827: @c doc-u<>
4828: @c doc-u=
4829: doc-u>
4830: doc-u>=
4831:
4832: doc-within
4833:
4834: doc-d<
4835: doc-d<=
4836: doc-d<>
4837: doc-d=
4838: doc-d>
4839: doc-d>=
4840:
4841: doc-d0<
4842: doc-d0<=
4843: doc-d0<>
4844: doc-d0=
4845: doc-d0>
4846: doc-d0>=
4847:
4848: doc-du<
4849: doc-du<=
4850: @c du<> and du= exist but are the same as d<> and d=
4851: @c doc-du<>
4852: @c doc-du=
4853: doc-du>
4854: doc-du>=
4855:
4856:
4857: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
4858: @subsection Mixed precision
4859: @cindex mixed precision arithmetic words
4860:
4861:
4862: doc-m+
4863: doc-*/
4864: doc-*/mod
4865: doc-m*
4866: doc-um*
4867: doc-m*/
4868: doc-um/mod
4869: doc-fm/mod
4870: doc-sm/rem
4871:
4872:
4873: @node Floating Point, , Mixed precision, Arithmetic
4874: @subsection Floating Point
4875: @cindex floating point arithmetic words
4876:
4877: For the rules used by the text interpreter for
4878: recognising floating-point numbers see @ref{Number Conversion}.
4879:
4880: Gforth has a separate floating point stack, but the documentation uses
4881: the unified notation.@footnote{It's easy to generate the separate
4882: notation from that by just separating the floating-point numbers out:
4883: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
4884: r3 )}.}
4885:
4886: @cindex floating-point arithmetic, pitfalls
4887: Floating point numbers have a number of unpleasant surprises for the
4888: unwary (e.g., floating point addition is not associative) and even a few
4889: for the wary. You should not use them unless you know what you are doing
4890: or you don't care that the results you get are totally bogus. If you
4891: want to learn about the problems of floating point numbers (and how to
4892: avoid them), you might start with @cite{David Goldberg,
4893: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
4894: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
4895: Surveys 23(1):5@minus{}48, March 1991}.
4896:
4897:
4898: doc-d>f
4899: doc-f>d
4900: doc-f+
4901: doc-f-
4902: doc-f*
4903: doc-f/
4904: doc-fnegate
4905: doc-fabs
4906: doc-fmax
4907: doc-fmin
4908: doc-floor
4909: doc-fround
4910: doc-f**
4911: doc-fsqrt
4912: doc-fexp
4913: doc-fexpm1
4914: doc-fln
4915: doc-flnp1
4916: doc-flog
4917: doc-falog
4918: doc-f2*
4919: doc-f2/
4920: doc-1/f
4921: doc-precision
4922: doc-set-precision
4923:
4924: @cindex angles in trigonometric operations
4925: @cindex trigonometric operations
4926: Angles in floating point operations are given in radians (a full circle
4927: has 2 pi radians).
4928:
4929: doc-fsin
4930: doc-fcos
4931: doc-fsincos
4932: doc-ftan
4933: doc-fasin
4934: doc-facos
4935: doc-fatan
4936: doc-fatan2
4937: doc-fsinh
4938: doc-fcosh
4939: doc-ftanh
4940: doc-fasinh
4941: doc-facosh
4942: doc-fatanh
4943: doc-pi
4944:
4945: @cindex equality of floats
4946: @cindex floating-point comparisons
4947: One particular problem with floating-point arithmetic is that comparison
4948: for equality often fails when you would expect it to succeed. For this
4949: reason approximate equality is often preferred (but you still have to
4950: know what you are doing). Also note that IEEE NaNs may compare
4951: differently from what you might expect. The comparison words are:
4952:
4953: doc-f~rel
4954: doc-f~abs
4955: doc-f~
4956: doc-f=
4957: doc-f<>
4958:
4959: doc-f<
4960: doc-f<=
4961: doc-f>
4962: doc-f>=
4963:
4964: doc-f0<
4965: doc-f0<=
4966: doc-f0<>
4967: doc-f0=
4968: doc-f0>
4969: doc-f0>=
4970:
4971:
4972: @node Stack Manipulation, Memory, Arithmetic, Words
4973: @section Stack Manipulation
4974: @cindex stack manipulation words
4975:
4976: @cindex floating-point stack in the standard
4977: Gforth maintains a number of separate stacks:
4978:
4979: @cindex data stack
4980: @cindex parameter stack
4981: @itemize @bullet
4982: @item
4983: A data stack (also known as the @dfn{parameter stack}) -- for
4984: characters, cells, addresses, and double cells.
4985:
4986: @cindex floating-point stack
4987: @item
4988: A floating point stack -- for holding floating point (FP) numbers.
4989:
4990: @cindex return stack
4991: @item
4992: A return stack -- for holding the return addresses of colon
4993: definitions and other (non-FP) data.
4994:
4995: @cindex locals stack
4996: @item
4997: A locals stack -- for holding local variables.
4998: @end itemize
4999:
5000: @menu
5001: * Data stack::
5002: * Floating point stack::
5003: * Return stack::
5004: * Locals stack::
5005: * Stack pointer manipulation::
5006: @end menu
5007:
5008: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
5009: @subsection Data stack
5010: @cindex data stack manipulation words
5011: @cindex stack manipulations words, data stack
5012:
5013:
5014: doc-drop
5015: doc-nip
5016: doc-dup
5017: doc-over
5018: doc-tuck
5019: doc-swap
5020: doc-pick
5021: doc-rot
5022: doc--rot
5023: doc-?dup
5024: doc-roll
5025: doc-2drop
5026: doc-2nip
5027: doc-2dup
5028: doc-2over
5029: doc-2tuck
5030: doc-2swap
5031: doc-2rot
5032:
5033:
5034: @node Floating point stack, Return stack, Data stack, Stack Manipulation
5035: @subsection Floating point stack
5036: @cindex floating-point stack manipulation words
5037: @cindex stack manipulation words, floating-point stack
5038:
5039: Whilst every sane Forth has a separate floating-point stack, it is not
5040: strictly required; an ANS Forth system could theoretically keep
5041: floating-point numbers on the data stack. As an additional difficulty,
5042: you don't know how many cells a floating-point number takes. It is
5043: reportedly possible to write words in a way that they work also for a
5044: unified stack model, but we do not recommend trying it. Instead, just
5045: say that your program has an environmental dependency on a separate
5046: floating-point stack.
5047:
5048: doc-floating-stack
5049:
5050: doc-fdrop
5051: doc-fnip
5052: doc-fdup
5053: doc-fover
5054: doc-ftuck
5055: doc-fswap
5056: doc-fpick
5057: doc-frot
5058:
5059:
5060: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
5061: @subsection Return stack
5062: @cindex return stack manipulation words
5063: @cindex stack manipulation words, return stack
5064:
5065: @cindex return stack and locals
5066: @cindex locals and return stack
5067: A Forth system is allowed to keep local variables on the
5068: return stack. This is reasonable, as local variables usually eliminate
5069: the need to use the return stack explicitly. So, if you want to produce
5070: a standard compliant program and you are using local variables in a
5071: word, forget about return stack manipulations in that word (refer to the
5072: standard document for the exact rules).
5073:
5074: doc->r
5075: doc-r>
5076: doc-r@
5077: doc-rdrop
5078: doc-2>r
5079: doc-2r>
5080: doc-2r@
5081: doc-2rdrop
5082:
5083:
5084: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
5085: @subsection Locals stack
5086:
5087: Gforth uses an extra locals stack. It is described, along with the
5088: reasons for its existence, in @ref{Locals implementation}.
5089:
5090: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
5091: @subsection Stack pointer manipulation
5092: @cindex stack pointer manipulation words
5093:
5094: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
5095: doc-sp0
5096: doc-sp@
5097: doc-sp!
5098: doc-fp0
5099: doc-fp@
5100: doc-fp!
5101: doc-rp0
5102: doc-rp@
5103: doc-rp!
5104: doc-lp0
5105: doc-lp@
5106: doc-lp!
5107:
5108:
5109: @node Memory, Control Structures, Stack Manipulation, Words
5110: @section Memory
5111: @cindex memory words
5112:
5113: @menu
5114: * Memory model::
5115: * Dictionary allocation::
5116: * Heap Allocation::
5117: * Memory Access::
5118: * Address arithmetic::
5119: * Memory Blocks::
5120: @end menu
5121:
5122: In addition to the standard Forth memory allocation words, there is also
5123: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
5124: garbage collector}.
5125:
5126: @node Memory model, Dictionary allocation, Memory, Memory
5127: @subsection ANS Forth and Gforth memory models
5128:
5129: @c The ANS Forth description is a mess (e.g., is the heap part of
5130: @c the dictionary?), so let's not stick to closely with it.
5131:
5132: ANS Forth considers a Forth system as consisting of several address
5133: spaces, of which only @dfn{data space} is managed and accessible with
5134: the memory words. Memory not necessarily in data space includes the
5135: stacks, the code (called code space) and the headers (called name
5136: space). In Gforth everything is in data space, but the code for the
5137: primitives is usually read-only.
5138:
5139: Data space is divided into a number of areas: The (data space portion of
5140: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
5141: refer to the search data structure embodied in word lists and headers,
5142: because it is used for looking up names, just as you would in a
5143: conventional dictionary.}, the heap, and a number of system-allocated
5144: buffers.
5145:
5146: @cindex address arithmetic restrictions, ANS vs. Gforth
5147: @cindex contiguous regions, ANS vs. Gforth
5148: In ANS Forth data space is also divided into contiguous regions. You
5149: can only use address arithmetic within a contiguous region, not between
5150: them. Usually each allocation gives you one contiguous region, but the
5151: dictionary allocation words have additional rules (@pxref{Dictionary
5152: allocation}).
5153:
5154: Gforth provides one big address space, and address arithmetic can be
5155: performed between any addresses. However, in the dictionary headers or
5156: code are interleaved with data, so almost the only contiguous data space
5157: regions there are those described by ANS Forth as contiguous; but you
5158: can be sure that the dictionary is allocated towards increasing
5159: addresses even between contiguous regions. The memory order of
5160: allocations in the heap is platform-dependent (and possibly different
5161: from one run to the next).
5162:
5163:
5164: @node Dictionary allocation, Heap Allocation, Memory model, Memory
5165: @subsection Dictionary allocation
5166: @cindex reserving data space
5167: @cindex data space - reserving some
5168:
5169: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
5170: you want to deallocate X, you also deallocate everything
5171: allocated after X.
5172:
5173: @cindex contiguous regions in dictionary allocation
5174: The allocations using the words below are contiguous and grow the region
5175: towards increasing addresses. Other words that allocate dictionary
5176: memory of any kind (i.e., defining words including @code{:noname}) end
5177: the contiguous region and start a new one.
5178:
5179: In ANS Forth only @code{create}d words are guaranteed to produce an
5180: address that is the start of the following contiguous region. In
5181: particular, the cell allocated by @code{variable} is not guaranteed to
5182: be contiguous with following @code{allot}ed memory.
5183:
5184: You can deallocate memory by using @code{allot} with a negative argument
5185: (with some restrictions, see @code{allot}). For larger deallocations use
5186: @code{marker}.
5187:
5188:
5189: doc-here
5190: doc-unused
5191: doc-allot
5192: doc-c,
5193: doc-f,
5194: doc-,
5195: doc-2,
5196:
5197: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
5198: course you should allocate memory in an aligned way, too. I.e., before
5199: allocating allocating a cell, @code{here} must be cell-aligned, etc.
5200: The words below align @code{here} if it is not already. Basically it is
5201: only already aligned for a type, if the last allocation was a multiple
5202: of the size of this type and if @code{here} was aligned for this type
5203: before.
5204:
5205: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
5206: ANS Forth (@code{maxalign}ed in Gforth).
5207:
5208: doc-align
5209: doc-falign
5210: doc-sfalign
5211: doc-dfalign
5212: doc-maxalign
5213: doc-cfalign
5214:
5215:
5216: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
5217: @subsection Heap allocation
5218: @cindex heap allocation
5219: @cindex dynamic allocation of memory
5220: @cindex memory-allocation word set
5221:
5222: @cindex contiguous regions and heap allocation
5223: Heap allocation supports deallocation of allocated memory in any
5224: order. Dictionary allocation is not affected by it (i.e., it does not
5225: end a contiguous region). In Gforth, these words are implemented using
5226: the standard C library calls malloc(), free() and resize().
5227:
5228: The memory region produced by one invocation of @code{allocate} or
5229: @code{resize} is internally contiguous. There is no contiguity between
5230: such a region and any other region (including others allocated from the
5231: heap).
5232:
5233: doc-allocate
5234: doc-free
5235: doc-resize
5236:
5237:
5238: @node Memory Access, Address arithmetic, Heap Allocation, Memory
5239: @subsection Memory Access
5240: @cindex memory access words
5241:
5242: doc-@
5243: doc-!
5244: doc-+!
5245: doc-c@
5246: doc-c!
5247: doc-2@
5248: doc-2!
5249: doc-f@
5250: doc-f!
5251: doc-sf@
5252: doc-sf!
5253: doc-df@
5254: doc-df!
5255:
5256:
5257: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5258: @subsection Address arithmetic
5259: @cindex address arithmetic words
5260:
5261: Address arithmetic is the foundation on which you can build data
5262: structures like arrays, records (@pxref{Structures}) and objects
5263: (@pxref{Object-oriented Forth}).
5264:
5265: @cindex address unit
5266: @cindex au (address unit)
5267: ANS Forth does not specify the sizes of the data types. Instead, it
5268: offers a number of words for computing sizes and doing address
5269: arithmetic. Address arithmetic is performed in terms of address units
5270: (aus); on most systems the address unit is one byte. Note that a
5271: character may have more than one au, so @code{chars} is no noop (on
5272: platforms where it is a noop, it compiles to nothing).
5273:
5274: The basic address arithmetic words are @code{+} and @code{-}. E.g., if
5275: you have the address of a cell, perform @code{1 cells +}, and you will
5276: have the address of the next cell.
5277:
5278: @cindex contiguous regions and address arithmetic
5279: In ANS Forth you can perform address arithmetic only within a contiguous
5280: region, i.e., if you have an address into one region, you can only add
5281: and subtract such that the result is still within the region; you can
5282: only subtract or compare addresses from within the same contiguous
5283: region. Reasons: several contiguous regions can be arranged in memory
5284: in any way; on segmented systems addresses may have unusual
5285: representations, such that address arithmetic only works within a
5286: region. Gforth provides a few more guarantees (linear address space,
5287: dictionary grows upwards), but in general I have found it easy to stay
5288: within contiguous regions (exception: computing and comparing to the
5289: address just beyond the end of an array).
5290:
5291: @cindex alignment of addresses for types
5292: ANS Forth also defines words for aligning addresses for specific
5293: types. Many computers require that accesses to specific data types
5294: must only occur at specific addresses; e.g., that cells may only be
5295: accessed at addresses divisible by 4. Even if a machine allows unaligned
5296: accesses, it can usually perform aligned accesses faster.
5297:
5298: For the performance-conscious: alignment operations are usually only
5299: necessary during the definition of a data structure, not during the
5300: (more frequent) accesses to it.
5301:
5302: ANS Forth defines no words for character-aligning addresses. This is not
5303: an oversight, but reflects the fact that addresses that are not
5304: char-aligned have no use in the standard and therefore will not be
5305: created.
5306:
5307: @cindex @code{CREATE} and alignment
5308: ANS Forth guarantees that addresses returned by @code{CREATE}d words
5309: are cell-aligned; in addition, Gforth guarantees that these addresses
5310: are aligned for all purposes.
5311:
5312: Note that the ANS Forth word @code{char} has nothing to do with address
5313: arithmetic.
5314:
5315:
5316: doc-chars
5317: doc-char+
5318: doc-cells
5319: doc-cell+
5320: doc-cell
5321: doc-aligned
5322: doc-floats
5323: doc-float+
5324: doc-float
5325: doc-faligned
5326: doc-sfloats
5327: doc-sfloat+
5328: doc-sfaligned
5329: doc-dfloats
5330: doc-dfloat+
5331: doc-dfaligned
5332: doc-maxaligned
5333: doc-cfaligned
5334: doc-address-unit-bits
5335:
5336:
5337: @node Memory Blocks, , Address arithmetic, Memory
5338: @subsection Memory Blocks
5339: @cindex memory block words
5340: @cindex character strings - moving and copying
5341:
5342: Memory blocks often represent character strings; For ways of storing
5343: character strings in memory see @ref{String Formats}. For other
5344: string-processing words see @ref{Displaying characters and strings}.
5345:
5346: A few of these words work on address unit blocks. In that case, you
5347: usually have to insert @code{CHARS} before the word when working on
5348: character strings. Most words work on character blocks, and expect a
5349: char-aligned address.
5350:
5351: When copying characters between overlapping memory regions, use
5352: @code{chars move} or choose carefully between @code{cmove} and
5353: @code{cmove>}.
5354:
5355: doc-move
5356: doc-erase
5357: doc-cmove
5358: doc-cmove>
5359: doc-fill
5360: doc-blank
5361: doc-compare
5362: doc-search
5363: doc--trailing
5364: doc-/string
5365: doc-bounds
5366:
5367: @comment TODO examples
5368:
5369:
5370: @node Control Structures, Defining Words, Memory, Words
5371: @section Control Structures
5372: @cindex control structures
5373:
5374: Control structures in Forth cannot be used interpretively, only in a
5375: colon definition@footnote{To be precise, they have no interpretation
5376: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5377: not like this limitation, but have not seen a satisfying way around it
5378: yet, although many schemes have been proposed.
5379:
5380: @menu
5381: * Selection:: IF ... ELSE ... ENDIF
5382: * Simple Loops:: BEGIN ...
5383: * Counted Loops:: DO
5384: * Arbitrary control structures::
5385: * Calls and returns::
5386: * Exception Handling::
5387: @end menu
5388:
5389: @node Selection, Simple Loops, Control Structures, Control Structures
5390: @subsection Selection
5391: @cindex selection control structures
5392: @cindex control structures for selection
5393:
5394: @cindex @code{IF} control structure
5395: @example
5396: @i{flag}
5397: IF
5398: @i{code}
5399: ENDIF
5400: @end example
5401: @noindent
5402:
5403: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5404: with any bit set represents truth) @i{code} is executed.
5405:
5406: @example
5407: @i{flag}
5408: IF
5409: @i{code1}
5410: ELSE
5411: @i{code2}
5412: ENDIF
5413: @end example
5414:
5415: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5416: executed.
5417:
5418: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5419: standard, and @code{ENDIF} is not, although it is quite popular. We
5420: recommend using @code{ENDIF}, because it is less confusing for people
5421: who also know other languages (and is not prone to reinforcing negative
5422: prejudices against Forth in these people). Adding @code{ENDIF} to a
5423: system that only supplies @code{THEN} is simple:
5424: @example
5425: : ENDIF POSTPONE then ; immediate
5426: @end example
5427:
5428: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5429: (adv.)} has the following meanings:
5430: @quotation
5431: ... 2b: following next after in order ... 3d: as a necessary consequence
5432: (if you were there, then you saw them).
5433: @end quotation
5434: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5435: and many other programming languages has the meaning 3d.]
5436:
5437: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
5438: you can avoid using @code{?dup}. Using these alternatives is also more
5439: efficient than using @code{?dup}. Definitions in ANS Forth
5440: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5441: @file{compat/control.fs}.
5442:
5443: @cindex @code{CASE} control structure
5444: @example
5445: @i{n}
5446: CASE
5447: @i{n1} OF @i{code1} ENDOF
5448: @i{n2} OF @i{code2} ENDOF
5449: @dots{}
5450: ( n ) @i{default-code} ( n )
5451: ENDCASE
5452: @end example
5453:
5454: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}. If no
5455: @i{ni} matches, the optional @i{default-code} is executed. The optional
5456: default case can be added by simply writing the code after the last
5457: @code{ENDOF}. It may use @i{n}, which is on top of the stack, but must
5458: not consume it.
5459:
5460: @progstyle
5461: To keep the code understandable, you should ensure that on all paths
5462: through a selection construct the stack is changed in the same way
5463: (wrt. number and types of stack items consumed and pushed).
5464:
5465: @node Simple Loops, Counted Loops, Selection, Control Structures
5466: @subsection Simple Loops
5467: @cindex simple loops
5468: @cindex loops without count
5469:
5470: @cindex @code{WHILE} loop
5471: @example
5472: BEGIN
5473: @i{code1}
5474: @i{flag}
5475: WHILE
5476: @i{code2}
5477: REPEAT
5478: @end example
5479:
5480: @i{code1} is executed and @i{flag} is computed. If it is true,
5481: @i{code2} is executed and the loop is restarted; If @i{flag} is
5482: false, execution continues after the @code{REPEAT}.
5483:
5484: @cindex @code{UNTIL} loop
5485: @example
5486: BEGIN
5487: @i{code}
5488: @i{flag}
5489: UNTIL
5490: @end example
5491:
5492: @i{code} is executed. The loop is restarted if @code{flag} is false.
5493:
5494: @progstyle
5495: To keep the code understandable, a complete iteration of the loop should
5496: not change the number and types of the items on the stacks.
5497:
5498: @cindex endless loop
5499: @cindex loops, endless
5500: @example
5501: BEGIN
5502: @i{code}
5503: AGAIN
5504: @end example
5505:
5506: This is an endless loop.
5507:
5508: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5509: @subsection Counted Loops
5510: @cindex counted loops
5511: @cindex loops, counted
5512: @cindex @code{DO} loops
5513:
5514: The basic counted loop is:
5515: @example
5516: @i{limit} @i{start}
5517: ?DO
5518: @i{body}
5519: LOOP
5520: @end example
5521:
5522: This performs one iteration for every integer, starting from @i{start}
5523: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
5524: accessed with @code{i}. For example, the loop:
5525: @example
5526: 10 0 ?DO
5527: i .
5528: LOOP
5529: @end example
5530: @noindent
5531: prints @code{0 1 2 3 4 5 6 7 8 9}
5532:
5533: The index of the innermost loop can be accessed with @code{i}, the index
5534: of the next loop with @code{j}, and the index of the third loop with
5535: @code{k}.
5536:
5537:
5538: doc-i
5539: doc-j
5540: doc-k
5541:
5542:
5543: The loop control data are kept on the return stack, so there are some
5544: restrictions on mixing return stack accesses and counted loop words. In
5545: particuler, if you put values on the return stack outside the loop, you
5546: cannot read them inside the loop@footnote{well, not in a way that is
5547: portable.}. If you put values on the return stack within a loop, you
5548: have to remove them before the end of the loop and before accessing the
5549: index of the loop.
5550:
5551: There are several variations on the counted loop:
5552:
5553: @itemize @bullet
5554: @item
5555: @code{LEAVE} leaves the innermost counted loop immediately; execution
5556: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5557:
5558: @example
5559: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5560: @end example
5561: prints @code{0 1 2 3}
5562:
5563:
5564: @item
5565: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5566: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5567: return stack so @code{EXIT} can get to its return address. For example:
5568:
5569: @example
5570: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5571: @end example
5572: prints @code{0 1 2 3}
5573:
5574:
5575: @item
5576: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
5577: (and @code{LOOP} iterates until they become equal by wrap-around
5578: arithmetic). This behaviour is usually not what you want. Therefore,
5579: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
5580: @code{?DO}), which do not enter the loop if @i{start} is greater than
5581: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
5582: unsigned loop parameters.
5583:
5584: @item
5585: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5586: the loop, independent of the loop parameters. Do not use @code{DO}, even
5587: if you know that the loop is entered in any case. Such knowledge tends
5588: to become invalid during maintenance of a program, and then the
5589: @code{DO} will make trouble.
5590:
5591: @item
5592: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5593: index by @i{n} instead of by 1. The loop is terminated when the border
5594: between @i{limit-1} and @i{limit} is crossed. E.g.:
5595:
5596: @example
5597: 4 0 +DO i . 2 +LOOP
5598: @end example
5599: @noindent
5600: prints @code{0 2}
5601:
5602: @example
5603: 4 1 +DO i . 2 +LOOP
5604: @end example
5605: @noindent
5606: prints @code{1 3}
5607:
5608: @item
5609: @cindex negative increment for counted loops
5610: @cindex counted loops with negative increment
5611: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
5612:
5613: @example
5614: -1 0 ?DO i . -1 +LOOP
5615: @end example
5616: @noindent
5617: prints @code{0 -1}
5618:
5619: @example
5620: 0 0 ?DO i . -1 +LOOP
5621: @end example
5622: prints nothing.
5623:
5624: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5625: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5626: index by @i{u} each iteration. The loop is terminated when the border
5627: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
5628: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5629:
5630: @example
5631: -2 0 -DO i . 1 -LOOP
5632: @end example
5633: @noindent
5634: prints @code{0 -1}
5635:
5636: @example
5637: -1 0 -DO i . 1 -LOOP
5638: @end example
5639: @noindent
5640: prints @code{0}
5641:
5642: @example
5643: 0 0 -DO i . 1 -LOOP
5644: @end example
5645: @noindent
5646: prints nothing.
5647:
5648: @end itemize
5649:
5650: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
5651: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5652: for these words that uses only standard words is provided in
5653: @file{compat/loops.fs}.
5654:
5655:
5656: @cindex @code{FOR} loops
5657: Another counted loop is:
5658: @example
5659: @i{n}
5660: FOR
5661: @i{body}
5662: NEXT
5663: @end example
5664: This is the preferred loop of native code compiler writers who are too
5665: lazy to optimize @code{?DO} loops properly. This loop structure is not
5666: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5667: @code{i} produces values starting with @i{n} and ending with 0. Other
5668: Forth systems may behave differently, even if they support @code{FOR}
5669: loops. To avoid problems, don't use @code{FOR} loops.
5670:
5671: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5672: @subsection Arbitrary control structures
5673: @cindex control structures, user-defined
5674:
5675: @cindex control-flow stack
5676: ANS Forth permits and supports using control structures in a non-nested
5677: way. Information about incomplete control structures is stored on the
5678: control-flow stack. This stack may be implemented on the Forth data
5679: stack, and this is what we have done in Gforth.
5680:
5681: @cindex @code{orig}, control-flow stack item
5682: @cindex @code{dest}, control-flow stack item
5683: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5684: entry represents a backward branch target. A few words are the basis for
5685: building any control structure possible (except control structures that
5686: need storage, like calls, coroutines, and backtracking).
5687:
5688:
5689: doc-if
5690: doc-ahead
5691: doc-then
5692: doc-begin
5693: doc-until
5694: doc-again
5695: doc-cs-pick
5696: doc-cs-roll
5697:
5698:
5699: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5700: manipulate the control-flow stack in a portable way. Without them, you
5701: would need to know how many stack items are occupied by a control-flow
5702: entry (many systems use one cell. In Gforth they currently take three,
5703: but this may change in the future).
5704:
5705: Some standard control structure words are built from these words:
5706:
5707:
5708: doc-else
5709: doc-while
5710: doc-repeat
5711:
5712:
5713: @noindent
5714: Gforth adds some more control-structure words:
5715:
5716:
5717: doc-endif
5718: doc-?dup-if
5719: doc-?dup-0=-if
5720:
5721:
5722: @noindent
5723: Counted loop words constitute a separate group of words:
5724:
5725:
5726: doc-?do
5727: doc-+do
5728: doc-u+do
5729: doc--do
5730: doc-u-do
5731: doc-do
5732: doc-for
5733: doc-loop
5734: doc-+loop
5735: doc--loop
5736: doc-next
5737: doc-leave
5738: doc-?leave
5739: doc-unloop
5740: doc-done
5741:
5742:
5743: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5744: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
5745: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5746: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5747: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5748: resolved (by using one of the loop-ending words or @code{DONE}).
5749:
5750: @noindent
5751: Another group of control structure words are:
5752:
5753:
5754: doc-case
5755: doc-endcase
5756: doc-of
5757: doc-endof
5758:
5759:
5760: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5761: @code{CS-ROLL}.
5762:
5763: @subsubsection Programming Style
5764: @cindex control structures programming style
5765: @cindex programming style, arbitrary control structures
5766:
5767: In order to ensure readability we recommend that you do not create
5768: arbitrary control structures directly, but define new control structure
5769: words for the control structure you want and use these words in your
5770: program. For example, instead of writing:
5771:
5772: @example
5773: BEGIN
5774: ...
5775: IF [ 1 CS-ROLL ]
5776: ...
5777: AGAIN THEN
5778: @end example
5779:
5780: @noindent
5781: we recommend defining control structure words, e.g.,
5782:
5783: @example
5784: : WHILE ( DEST -- ORIG DEST )
5785: POSTPONE IF
5786: 1 CS-ROLL ; immediate
5787:
5788: : REPEAT ( orig dest -- )
5789: POSTPONE AGAIN
5790: POSTPONE THEN ; immediate
5791: @end example
5792:
5793: @noindent
5794: and then using these to create the control structure:
5795:
5796: @example
5797: BEGIN
5798: ...
5799: WHILE
5800: ...
5801: REPEAT
5802: @end example
5803:
5804: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5805: @code{WHILE} are predefined, so in this example it would not be
5806: necessary to define them.
5807:
5808: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5809: @subsection Calls and returns
5810: @cindex calling a definition
5811: @cindex returning from a definition
5812:
5813: @cindex recursive definitions
5814: A definition can be called simply be writing the name of the definition
5815: to be called. Normally a definition is invisible during its own
5816: definition. If you want to write a directly recursive definition, you
5817: can use @code{recursive} to make the current definition visible, or
5818: @code{recurse} to call the current definition directly.
5819:
5820:
5821: doc-recursive
5822: doc-recurse
5823:
5824:
5825: @comment TODO add example of the two recursion methods
5826: @quotation
5827: @progstyle
5828: I prefer using @code{recursive} to @code{recurse}, because calling the
5829: definition by name is more descriptive (if the name is well-chosen) than
5830: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5831: implementation, it is much better to read (and think) ``now sort the
5832: partitions'' than to read ``now do a recursive call''.
5833: @end quotation
5834:
5835: For mutual recursion, use @code{Defer}red words, like this:
5836:
5837: @example
5838: Defer foo
5839:
5840: : bar ( ... -- ... )
5841: ... foo ... ;
5842:
5843: :noname ( ... -- ... )
5844: ... bar ... ;
5845: IS foo
5846: @end example
5847:
5848: Deferred words are discussed in more detail in @ref{Deferred words}.
5849:
5850: The current definition returns control to the calling definition when
5851: the end of the definition is reached or @code{EXIT} is encountered.
5852:
5853: doc-exit
5854: doc-;s
5855:
5856:
5857: @node Exception Handling, , Calls and returns, Control Structures
5858: @subsection Exception Handling
5859: @cindex exceptions
5860:
5861: @c quit is a very bad idea for error handling,
5862: @c because it does not translate into a THROW
5863: @c it also does not belong into this chapter
5864:
5865: If a word detects an error condition that it cannot handle, it can
5866: @code{throw} an exception. In the simplest case, this will terminate
5867: your program, and report an appropriate error.
5868:
5869: doc-throw
5870:
5871: @code{Throw} consumes a cell-sized error number on the stack. There are
5872: some predefined error numbers in ANS Forth (see @file{errors.fs}). In
5873: Gforth (and most other systems) you can use the iors produced by various
5874: words as error numbers (e.g., a typical use of @code{allocate} is
5875: @code{allocate throw}). Gforth also provides the word @code{exception}
5876: to define your own error numbers (with decent error reporting); an ANS
5877: Forth version of this word (but without the error messages) is available
5878: in @code{compat/except.fs}. And finally, you can use your own error
5879: numbers (anything outside the range -4095..0), but won't get nice error
5880: messages, only numbers. For example, try:
5881:
5882: @example
5883: -10 throw \ ANS defined
5884: -267 throw \ system defined
5885: s" my error" exception throw \ user defined
5886: 7 throw \ arbitrary number
5887: @end example
5888:
5889: doc---exception-exception
5890:
5891: A common idiom to @code{THROW} a specific error if a flag is true is
5892: this:
5893:
5894: @example
5895: @code{( flag ) 0<> @i{errno} and throw}
5896: @end example
5897:
5898: Your program can provide exception handlers to catch exceptions. An
5899: exception handler can be used to correct the problem, or to clean up
5900: some data structures and just throw the exception to the next exception
5901: handler. Note that @code{throw} jumps to the dynamically innermost
5902: exception handler. The system's exception handler is outermost, and just
5903: prints an error and restarts command-line interpretation (or, in batch
5904: mode (i.e., while processing the shell command line), leaves Gforth).
5905:
5906: The ANS Forth way to catch exceptions is @code{catch}:
5907:
5908: doc-catch
5909:
5910: The most common use of exception handlers is to clean up the state when
5911: an error happens. E.g.,
5912:
5913: @example
5914: base @ >r hex \ actually the hex should be inside foo, or we h
5915: ['] foo catch ( nerror|0 )
5916: r> base !
5917: ( nerror|0 ) throw \ pass it on
5918: @end example
5919:
5920: A use of @code{catch} for handling the error @code{myerror} might look
5921: like this:
5922:
5923: @example
5924: ['] foo catch
5925: CASE
5926: myerror OF ... ( do something about it ) ENDOF
5927: dup throw \ default: pass other errors on, do nothing on non-errors
5928: ENDCASE
5929: @end example
5930:
5931: Having to wrap the code into a separate word is often cumbersome,
5932: therefore Gforth provides an alternative syntax:
5933:
5934: @example
5935: TRY
5936: @i{code1}
5937: RECOVER \ optional
5938: @i{code2} \ optional
5939: ENDTRY
5940: @end example
5941:
5942: This performs @i{Code1}. If @i{code1} completes normally, execution
5943: continues after the @code{endtry}. If @i{Code1} throws, the stacks are
5944: reset to the state during @code{try}, the throw value is pushed on the
5945: data stack, and execution constinues at @i{code2}, and finally falls
5946: through the @code{endtry} into the following code.
5947:
5948: doc-try
5949: doc-recover
5950: doc-endtry
5951:
5952: The cleanup example from above in this syntax:
5953:
5954: @example
5955: base @ >r TRY
5956: hex foo \ now the hex is placed correctly
5957: 0 \ value for throw
5958: RECOVER ENDTRY
5959: r> base ! throw
5960: @end example
5961:
5962: And here's the error handling example:
5963:
5964: @example
5965: TRY
5966: foo
5967: RECOVER
5968: CASE
5969: myerror OF ... ( do something about it ) ENDOF
5970: throw \ pass other errors on
5971: ENDCASE
5972: ENDTRY
5973: @end example
5974:
5975: @progstyle
5976: As usual, you should ensure that the stack depth is statically known at
5977: the end: either after the @code{throw} for passing on errors, or after
5978: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
5979: selection construct for handling the error).
5980:
5981: There are two alternatives to @code{throw}: @code{Abort"} is conditional
5982: and you can provide an error message. @code{Abort} just produces an
5983: ``Aborted'' error.
5984:
5985: The problem with these words is that exception handlers cannot
5986: differentiate between different @code{abort"}s; they just look like
5987: @code{-2 throw} to them (the error message cannot be accessed by
5988: standard programs). Similar @code{abort} looks like @code{-1 throw} to
5989: exception handlers.
5990:
5991: doc-abort"
5992: doc-abort
5993:
5994:
5995:
5996: @c -------------------------------------------------------------
5997: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
5998: @section Defining Words
5999: @cindex defining words
6000:
6001: Defining words are used to extend Forth by creating new entries in the dictionary.
6002:
6003: @menu
6004: * CREATE::
6005: * Variables:: Variables and user variables
6006: * Constants::
6007: * Values:: Initialised variables
6008: * Colon Definitions::
6009: * Anonymous Definitions:: Definitions without names
6010: * Supplying names:: Passing definition names as strings
6011: * User-defined Defining Words::
6012: * Deferred words:: Allow forward references
6013: * Aliases::
6014: @end menu
6015:
6016: @node CREATE, Variables, Defining Words, Defining Words
6017: @subsection @code{CREATE}
6018: @cindex simple defining words
6019: @cindex defining words, simple
6020:
6021: Defining words are used to create new entries in the dictionary. The
6022: simplest defining word is @code{CREATE}. @code{CREATE} is used like
6023: this:
6024:
6025: @example
6026: CREATE new-word1
6027: @end example
6028:
6029: @code{CREATE} is a parsing word, i.e., it takes an argument from the
6030: input stream (@code{new-word1} in our example). It generates a
6031: dictionary entry for @code{new-word1}. When @code{new-word1} is
6032: executed, all that it does is leave an address on the stack. The address
6033: represents the value of the data space pointer (@code{HERE}) at the time
6034: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
6035: associating a name with the address of a region of memory.
6036:
6037: doc-create
6038:
6039: Note that in ANS Forth guarantees only for @code{create} that its body
6040: is in dictionary data space (i.e., where @code{here}, @code{allot}
6041: etc. work, @pxref{Dictionary allocation}). Also, in ANS Forth only
6042: @code{create}d words can be modified with @code{does>}
6043: (@pxref{User-defined Defining Words}). And in ANS Forth @code{>body}
6044: can only be applied to @code{create}d words.
6045:
6046: By extending this example to reserve some memory in data space, we end
6047: up with something like a @i{variable}. Here are two different ways to do
6048: it:
6049:
6050: @example
6051: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
6052: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
6053: @end example
6054:
6055: The variable can be examined and modified using @code{@@} (``fetch'') and
6056: @code{!} (``store'') like this:
6057:
6058: @example
6059: new-word2 @@ . \ get address, fetch from it and display
6060: 1234 new-word2 ! \ new value, get address, store to it
6061: @end example
6062:
6063: @cindex arrays
6064: A similar mechanism can be used to create arrays. For example, an
6065: 80-character text input buffer:
6066:
6067: @example
6068: CREATE text-buf 80 chars allot
6069:
6070: text-buf 0 chars c@@ \ the 1st character (offset 0)
6071: text-buf 3 chars c@@ \ the 4th character (offset 3)
6072: @end example
6073:
6074: You can build arbitrarily complex data structures by allocating
6075: appropriate areas of memory. For further discussions of this, and to
6076: learn about some Gforth tools that make it easier,
6077: @xref{Structures}.
6078:
6079:
6080: @node Variables, Constants, CREATE, Defining Words
6081: @subsection Variables
6082: @cindex variables
6083:
6084: The previous section showed how a sequence of commands could be used to
6085: generate a variable. As a final refinement, the whole code sequence can
6086: be wrapped up in a defining word (pre-empting the subject of the next
6087: section), making it easier to create new variables:
6088:
6089: @example
6090: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
6091: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
6092:
6093: myvariableX foo \ variable foo starts off with an unknown value
6094: myvariable0 joe \ whilst joe is initialised to 0
6095:
6096: 45 3 * foo ! \ set foo to 135
6097: 1234 joe ! \ set joe to 1234
6098: 3 joe +! \ increment joe by 3.. to 1237
6099: @end example
6100:
6101: Not surprisingly, there is no need to define @code{myvariable}, since
6102: Forth already has a definition @code{Variable}. ANS Forth does not
6103: guarantee that a @code{Variable} is initialised when it is created
6104: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
6105: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
6106: like @code{myvariable0}). Forth also provides @code{2Variable} and
6107: @code{fvariable} for double and floating-point variables, respectively
6108: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
6109: store a boolean, you can use @code{on} and @code{off} to toggle its
6110: state.
6111:
6112: doc-variable
6113: doc-2variable
6114: doc-fvariable
6115:
6116: @cindex user variables
6117: @cindex user space
6118: The defining word @code{User} behaves in the same way as @code{Variable}.
6119: The difference is that it reserves space in @i{user (data) space} rather
6120: than normal data space. In a Forth system that has a multi-tasker, each
6121: task has its own set of user variables.
6122:
6123: doc-user
6124: @c doc-udp
6125: @c doc-uallot
6126:
6127: @comment TODO is that stuff about user variables strictly correct? Is it
6128: @comment just terminal tasks that have user variables?
6129: @comment should document tasker.fs (with some examples) elsewhere
6130: @comment in this manual, then expand on user space and user variables.
6131:
6132: @node Constants, Values, Variables, Defining Words
6133: @subsection Constants
6134: @cindex constants
6135:
6136: @code{Constant} allows you to declare a fixed value and refer to it by
6137: name. For example:
6138:
6139: @example
6140: 12 Constant INCHES-PER-FOOT
6141: 3E+08 fconstant SPEED-O-LIGHT
6142: @end example
6143:
6144: A @code{Variable} can be both read and written, so its run-time
6145: behaviour is to supply an address through which its current value can be
6146: manipulated. In contrast, the value of a @code{Constant} cannot be
6147: changed once it has been declared@footnote{Well, often it can be -- but
6148: not in a Standard, portable way. It's safer to use a @code{Value} (read
6149: on).} so it's not necessary to supply the address -- it is more
6150: efficient to return the value of the constant directly. That's exactly
6151: what happens; the run-time effect of a constant is to put its value on
6152: the top of the stack (You can find one
6153: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
6154:
6155: Forth also provides @code{2Constant} and @code{fconstant} for defining
6156: double and floating-point constants, respectively.
6157:
6158: doc-constant
6159: doc-2constant
6160: doc-fconstant
6161:
6162: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
6163: @c nac-> How could that not be true in an ANS Forth? You can't define a
6164: @c constant, use it and then delete the definition of the constant..
6165:
6166: @c anton->An ANS Forth system can compile a constant to a literal; On
6167: @c decompilation you would see only the number, just as if it had been used
6168: @c in the first place. The word will stay, of course, but it will only be
6169: @c used by the text interpreter (no run-time duties, except when it is
6170: @c POSTPONEd or somesuch).
6171:
6172: @c nac:
6173: @c I agree that it's rather deep, but IMO it is an important difference
6174: @c relative to other programming languages.. often it's annoying: it
6175: @c certainly changes my programming style relative to C.
6176:
6177: @c anton: In what way?
6178:
6179: Constants in Forth behave differently from their equivalents in other
6180: programming languages. In other languages, a constant (such as an EQU in
6181: assembler or a #define in C) only exists at compile-time; in the
6182: executable program the constant has been translated into an absolute
6183: number and, unless you are using a symbolic debugger, it's impossible to
6184: know what abstract thing that number represents. In Forth a constant has
6185: an entry in the header space and remains there after the code that uses
6186: it has been defined. In fact, it must remain in the dictionary since it
6187: has run-time duties to perform. For example:
6188:
6189: @example
6190: 12 Constant INCHES-PER-FOOT
6191: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
6192: @end example
6193:
6194: @cindex in-lining of constants
6195: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
6196: associated with the constant @code{INCHES-PER-FOOT}. If you use
6197: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
6198: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
6199: attempt to optimise constants by in-lining them where they are used. You
6200: can force Gforth to in-line a constant like this:
6201:
6202: @example
6203: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
6204: @end example
6205:
6206: If you use @code{see} to decompile @i{this} version of
6207: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
6208: longer present. To understand how this works, read
6209: @ref{Interpret/Compile states}, and @ref{Literals}.
6210:
6211: In-lining constants in this way might improve execution time
6212: fractionally, and can ensure that a constant is now only referenced at
6213: compile-time. However, the definition of the constant still remains in
6214: the dictionary. Some Forth compilers provide a mechanism for controlling
6215: a second dictionary for holding transient words such that this second
6216: dictionary can be deleted later in order to recover memory
6217: space. However, there is no standard way of doing this.
6218:
6219:
6220: @node Values, Colon Definitions, Constants, Defining Words
6221: @subsection Values
6222: @cindex values
6223:
6224: A @code{Value} behaves like a @code{Constant}, but it can be changed.
6225: @code{TO} is a parsing word that changes a @code{Values}. In Gforth
6226: (not in ANS Forth) you can access (and change) a @code{value} also with
6227: @code{>body}.
6228:
6229: Here are some
6230: examples:
6231:
6232: @example
6233: 12 Value APPLES \ Define APPLES with an initial value of 12
6234: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6235: 1 ' APPLES >body +! \ Increment APPLES. Non-standard usage.
6236: APPLES \ puts 35 on the top of the stack.
6237: @end example
6238:
6239: doc-value
6240: doc-to
6241:
6242:
6243:
6244: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6245: @subsection Colon Definitions
6246: @cindex colon definitions
6247:
6248: @example
6249: : name ( ... -- ... )
6250: word1 word2 word3 ;
6251: @end example
6252:
6253: @noindent
6254: Creates a word called @code{name} that, upon execution, executes
6255: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
6256:
6257: The explanation above is somewhat superficial. For simple examples of
6258: colon definitions see @ref{Your first definition}. For an in-depth
6259: discussion of some of the issues involved, @xref{Interpretation and
6260: Compilation Semantics}.
6261:
6262: doc-:
6263: doc-;
6264:
6265:
6266: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
6267: @subsection Anonymous Definitions
6268: @cindex colon definitions
6269: @cindex defining words without name
6270:
6271: Sometimes you want to define an @dfn{anonymous word}; a word without a
6272: name. You can do this with:
6273:
6274: doc-:noname
6275:
6276: This leaves the execution token for the word on the stack after the
6277: closing @code{;}. Here's an example in which a deferred word is
6278: initialised with an @code{xt} from an anonymous colon definition:
6279:
6280: @example
6281: Defer deferred
6282: :noname ( ... -- ... )
6283: ... ;
6284: IS deferred
6285: @end example
6286:
6287: @noindent
6288: Gforth provides an alternative way of doing this, using two separate
6289: words:
6290:
6291: doc-noname
6292: @cindex execution token of last defined word
6293: doc-lastxt
6294:
6295: @noindent
6296: The previous example can be rewritten using @code{noname} and
6297: @code{lastxt}:
6298:
6299: @example
6300: Defer deferred
6301: noname : ( ... -- ... )
6302: ... ;
6303: lastxt IS deferred
6304: @end example
6305:
6306: @noindent
6307: @code{noname} works with any defining word, not just @code{:}.
6308:
6309: @code{lastxt} also works when the last word was not defined as
6310: @code{noname}. It does not work for combined words, though. It also has
6311: the useful property that is is valid as soon as the header for a
6312: definition has been built. Thus:
6313:
6314: @example
6315: lastxt . : foo [ lastxt . ] ; ' foo .
6316: @end example
6317:
6318: @noindent
6319: prints 3 numbers; the last two are the same.
6320:
6321: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
6322: @subsection Supplying the name of a defined word
6323: @cindex names for defined words
6324: @cindex defining words, name given in a string
6325:
6326: By default, a defining word takes the name for the defined word from the
6327: input stream. Sometimes you want to supply the name from a string. You
6328: can do this with:
6329:
6330: doc-nextname
6331:
6332: For example:
6333:
6334: @example
6335: s" foo" nextname create
6336: @end example
6337:
6338: @noindent
6339: is equivalent to:
6340:
6341: @example
6342: create foo
6343: @end example
6344:
6345: @noindent
6346: @code{nextname} works with any defining word.
6347:
6348:
6349: @node User-defined Defining Words, Deferred words, Supplying names, Defining Words
6350: @subsection User-defined Defining Words
6351: @cindex user-defined defining words
6352: @cindex defining words, user-defined
6353:
6354: You can create a new defining word by wrapping defining-time code around
6355: an existing defining word and putting the sequence in a colon
6356: definition.
6357:
6358: @c anton: This example is very complex and leads in a quite different
6359: @c direction from the CREATE-DOES> stuff that follows. It should probably
6360: @c be done elsewhere, or as a subsubsection of this subsection (or as a
6361: @c subsection of Defining Words)
6362:
6363: For example, suppose that you have a word @code{stats} that
6364: gathers statistics about colon definitions given the @i{xt} of the
6365: definition, and you want every colon definition in your application to
6366: make a call to @code{stats}. You can define and use a new version of
6367: @code{:} like this:
6368:
6369: @example
6370: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6371: ... ; \ other code
6372:
6373: : my: : lastxt postpone literal ['] stats compile, ;
6374:
6375: my: foo + - ;
6376: @end example
6377:
6378: When @code{foo} is defined using @code{my:} these steps occur:
6379:
6380: @itemize @bullet
6381: @item
6382: @code{my:} is executed.
6383: @item
6384: The @code{:} within the definition (the one between @code{my:} and
6385: @code{lastxt}) is executed, and does just what it always does; it parses
6386: the input stream for a name, builds a dictionary header for the name
6387: @code{foo} and switches @code{state} from interpret to compile.
6388: @item
6389: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
6390: being defined -- @code{foo} -- onto the stack.
6391: @item
6392: The code that was produced by @code{postpone literal} is executed; this
6393: causes the value on the stack to be compiled as a literal in the code
6394: area of @code{foo}.
6395: @item
6396: The code @code{['] stats} compiles a literal into the definition of
6397: @code{my:}. When @code{compile,} is executed, that literal -- the
6398: execution token for @code{stats} -- is layed down in the code area of
6399: @code{foo} , following the literal@footnote{Strictly speaking, the
6400: mechanism that @code{compile,} uses to convert an @i{xt} into something
6401: in the code area is implementation-dependent. A threaded implementation
6402: might spit out the execution token directly whilst another
6403: implementation might spit out a native code sequence.}.
6404: @item
6405: At this point, the execution of @code{my:} is complete, and control
6406: returns to the text interpreter. The text interpreter is in compile
6407: state, so subsequent text @code{+ -} is compiled into the definition of
6408: @code{foo} and the @code{;} terminates the definition as always.
6409: @end itemize
6410:
6411: You can use @code{see} to decompile a word that was defined using
6412: @code{my:} and see how it is different from a normal @code{:}
6413: definition. For example:
6414:
6415: @example
6416: : bar + - ; \ like foo but using : rather than my:
6417: see bar
6418: : bar
6419: + - ;
6420: see foo
6421: : foo
6422: 107645672 stats + - ;
6423:
6424: \ use ' stats . to show that 107645672 is the xt for stats
6425: @end example
6426:
6427: You can use techniques like this to make new defining words in terms of
6428: @i{any} existing defining word.
6429:
6430:
6431: @cindex defining defining words
6432: @cindex @code{CREATE} ... @code{DOES>}
6433: If you want the words defined with your defining words to behave
6434: differently from words defined with standard defining words, you can
6435: write your defining word like this:
6436:
6437: @example
6438: : def-word ( "name" -- )
6439: CREATE @i{code1}
6440: DOES> ( ... -- ... )
6441: @i{code2} ;
6442:
6443: def-word name
6444: @end example
6445:
6446: @cindex child words
6447: This fragment defines a @dfn{defining word} @code{def-word} and then
6448: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6449: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6450: is not executed at this time. The word @code{name} is sometimes called a
6451: @dfn{child} of @code{def-word}.
6452:
6453: When you execute @code{name}, the address of the body of @code{name} is
6454: put on the data stack and @i{code2} is executed (the address of the body
6455: of @code{name} is the address @code{HERE} returns immediately after the
6456: @code{CREATE}, i.e., the address a @code{create}d word returns by
6457: default).
6458:
6459: @c anton:
6460: @c www.dictionary.com says:
6461: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
6462: @c several generations of absence, usually caused by the chance
6463: @c recombination of genes. 2.An individual or a part that exhibits
6464: @c atavism. Also called throwback. 3.The return of a trait or recurrence
6465: @c of previous behavior after a period of absence.
6466: @c
6467: @c Doesn't seem to fit.
6468:
6469: @c @cindex atavism in child words
6470: You can use @code{def-word} to define a set of child words that behave
6471: similarly; they all have a common run-time behaviour determined by
6472: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
6473: body of the child word. The structure of the data is common to all
6474: children of @code{def-word}, but the data values are specific -- and
6475: private -- to each child word. When a child word is executed, the
6476: address of its private data area is passed as a parameter on TOS to be
6477: used and manipulated@footnote{It is legitimate both to read and write to
6478: this data area.} by @i{code2}.
6479:
6480: The two fragments of code that make up the defining words act (are
6481: executed) at two completely separate times:
6482:
6483: @itemize @bullet
6484: @item
6485: At @i{define time}, the defining word executes @i{code1} to generate a
6486: child word
6487: @item
6488: At @i{child execution time}, when a child word is invoked, @i{code2}
6489: is executed, using parameters (data) that are private and specific to
6490: the child word.
6491: @end itemize
6492:
6493: Another way of understanding the behaviour of @code{def-word} and
6494: @code{name} is to say that, if you make the following definitions:
6495: @example
6496: : def-word1 ( "name" -- )
6497: CREATE @i{code1} ;
6498:
6499: : action1 ( ... -- ... )
6500: @i{code2} ;
6501:
6502: def-word1 name1
6503: @end example
6504:
6505: @noindent
6506: Then using @code{name1 action1} is equivalent to using @code{name}.
6507:
6508: The classic example is that you can define @code{CONSTANT} in this way:
6509:
6510: @example
6511: : CONSTANT ( w "name" -- )
6512: CREATE ,
6513: DOES> ( -- w )
6514: @@ ;
6515: @end example
6516:
6517: @comment There is a beautiful description of how this works and what
6518: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6519: @comment commentary on the Counting Fruits problem.
6520:
6521: When you create a constant with @code{5 CONSTANT five}, a set of
6522: define-time actions take place; first a new word @code{five} is created,
6523: then the value 5 is laid down in the body of @code{five} with
6524: @code{,}. When @code{five} is executed, the address of the body is put on
6525: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6526: no code of its own; it simply contains a data field and a pointer to the
6527: code that follows @code{DOES>} in its defining word. That makes words
6528: created in this way very compact.
6529:
6530: The final example in this section is intended to remind you that space
6531: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6532: both read and written by a Standard program@footnote{Exercise: use this
6533: example as a starting point for your own implementation of @code{Value}
6534: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6535: @code{[']}.}:
6536:
6537: @example
6538: : foo ( "name" -- )
6539: CREATE -1 ,
6540: DOES> ( -- )
6541: @@ . ;
6542:
6543: foo first-word
6544: foo second-word
6545:
6546: 123 ' first-word >BODY !
6547: @end example
6548:
6549: If @code{first-word} had been a @code{CREATE}d word, we could simply
6550: have executed it to get the address of its data field. However, since it
6551: was defined to have @code{DOES>} actions, its execution semantics are to
6552: perform those @code{DOES>} actions. To get the address of its data field
6553: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6554: translate the xt into the address of the data field. When you execute
6555: @code{first-word}, it will display @code{123}. When you execute
6556: @code{second-word} it will display @code{-1}.
6557:
6558: @cindex stack effect of @code{DOES>}-parts
6559: @cindex @code{DOES>}-parts, stack effect
6560: In the examples above the stack comment after the @code{DOES>} specifies
6561: the stack effect of the defined words, not the stack effect of the
6562: following code (the following code expects the address of the body on
6563: the top of stack, which is not reflected in the stack comment). This is
6564: the convention that I use and recommend (it clashes a bit with using
6565: locals declarations for stack effect specification, though).
6566:
6567: @menu
6568: * CREATE..DOES> applications::
6569: * CREATE..DOES> details::
6570: * Advanced does> usage example::
6571: * @code{Const-does>}::
6572: @end menu
6573:
6574: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
6575: @subsubsection Applications of @code{CREATE..DOES>}
6576: @cindex @code{CREATE} ... @code{DOES>}, applications
6577:
6578: You may wonder how to use this feature. Here are some usage patterns:
6579:
6580: @cindex factoring similar colon definitions
6581: When you see a sequence of code occurring several times, and you can
6582: identify a meaning, you will factor it out as a colon definition. When
6583: you see similar colon definitions, you can factor them using
6584: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6585: that look very similar:
6586: @example
6587: : ori, ( reg-target reg-source n -- )
6588: 0 asm-reg-reg-imm ;
6589: : andi, ( reg-target reg-source n -- )
6590: 1 asm-reg-reg-imm ;
6591: @end example
6592:
6593: @noindent
6594: This could be factored with:
6595: @example
6596: : reg-reg-imm ( op-code -- )
6597: CREATE ,
6598: DOES> ( reg-target reg-source n -- )
6599: @@ asm-reg-reg-imm ;
6600:
6601: 0 reg-reg-imm ori,
6602: 1 reg-reg-imm andi,
6603: @end example
6604:
6605: @cindex currying
6606: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6607: supply a part of the parameters for a word (known as @dfn{currying} in
6608: the functional language community). E.g., @code{+} needs two
6609: parameters. Creating versions of @code{+} with one parameter fixed can
6610: be done like this:
6611:
6612: @example
6613: : curry+ ( n1 "name" -- )
6614: CREATE ,
6615: DOES> ( n2 -- n1+n2 )
6616: @@ + ;
6617:
6618: 3 curry+ 3+
6619: -2 curry+ 2-
6620: @end example
6621:
6622:
6623: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
6624: @subsubsection The gory details of @code{CREATE..DOES>}
6625: @cindex @code{CREATE} ... @code{DOES>}, details
6626:
6627: doc-does>
6628:
6629: @cindex @code{DOES>} in a separate definition
6630: This means that you need not use @code{CREATE} and @code{DOES>} in the
6631: same definition; you can put the @code{DOES>}-part in a separate
6632: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
6633: @example
6634: : does1
6635: DOES> ( ... -- ... )
6636: ... ;
6637:
6638: : does2
6639: DOES> ( ... -- ... )
6640: ... ;
6641:
6642: : def-word ( ... -- ... )
6643: create ...
6644: IF
6645: does1
6646: ELSE
6647: does2
6648: ENDIF ;
6649: @end example
6650:
6651: In this example, the selection of whether to use @code{does1} or
6652: @code{does2} is made at definition-time; at the time that the child word is
6653: @code{CREATE}d.
6654:
6655: @cindex @code{DOES>} in interpretation state
6656: In a standard program you can apply a @code{DOES>}-part only if the last
6657: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6658: will override the behaviour of the last word defined in any case. In a
6659: standard program, you can use @code{DOES>} only in a colon
6660: definition. In Gforth, you can also use it in interpretation state, in a
6661: kind of one-shot mode; for example:
6662: @example
6663: CREATE name ( ... -- ... )
6664: @i{initialization}
6665: DOES>
6666: @i{code} ;
6667: @end example
6668:
6669: @noindent
6670: is equivalent to the standard:
6671: @example
6672: :noname
6673: DOES>
6674: @i{code} ;
6675: CREATE name EXECUTE ( ... -- ... )
6676: @i{initialization}
6677: @end example
6678:
6679: doc->body
6680:
6681: @node Advanced does> usage example, @code{Const-does>}, CREATE..DOES> details, User-defined Defining Words
6682: @subsubsection Advanced does> usage example
6683:
6684: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6685: for disassembling instructions, that follow a very repetetive scheme:
6686:
6687: @example
6688: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6689: @var{entry-num} cells @var{table} + !
6690: @end example
6691:
6692: Of course, this inspires the idea to factor out the commonalities to
6693: allow a definition like
6694:
6695: @example
6696: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6697: @end example
6698:
6699: The parameters @var{disasm-operands} and @var{table} are usually
6700: correlated. Moreover, before I wrote the disassembler, there already
6701: existed code that defines instructions like this:
6702:
6703: @example
6704: @var{entry-num} @var{inst-format} @var{inst-name}
6705: @end example
6706:
6707: This code comes from the assembler and resides in
6708: @file{arch/mips/insts.fs}.
6709:
6710: So I had to define the @var{inst-format} words that performed the scheme
6711: above when executed. At first I chose to use run-time code-generation:
6712:
6713: @example
6714: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6715: :noname Postpone @var{disasm-operands}
6716: name Postpone sliteral Postpone type Postpone ;
6717: swap cells @var{table} + ! ;
6718: @end example
6719:
6720: Note that this supplies the other two parameters of the scheme above.
6721:
6722: An alternative would have been to write this using
6723: @code{create}/@code{does>}:
6724:
6725: @example
6726: : @var{inst-format} ( entry-num "name" -- )
6727: here name string, ( entry-num c-addr ) \ parse and save "name"
6728: noname create , ( entry-num )
6729: lastxt swap cells @var{table} + !
6730: does> ( addr w -- )
6731: \ disassemble instruction w at addr
6732: @@ >r
6733: @var{disasm-operands}
6734: r> count type ;
6735: @end example
6736:
6737: Somehow the first solution is simpler, mainly because it's simpler to
6738: shift a string from definition-time to use-time with @code{sliteral}
6739: than with @code{string,} and friends.
6740:
6741: I wrote a lot of words following this scheme and soon thought about
6742: factoring out the commonalities among them. Note that this uses a
6743: two-level defining word, i.e., a word that defines ordinary defining
6744: words.
6745:
6746: This time a solution involving @code{postpone} and friends seemed more
6747: difficult (try it as an exercise), so I decided to use a
6748: @code{create}/@code{does>} word; since I was already at it, I also used
6749: @code{create}/@code{does>} for the lower level (try using
6750: @code{postpone} etc. as an exercise), resulting in the following
6751: definition:
6752:
6753: @example
6754: : define-format ( disasm-xt table-xt -- )
6755: \ define an instruction format that uses disasm-xt for
6756: \ disassembling and enters the defined instructions into table
6757: \ table-xt
6758: create 2,
6759: does> ( u "inst" -- )
6760: \ defines an anonymous word for disassembling instruction inst,
6761: \ and enters it as u-th entry into table-xt
6762: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6763: noname create 2, \ define anonymous word
6764: execute lastxt swap ! \ enter xt of defined word into table-xt
6765: does> ( addr w -- )
6766: \ disassemble instruction w at addr
6767: 2@@ >r ( addr w disasm-xt R: c-addr )
6768: execute ( R: c-addr ) \ disassemble operands
6769: r> count type ; \ print name
6770: @end example
6771:
6772: Note that the tables here (in contrast to above) do the @code{cells +}
6773: by themselves (that's why you have to pass an xt). This word is used in
6774: the following way:
6775:
6776: @example
6777: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6778: @end example
6779:
6780: As shown above, the defined instruction format is then used like this:
6781:
6782: @example
6783: @var{entry-num} @var{inst-format} @var{inst-name}
6784: @end example
6785:
6786: In terms of currying, this kind of two-level defining word provides the
6787: parameters in three stages: first @var{disasm-operands} and @var{table},
6788: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6789: the instruction to be disassembled.
6790:
6791: Of course this did not quite fit all the instruction format names used
6792: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6793: the parameters into the right form.
6794:
6795: If you have trouble following this section, don't worry. First, this is
6796: involved and takes time (and probably some playing around) to
6797: understand; second, this is the first two-level
6798: @code{create}/@code{does>} word I have written in seventeen years of
6799: Forth; and if I did not have @file{insts.fs} to start with, I may well
6800: have elected to use just a one-level defining word (with some repeating
6801: of parameters when using the defining word). So it is not necessary to
6802: understand this, but it may improve your understanding of Forth.
6803:
6804:
6805: @node @code{Const-does>}, , Advanced does> usage example, User-defined Defining Words
6806: @subsubsection @code{Const-does>}
6807:
6808: A frequent use of @code{create}...@code{does>} is for transferring some
6809: values from definition-time to run-time. Gforth supports this use with
6810:
6811: doc-const-does>
6812:
6813: A typical use of this word is:
6814:
6815: @example
6816: : curry+ ( n1 "name" -- )
6817: 1 0 CONST-DOES> ( n2 -- n1+n2 )
6818: + ;
6819:
6820: 3 curry+ 3+
6821: @end example
6822:
6823: Here the @code{1 0} means that 1 cell and 0 floats are transferred from
6824: definition to run-time.
6825:
6826: The advantages of using @code{const-does>} are:
6827:
6828: @itemize
6829:
6830: @item
6831: You don't have to deal with storing and retrieving the values, i.e.,
6832: your program becomes more writable and readable.
6833:
6834: @item
6835: When using @code{does>}, you have to introduce a @code{@@} that cannot
6836: be optimized away (because you could change the data using
6837: @code{>body}...@code{!}); @code{const-does>} avoids this problem.
6838:
6839: @end itemize
6840:
6841: An ANS Forth implementation of @code{const-does>} is available in
6842: @file{compat/const-does.fs}.
6843:
6844:
6845: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6846: @subsection Deferred words
6847: @cindex deferred words
6848:
6849: The defining word @code{Defer} allows you to define a word by name
6850: without defining its behaviour; the definition of its behaviour is
6851: deferred. Here are two situation where this can be useful:
6852:
6853: @itemize @bullet
6854: @item
6855: Where you want to allow the behaviour of a word to be altered later, and
6856: for all precompiled references to the word to change when its behaviour
6857: is changed.
6858: @item
6859: For mutual recursion; @xref{Calls and returns}.
6860: @end itemize
6861:
6862: In the following example, @code{foo} always invokes the version of
6863: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6864: always invokes the version that prints ``@code{Hello}''. There is no way
6865: of getting @code{foo} to use the later version without re-ordering the
6866: source code and recompiling it.
6867:
6868: @example
6869: : greet ." Good morning" ;
6870: : foo ... greet ... ;
6871: : greet ." Hello" ;
6872: : bar ... greet ... ;
6873: @end example
6874:
6875: This problem can be solved by defining @code{greet} as a @code{Defer}red
6876: word. The behaviour of a @code{Defer}red word can be defined and
6877: redefined at any time by using @code{IS} to associate the xt of a
6878: previously-defined word with it. The previous example becomes:
6879:
6880: @example
6881: Defer greet ( -- )
6882: : foo ... greet ... ;
6883: : bar ... greet ... ;
6884: : greet1 ( -- ) ." Good morning" ;
6885: : greet2 ( -- ) ." Hello" ;
6886: ' greet2 <IS> greet \ make greet behave like greet2
6887: @end example
6888:
6889: @progstyle
6890: You should write a stack comment for every deferred word, and put only
6891: XTs into deferred words that conform to this stack effect. Otherwise
6892: it's too difficult to use the deferred word.
6893:
6894: A deferred word can be used to improve the statistics-gathering example
6895: from @ref{User-defined Defining Words}; rather than edit the
6896: application's source code to change every @code{:} to a @code{my:}, do
6897: this:
6898:
6899: @example
6900: : real: : ; \ retain access to the original
6901: defer : \ redefine as a deferred word
6902: ' my: <IS> : \ use special version of :
6903: \
6904: \ load application here
6905: \
6906: ' real: <IS> : \ go back to the original
6907: @end example
6908:
6909:
6910: One thing to note is that @code{<IS>} consumes its name when it is
6911: executed. If you want to specify the name at compile time, use
6912: @code{[IS]}:
6913:
6914: @example
6915: : set-greet ( xt -- )
6916: [IS] greet ;
6917:
6918: ' greet1 set-greet
6919: @end example
6920:
6921: A deferred word can only inherit execution semantics from the xt
6922: (because that is all that an xt can represent -- for more discussion of
6923: this @pxref{Tokens for Words}); by default it will have default
6924: interpretation and compilation semantics deriving from this execution
6925: semantics. However, you can change the interpretation and compilation
6926: semantics of the deferred word in the usual ways:
6927:
6928: @example
6929: : bar .... ; compile-only
6930: Defer fred immediate
6931: Defer jim
6932:
6933: ' bar <IS> jim \ jim has default semantics
6934: ' bar <IS> fred \ fred is immediate
6935: @end example
6936:
6937: doc-defer
6938: doc-<is>
6939: doc-[is]
6940: doc-is
6941: @comment TODO document these: what's defers [is]
6942: doc-what's
6943: doc-defers
6944:
6945: @c Use @code{words-deferred} to see a list of deferred words.
6946:
6947: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
6948: are provided in @file{compat/defer.fs}.
6949:
6950:
6951: @node Aliases, , Deferred words, Defining Words
6952: @subsection Aliases
6953: @cindex aliases
6954:
6955: The defining word @code{Alias} allows you to define a word by name that
6956: has the same behaviour as some other word. Here are two situation where
6957: this can be useful:
6958:
6959: @itemize @bullet
6960: @item
6961: When you want access to a word's definition from a different word list
6962: (for an example of this, see the definition of the @code{Root} word list
6963: in the Gforth source).
6964: @item
6965: When you want to create a synonym; a definition that can be known by
6966: either of two names (for example, @code{THEN} and @code{ENDIF} are
6967: aliases).
6968: @end itemize
6969:
6970: Like deferred words, an alias has default compilation and interpretation
6971: semantics at the beginning (not the modifications of the other word),
6972: but you can change them in the usual ways (@code{immediate},
6973: @code{compile-only}). For example:
6974:
6975: @example
6976: : foo ... ; immediate
6977:
6978: ' foo Alias bar \ bar is not an immediate word
6979: ' foo Alias fooby immediate \ fooby is an immediate word
6980: @end example
6981:
6982: Words that are aliases have the same xt, different headers in the
6983: dictionary, and consequently different name tokens (@pxref{Tokens for
6984: Words}) and possibly different immediate flags. An alias can only have
6985: default or immediate compilation semantics; you can define aliases for
6986: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
6987:
6988: doc-alias
6989:
6990:
6991: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6992: @section Interpretation and Compilation Semantics
6993: @cindex semantics, interpretation and compilation
6994:
6995: @c !! state and ' are used without explanation
6996: @c example for immediate/compile-only? or is the tutorial enough
6997:
6998: @cindex interpretation semantics
6999: The @dfn{interpretation semantics} of a (named) word are what the text
7000: interpreter does when it encounters the word in interpret state. It also
7001: appears in some other contexts, e.g., the execution token returned by
7002: @code{' @i{word}} identifies the interpretation semantics of @i{word}
7003: (in other words, @code{' @i{word} execute} is equivalent to
7004: interpret-state text interpretation of @code{@i{word}}).
7005:
7006: @cindex compilation semantics
7007: The @dfn{compilation semantics} of a (named) word are what the text
7008: interpreter does when it encounters the word in compile state. It also
7009: appears in other contexts, e.g, @code{POSTPONE @i{word}}
7010: compiles@footnote{In standard terminology, ``appends to the current
7011: definition''.} the compilation semantics of @i{word}.
7012:
7013: @cindex execution semantics
7014: The standard also talks about @dfn{execution semantics}. They are used
7015: only for defining the interpretation and compilation semantics of many
7016: words. By default, the interpretation semantics of a word are to
7017: @code{execute} its execution semantics, and the compilation semantics of
7018: a word are to @code{compile,} its execution semantics.@footnote{In
7019: standard terminology: The default interpretation semantics are its
7020: execution semantics; the default compilation semantics are to append its
7021: execution semantics to the execution semantics of the current
7022: definition.}
7023:
7024: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
7025: the text interpreter, ticked, or @code{postpone}d, so they have no
7026: interpretation or compilation semantics. Their behaviour is represented
7027: by their XT (@pxref{Tokens for Words}), and we call it execution
7028: semantics, too.
7029:
7030: @comment TODO expand, make it co-operate with new sections on text interpreter.
7031:
7032: @cindex immediate words
7033: @cindex compile-only words
7034: You can change the semantics of the most-recently defined word:
7035:
7036:
7037: doc-immediate
7038: doc-compile-only
7039: doc-restrict
7040:
7041: By convention, words with non-default compilation semantics (e.g.,
7042: immediate words) often have names surrounded with brackets (e.g.,
7043: @code{[']}, @pxref{Execution token}).
7044:
7045: Note that ticking (@code{'}) a compile-only word gives an error
7046: (``Interpreting a compile-only word'').
7047:
7048: @menu
7049: * Combined words::
7050: @end menu
7051:
7052:
7053: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
7054: @subsection Combined Words
7055: @cindex combined words
7056:
7057: Gforth allows you to define @dfn{combined words} -- words that have an
7058: arbitrary combination of interpretation and compilation semantics.
7059:
7060: doc-interpret/compile:
7061:
7062: This feature was introduced for implementing @code{TO} and @code{S"}. I
7063: recommend that you do not define such words, as cute as they may be:
7064: they make it hard to get at both parts of the word in some contexts.
7065: E.g., assume you want to get an execution token for the compilation
7066: part. Instead, define two words, one that embodies the interpretation
7067: part, and one that embodies the compilation part. Once you have done
7068: that, you can define a combined word with @code{interpret/compile:} for
7069: the convenience of your users.
7070:
7071: You might try to use this feature to provide an optimizing
7072: implementation of the default compilation semantics of a word. For
7073: example, by defining:
7074: @example
7075: :noname
7076: foo bar ;
7077: :noname
7078: POSTPONE foo POSTPONE bar ;
7079: interpret/compile: opti-foobar
7080: @end example
7081:
7082: @noindent
7083: as an optimizing version of:
7084:
7085: @example
7086: : foobar
7087: foo bar ;
7088: @end example
7089:
7090: Unfortunately, this does not work correctly with @code{[compile]},
7091: because @code{[compile]} assumes that the compilation semantics of all
7092: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
7093: opti-foobar} would compile compilation semantics, whereas
7094: @code{[compile] foobar} would compile interpretation semantics.
7095:
7096: @cindex state-smart words (are a bad idea)
7097: @anchor{state-smartness}
7098: Some people try to use @dfn{state-smart} words to emulate the feature provided
7099: by @code{interpret/compile:} (words are state-smart if they check
7100: @code{STATE} during execution). E.g., they would try to code
7101: @code{foobar} like this:
7102:
7103: @example
7104: : foobar
7105: STATE @@
7106: IF ( compilation state )
7107: POSTPONE foo POSTPONE bar
7108: ELSE
7109: foo bar
7110: ENDIF ; immediate
7111: @end example
7112:
7113: Although this works if @code{foobar} is only processed by the text
7114: interpreter, it does not work in other contexts (like @code{'} or
7115: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
7116: for a state-smart word, not for the interpretation semantics of the
7117: original @code{foobar}; when you execute this execution token (directly
7118: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
7119: state, the result will not be what you expected (i.e., it will not
7120: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
7121: write them@footnote{For a more detailed discussion of this topic, see
7122: M. Anton Ertl,
7123: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
7124: it is Evil and How to Exorcise it}}, EuroForth '98.}!
7125:
7126: @cindex defining words with arbitrary semantics combinations
7127: It is also possible to write defining words that define words with
7128: arbitrary combinations of interpretation and compilation semantics. In
7129: general, they look like this:
7130:
7131: @example
7132: : def-word
7133: create-interpret/compile
7134: @i{code1}
7135: interpretation>
7136: @i{code2}
7137: <interpretation
7138: compilation>
7139: @i{code3}
7140: <compilation ;
7141: @end example
7142:
7143: For a @i{word} defined with @code{def-word}, the interpretation
7144: semantics are to push the address of the body of @i{word} and perform
7145: @i{code2}, and the compilation semantics are to push the address of
7146: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
7147: can also be defined like this (except that the defined constants don't
7148: behave correctly when @code{[compile]}d):
7149:
7150: @example
7151: : constant ( n "name" -- )
7152: create-interpret/compile
7153: ,
7154: interpretation> ( -- n )
7155: @@
7156: <interpretation
7157: compilation> ( compilation. -- ; run-time. -- n )
7158: @@ postpone literal
7159: <compilation ;
7160: @end example
7161:
7162:
7163: doc-create-interpret/compile
7164: doc-interpretation>
7165: doc-<interpretation
7166: doc-compilation>
7167: doc-<compilation
7168:
7169:
7170: Words defined with @code{interpret/compile:} and
7171: @code{create-interpret/compile} have an extended header structure that
7172: differs from other words; however, unless you try to access them with
7173: plain address arithmetic, you should not notice this. Words for
7174: accessing the header structure usually know how to deal with this; e.g.,
7175: @code{'} @i{word} @code{>body} also gives you the body of a word created
7176: with @code{create-interpret/compile}.
7177:
7178:
7179: @c -------------------------------------------------------------
7180: @node Tokens for Words, Compiling words, Interpretation and Compilation Semantics, Words
7181: @section Tokens for Words
7182: @cindex tokens for words
7183:
7184: This section describes the creation and use of tokens that represent
7185: words.
7186:
7187: @menu
7188: * Execution token:: represents execution/interpretation semantics
7189: * Compilation token:: represents compilation semantics
7190: * Name token:: represents named words
7191: @end menu
7192:
7193: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
7194: @subsection Execution token
7195:
7196: @cindex xt
7197: @cindex execution token
7198: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
7199: You can use @code{execute} to invoke this behaviour.
7200:
7201: @cindex tick (')
7202: You can use @code{'} to get an execution token that represents the
7203: interpretation semantics of a named word:
7204:
7205: @example
7206: 5 ' . ( n xt )
7207: execute ( ) \ execute the xt (i.e., ".")
7208: @end example
7209:
7210: doc-'
7211:
7212: @code{'} parses at run-time; there is also a word @code{[']} that parses
7213: when it is compiled, and compiles the resulting XT:
7214:
7215: @example
7216: : foo ['] . execute ;
7217: 5 foo
7218: : bar ' execute ; \ by contrast,
7219: 5 bar . \ ' parses "." when bar executes
7220: @end example
7221:
7222: doc-[']
7223:
7224: If you want the execution token of @i{word}, write @code{['] @i{word}}
7225: in compiled code and @code{' @i{word}} in interpreted code. Gforth's
7226: @code{'} and @code{[']} behave somewhat unusually by complaining about
7227: compile-only words (because these words have no interpretation
7228: semantics). You might get what you want by using @code{COMP' @i{word}
7229: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
7230: token}).
7231:
7232: Another way to get an XT is @code{:noname} or @code{lastxt}
7233: (@pxref{Anonymous Definitions}). For anonymous words this gives an xt
7234: for the only behaviour the word has (the execution semantics). For
7235: named words, @code{lastxt} produces an XT for the same behaviour it
7236: would produce if the word was defined anonymously.
7237:
7238: @example
7239: :noname ." hello" ;
7240: execute
7241: @end example
7242:
7243: An XT occupies one cell and can be manipulated like any other cell.
7244:
7245: @cindex code field address
7246: @cindex CFA
7247: In ANS Forth the XT is just an abstract data type (i.e., defined by the
7248: operations that produce or consume it). For old hands: In Gforth, the
7249: XT is implemented as a code field address (CFA).
7250:
7251: doc-execute
7252: doc-perform
7253:
7254: @node Compilation token, Name token, Execution token, Tokens for Words
7255: @subsection Compilation token
7256:
7257: @cindex compilation token
7258: @cindex CT (compilation token)
7259: Gforth represents the compilation semantics of a named word by a
7260: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
7261: @i{xt} is an execution token. The compilation semantics represented by
7262: the compilation token can be performed with @code{execute}, which
7263: consumes the whole compilation token, with an additional stack effect
7264: determined by the represented compilation semantics.
7265:
7266: At present, the @i{w} part of a compilation token is an execution token,
7267: and the @i{xt} part represents either @code{execute} or
7268: @code{compile,}@footnote{Depending upon the compilation semantics of the
7269: word. If the word has default compilation semantics, the @i{xt} will
7270: represent @code{compile,}. Otherwise (e.g., for immediate words), the
7271: @i{xt} will represent @code{execute}.}. However, don't rely on that
7272: knowledge, unless necessary; future versions of Gforth may introduce
7273: unusual compilation tokens (e.g., a compilation token that represents
7274: the compilation semantics of a literal).
7275:
7276: You can perform the compilation semantics represented by the compilation
7277: token with @code{execute}. You can compile the compilation semantics
7278: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
7279: equivalent to @code{postpone @i{word}}.
7280:
7281: doc-[comp']
7282: doc-comp'
7283: doc-postpone,
7284:
7285: @node Name token, , Compilation token, Tokens for Words
7286: @subsection Name token
7287:
7288: @cindex name token
7289: @cindex name field address
7290: @cindex NFA
7291: Gforth represents named words by the @dfn{name token}, (@i{nt}). In
7292: Gforth, the abstract data type @emph{name token} is implemented as a
7293: name field address (NFA).
7294:
7295: doc-find-name
7296: doc-name>int
7297: doc-name?int
7298: doc-name>comp
7299: doc-name>string
7300:
7301: @c ----------------------------------------------------------
7302: @node Compiling words, The Text Interpreter, Tokens for Words, Words
7303: @section Compiling words
7304: @cindex compiling words
7305: @cindex macros
7306:
7307: In contrast to most other languages, Forth has no strict boundary
7308: between compilation and run-time. E.g., you can run arbitrary code
7309: between defining words (or for computing data used by defining words
7310: like @code{constant}). Moreover, @code{Immediate} (@pxref{Interpretation
7311: and Compilation Semantics} and @code{[}...@code{]} (see below) allow
7312: running arbitrary code while compiling a colon definition (exception:
7313: you must not allot dictionary space).
7314:
7315: @menu
7316: * Literals:: Compiling data values
7317: * Macros:: Compiling words
7318: @end menu
7319:
7320: @node Literals, Macros, Compiling words, Compiling words
7321: @subsection Literals
7322: @cindex Literals
7323:
7324: The simplest and most frequent example is to compute a literal during
7325: compilation. E.g., the following definition prints an array of strings,
7326: one string per line:
7327:
7328: @example
7329: : .strings ( addr u -- ) \ gforth
7330: 2* cells bounds U+DO
7331: cr i 2@@ type
7332: 2 cells +LOOP ;
7333: @end example
7334:
7335: With a simple-minded compiler like Gforth's, this computes @code{2
7336: cells} on every loop iteration. You can compute this value once and for
7337: all at compile time and compile it into the definition like this:
7338:
7339: @example
7340: : .strings ( addr u -- ) \ gforth
7341: 2* cells bounds U+DO
7342: cr i 2@@ type
7343: [ 2 cells ] literal +LOOP ;
7344: @end example
7345:
7346: @code{[} switches the text interpreter to interpret state (you will get
7347: an @code{ok} prompt if you type this example interactively and insert a
7348: newline between @code{[} and @code{]}), so it performs the
7349: interpretation semantics of @code{2 cells}; this computes a number.
7350: @code{]} switches the text interpreter back into compile state. It then
7351: performs @code{Literal}'s compilation semantics, which are to compile
7352: this number into the current word. You can decompile the word with
7353: @code{see .strings} to see the effect on the compiled code.
7354:
7355: You can also optimize the @code{2* cells} into @code{[ 2 cells ] literal
7356: *} in this way.
7357:
7358: doc-[
7359: doc-]
7360: doc-literal
7361: doc-]L
7362:
7363: There are also words for compiling other data types than single cells as
7364: literals:
7365:
7366: doc-2literal
7367: doc-fliteral
7368: doc-sliteral
7369:
7370: @cindex colon-sys, passing data across @code{:}
7371: @cindex @code{:}, passing data across
7372: You might be tempted to pass data from outside a colon definition to the
7373: inside on the data stack. This does not work, because @code{:} puhes a
7374: colon-sys, making stuff below unaccessible. E.g., this does not work:
7375:
7376: @example
7377: 5 : foo literal ; \ error: "unstructured"
7378: @end example
7379:
7380: Instead, you have to pass the value in some other way, e.g., through a
7381: variable:
7382:
7383: @example
7384: variable temp
7385: 5 temp !
7386: : foo [ temp @@ ] literal ;
7387: @end example
7388:
7389:
7390: @node Macros, , Literals, Compiling words
7391: @subsection Macros
7392: @cindex Macros
7393: @cindex compiling compilation semantics
7394:
7395: @code{Literal} and friends compile data values into the current
7396: definition. You can also write words that compile other words into the
7397: current definition. E.g.,
7398:
7399: @example
7400: : compile-+ ( -- ) \ compiled code: ( n1 n2 -- n )
7401: POSTPONE + ;
7402:
7403: : foo ( n1 n2 -- n )
7404: [ compile-+ ] ;
7405: 1 2 foo .
7406: @end example
7407:
7408: This is equivalent to @code{: foo + ;} (@code{see foo} to check this).
7409: What happens in this example? @code{Postpone} compiles the compilation
7410: semantics of @code{+} into @code{compile-+}; later the text interpreter
7411: executes @code{compile-+} and thus the compilation semantics of +, which
7412: compile (the execution semantics of) @code{+} into
7413: @code{foo}.@footnote{A recent RFI answer requires that compiling words
7414: should only be executed in compile state, so this example is not
7415: guaranteed to work on all standard systems, but on any decent system it
7416: will work.}
7417:
7418: doc-postpone
7419: doc-[compile]
7420:
7421: Compiling words like @code{compile-+} are usually immediate (or similar)
7422: so you do not have to switch to interpret state to execute them;
7423: mopifying the last example accordingly produces:
7424:
7425: @example
7426: : [compile-+] ( compilation: --; interpretation: -- )
7427: \ compiled code: ( n1 n2 -- n )
7428: POSTPONE + ; immediate
7429:
7430: : foo ( n1 n2 -- n )
7431: [compile-+] ;
7432: 1 2 foo .
7433: @end example
7434:
7435: Immediate compiling words are similar to macros in other languages (in
7436: particular, Lisp). The important differences to macros in, e.g., C are:
7437:
7438: @itemize @bullet
7439:
7440: @item
7441: You use the same language for defining and processing macros, not a
7442: separate preprocessing language and processor.
7443:
7444: @item
7445: Consequently, the full power of Forth is available in macro definitions.
7446: E.g., you can perform arbitrarily complex computations, or generate
7447: different code conditionally or in a loop (e.g., @pxref{Advanced macros
7448: Tutorial}). This power is very useful when writing a parser generators
7449: or other code-generating software.
7450:
7451: @item
7452: Macros defined using @code{postpone} etc. deal with the language at a
7453: higher level than strings; name binding happens at macro definition
7454: time, so you can avoid the pitfalls of name collisions that can happen
7455: in C macros. Of course, Forth is a liberal language and also allows to
7456: shoot yourself in the foot with text-interpreted macros like
7457:
7458: @example
7459: : [compile-+] s" +" evaluate ; immediate
7460: @end example
7461:
7462: Apart from binding the name at macro use time, using @code{evaluate}
7463: also makes your definition @code{state}-smart (@pxref{state-smartness}).
7464: @end itemize
7465:
7466: You may want the macro to compile a number into a word. The word to do
7467: it is @code{literal}, but you have to @code{postpone} it, so its
7468: compilation semantics take effect when the macro is executed, not when
7469: it is compiled:
7470:
7471: @example
7472: : [compile-5] ( -- ) \ compiled code: ( -- n )
7473: 5 POSTPONE literal ; immediate
7474:
7475: : foo [compile-5] ;
7476: foo .
7477: @end example
7478:
7479: You may want to pass parameters to a macro, that the macro should
7480: compile into the current definition. If the parameter is a number, then
7481: you can use @code{postpone literal} (similar for other values).
7482:
7483: If you want to pass a word that is to be compiled, the usual way is to
7484: pass an execution token and @code{compile,} it:
7485:
7486: @example
7487: : twice1 ( xt -- ) \ compiled code: ... -- ...
7488: dup compile, compile, ;
7489:
7490: : 2+ ( n1 -- n2 )
7491: [ ' 1+ twice1 ] ;
7492: @end example
7493:
7494: doc-compile,
7495:
7496: An alternative available in Gforth, that allows you to pass compile-only
7497: words as parameters is to use the compilation token (@pxref{Compilation
7498: token}). The same example in this technique:
7499:
7500: @example
7501: : twice ( ... ct -- ... ) \ compiled code: ... -- ...
7502: 2dup 2>r execute 2r> execute ;
7503:
7504: : 2+ ( n1 -- n2 )
7505: [ comp' 1+ twice ] ;
7506: @end example
7507:
7508: In the example above @code{2>r} and @code{2r>} ensure that @code{twice}
7509: works even if the executed compilation semantics has an effect on the
7510: data stack.
7511:
7512: You can also define complete definitions with these words; this provides
7513: an alternative to using @code{does>} (@pxref{User-defined Defining
7514: Words}). E.g., instead of
7515:
7516: @example
7517: : curry+ ( n1 "name" -- )
7518: CREATE ,
7519: DOES> ( n2 -- n1+n2 )
7520: @@ + ;
7521: @end example
7522:
7523: you could define
7524:
7525: @example
7526: : curry+ ( n1 "name" -- )
7527: \ name execution: ( n2 -- n1+n2 )
7528: >r : r> POSTPONE literal POSTPONE + POSTPONE ; ;
7529:
7530: -3 curry+ 3-
7531: see 3-
7532: @end example
7533:
7534: The sequence @code{>r : r>} is necessary, because @code{:} puts a
7535: colon-sys on the data stack that makes everything below it unaccessible.
7536:
7537: This way of writing defining words is sometimes more, sometimes less
7538: convenient than using @code{does>} (@pxref{Advanced does> usage
7539: example}). One advantage of this method is that it can be optimized
7540: better, because the compiler knows that the value compiled with
7541: @code{literal} is fixed, whereas the data associated with a
7542: @code{create}d word can be changed.
7543:
7544: @c ----------------------------------------------------------
7545: @node The Text Interpreter, Word Lists, Compiling words, Words
7546: @section The Text Interpreter
7547: @cindex interpreter - outer
7548: @cindex text interpreter
7549: @cindex outer interpreter
7550:
7551: @c Should we really describe all these ugly details? IMO the text
7552: @c interpreter should be much cleaner, but that may not be possible within
7553: @c ANS Forth. - anton
7554: @c nac-> I wanted to explain how it works to show how you can exploit
7555: @c it in your own programs. When I was writing a cross-compiler, figuring out
7556: @c some of these gory details was very helpful to me. None of the textbooks
7557: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
7558: @c seems to positively avoid going into too much detail for some of
7559: @c the internals.
7560:
7561: @c anton: ok. I wonder, though, if this is the right place; for some stuff
7562: @c it is; for the ugly details, I would prefer another place. I wonder
7563: @c whether we should have a chapter before "Words" that describes some
7564: @c basic concepts referred to in words, and a chapter after "Words" that
7565: @c describes implementation details.
7566:
7567: The text interpreter@footnote{This is an expanded version of the
7568: material in @ref{Introducing the Text Interpreter}.} is an endless loop
7569: that processes input from the current input device. It is also called
7570: the outer interpreter, in contrast to the inner interpreter
7571: (@pxref{Engine}) which executes the compiled Forth code on interpretive
7572: implementations.
7573:
7574: @cindex interpret state
7575: @cindex compile state
7576: The text interpreter operates in one of two states: @dfn{interpret
7577: state} and @dfn{compile state}. The current state is defined by the
7578: aptly-named variable @code{state}.
7579:
7580: This section starts by describing how the text interpreter behaves when
7581: it is in interpret state, processing input from the user input device --
7582: the keyboard. This is the mode that a Forth system is in after it starts
7583: up.
7584:
7585: @cindex input buffer
7586: @cindex terminal input buffer
7587: The text interpreter works from an area of memory called the @dfn{input
7588: buffer}@footnote{When the text interpreter is processing input from the
7589: keyboard, this area of memory is called the @dfn{terminal input buffer}
7590: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7591: @code{#TIB}.}, which stores your keyboard input when you press the
7592: @key{RET} key. Starting at the beginning of the input buffer, it skips
7593: leading spaces (called @dfn{delimiters}) then parses a string (a
7594: sequence of non-space characters) until it reaches either a space
7595: character or the end of the buffer. Having parsed a string, it makes two
7596: attempts to process it:
7597:
7598: @cindex dictionary
7599: @itemize @bullet
7600: @item
7601: It looks for the string in a @dfn{dictionary} of definitions. If the
7602: string is found, the string names a @dfn{definition} (also known as a
7603: @dfn{word}) and the dictionary search returns information that allows
7604: the text interpreter to perform the word's @dfn{interpretation
7605: semantics}. In most cases, this simply means that the word will be
7606: executed.
7607: @item
7608: If the string is not found in the dictionary, the text interpreter
7609: attempts to treat it as a number, using the rules described in
7610: @ref{Number Conversion}. If the string represents a legal number in the
7611: current radix, the number is pushed onto a parameter stack (the data
7612: stack for integers, the floating-point stack for floating-point
7613: numbers).
7614: @end itemize
7615:
7616: If both attempts fail, or if the word is found in the dictionary but has
7617: no interpretation semantics@footnote{This happens if the word was
7618: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7619: remainder of the input buffer, issues an error message and waits for
7620: more input. If one of the attempts succeeds, the text interpreter
7621: repeats the parsing process until the whole of the input buffer has been
7622: processed, at which point it prints the status message ``@code{ ok}''
7623: and waits for more input.
7624:
7625: @c anton: this should be in the input stream subsection (or below it)
7626:
7627: @cindex parse area
7628: The text interpreter keeps track of its position in the input buffer by
7629: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7630: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7631: of the input buffer. The region from offset @code{>IN @@} to the end of
7632: the input buffer is called the @dfn{parse area}@footnote{In other words,
7633: the text interpreter processes the contents of the input buffer by
7634: parsing strings from the parse area until the parse area is empty.}.
7635: This example shows how @code{>IN} changes as the text interpreter parses
7636: the input buffer:
7637:
7638: @example
7639: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7640: CR ." ->" TYPE ." <-" ; IMMEDIATE
7641:
7642: 1 2 3 remaining + remaining .
7643:
7644: : foo 1 2 3 remaining SWAP remaining ;
7645: @end example
7646:
7647: @noindent
7648: The result is:
7649:
7650: @example
7651: ->+ remaining .<-
7652: ->.<-5 ok
7653:
7654: ->SWAP remaining ;-<
7655: ->;<- ok
7656: @end example
7657:
7658: @cindex parsing words
7659: The value of @code{>IN} can also be modified by a word in the input
7660: buffer that is executed by the text interpreter. This means that a word
7661: can ``trick'' the text interpreter into either skipping a section of the
7662: input buffer@footnote{This is how parsing words work.} or into parsing a
7663: section twice. For example:
7664:
7665: @example
7666: : lat ." <<foo>>" ;
7667: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
7668: @end example
7669:
7670: @noindent
7671: When @code{flat} is executed, this output is produced@footnote{Exercise
7672: for the reader: what would happen if the @code{3} were replaced with
7673: @code{4}?}:
7674:
7675: @example
7676: <<bar>><<foo>>
7677: @end example
7678:
7679: This technique can be used to work around some of the interoperability
7680: problems of parsing words. Of course, it's better to avoid parsing
7681: words where possible.
7682:
7683: @noindent
7684: Two important notes about the behaviour of the text interpreter:
7685:
7686: @itemize @bullet
7687: @item
7688: It processes each input string to completion before parsing additional
7689: characters from the input buffer.
7690: @item
7691: It treats the input buffer as a read-only region (and so must your code).
7692: @end itemize
7693:
7694: @noindent
7695: When the text interpreter is in compile state, its behaviour changes in
7696: these ways:
7697:
7698: @itemize @bullet
7699: @item
7700: If a parsed string is found in the dictionary, the text interpreter will
7701: perform the word's @dfn{compilation semantics}. In most cases, this
7702: simply means that the execution semantics of the word will be appended
7703: to the current definition.
7704: @item
7705: When a number is encountered, it is compiled into the current definition
7706: (as a literal) rather than being pushed onto a parameter stack.
7707: @item
7708: If an error occurs, @code{state} is modified to put the text interpreter
7709: back into interpret state.
7710: @item
7711: Each time a line is entered from the keyboard, Gforth prints
7712: ``@code{ compiled}'' rather than `` @code{ok}''.
7713: @end itemize
7714:
7715: @cindex text interpreter - input sources
7716: When the text interpreter is using an input device other than the
7717: keyboard, its behaviour changes in these ways:
7718:
7719: @itemize @bullet
7720: @item
7721: When the parse area is empty, the text interpreter attempts to refill
7722: the input buffer from the input source. When the input source is
7723: exhausted, the input source is set back to the previous input source.
7724: @item
7725: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7726: time the parse area is emptied.
7727: @item
7728: If an error occurs, the input source is set back to the user input
7729: device.
7730: @end itemize
7731:
7732: You can read about this in more detail in @ref{Input Sources}.
7733:
7734: doc->in
7735: doc-source
7736:
7737: doc-tib
7738: doc-#tib
7739:
7740:
7741: @menu
7742: * Input Sources::
7743: * Number Conversion::
7744: * Interpret/Compile states::
7745: * Interpreter Directives::
7746: @end menu
7747:
7748: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7749: @subsection Input Sources
7750: @cindex input sources
7751: @cindex text interpreter - input sources
7752:
7753: By default, the text interpreter processes input from the user input
7754: device (the keyboard) when Forth starts up. The text interpreter can
7755: process input from any of these sources:
7756:
7757: @itemize @bullet
7758: @item
7759: The user input device -- the keyboard.
7760: @item
7761: A file, using the words described in @ref{Forth source files}.
7762: @item
7763: A block, using the words described in @ref{Blocks}.
7764: @item
7765: A text string, using @code{evaluate}.
7766: @end itemize
7767:
7768: A program can identify the current input device from the values of
7769: @code{source-id} and @code{blk}.
7770:
7771:
7772: doc-source-id
7773: doc-blk
7774:
7775: doc-save-input
7776: doc-restore-input
7777:
7778: doc-evaluate
7779:
7780:
7781:
7782: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
7783: @subsection Number Conversion
7784: @cindex number conversion
7785: @cindex double-cell numbers, input format
7786: @cindex input format for double-cell numbers
7787: @cindex single-cell numbers, input format
7788: @cindex input format for single-cell numbers
7789: @cindex floating-point numbers, input format
7790: @cindex input format for floating-point numbers
7791:
7792: This section describes the rules that the text interpreter uses when it
7793: tries to convert a string into a number.
7794:
7795: Let <digit> represent any character that is a legal digit in the current
7796: number base@footnote{For example, 0-9 when the number base is decimal or
7797: 0-9, A-F when the number base is hexadecimal.}.
7798:
7799: Let <decimal digit> represent any character in the range 0-9.
7800:
7801: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7802: in the braces (@i{a} or @i{b} or neither).
7803:
7804: Let * represent any number of instances of the previous character
7805: (including none).
7806:
7807: Let any other character represent itself.
7808:
7809: @noindent
7810: Now, the conversion rules are:
7811:
7812: @itemize @bullet
7813: @item
7814: A string of the form <digit><digit>* is treated as a single-precision
7815: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
7816: @item
7817: A string of the form -<digit><digit>* is treated as a single-precision
7818: (cell-sized) negative integer, and is represented using 2's-complement
7819: arithmetic. Examples are -45 -5681 -0
7820: @item
7821: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
7822: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7823: (all three of these represent the same number).
7824: @item
7825: A string of the form -<digit><digit>*.<digit>* is treated as a
7826: double-precision (double-cell-sized) negative integer, and is
7827: represented using 2's-complement arithmetic. Examples are -3465. -3.465
7828: -34.65 (all three of these represent the same number).
7829: @item
7830: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7831: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
7832: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
7833: number) +12.E-4
7834: @end itemize
7835:
7836: By default, the number base used for integer number conversion is given
7837: by the contents of the variable @code{base}. Note that a lot of
7838: confusion can result from unexpected values of @code{base}. If you
7839: change @code{base} anywhere, make sure to save the old value and restore
7840: it afterwards. In general I recommend keeping @code{base} decimal, and
7841: using the prefixes described below for the popular non-decimal bases.
7842:
7843: doc-dpl
7844: doc-base
7845: doc-hex
7846: doc-decimal
7847:
7848:
7849: @cindex '-prefix for character strings
7850: @cindex &-prefix for decimal numbers
7851: @cindex %-prefix for binary numbers
7852: @cindex $-prefix for hexadecimal numbers
7853: Gforth allows you to override the value of @code{base} by using a
7854: prefix@footnote{Some Forth implementations provide a similar scheme by
7855: implementing @code{$} etc. as parsing words that process the subsequent
7856: number in the input stream and push it onto the stack. For example, see
7857: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7858: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7859: is required between the prefix and the number.} before the first digit
7860: of an (integer) number. Four prefixes are supported:
7861:
7862: @itemize @bullet
7863: @item
7864: @code{&} -- decimal
7865: @item
7866: @code{%} -- binary
7867: @item
7868: @code{$} -- hexadecimal
7869: @item
7870: @code{'} -- base @code{max-char+1}
7871: @end itemize
7872:
7873: Here are some examples, with the equivalent decimal number shown after
7874: in braces:
7875:
7876: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7877: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
7878: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
7879: &905 (905), $abc (2478), $ABC (2478).
7880:
7881: @cindex number conversion - traps for the unwary
7882: @noindent
7883: Number conversion has a number of traps for the unwary:
7884:
7885: @itemize @bullet
7886: @item
7887: You cannot determine the current number base using the code sequence
7888: @code{base @@ .} -- the number base is always 10 in the current number
7889: base. Instead, use something like @code{base @@ dec.}
7890: @item
7891: If the number base is set to a value greater than 14 (for example,
7892: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7893: it to be intepreted as either a single-precision integer or a
7894: floating-point number (Gforth treats it as an integer). The ambiguity
7895: can be resolved by explicitly stating the sign of the mantissa and/or
7896: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7897: ambiguity arises; either representation will be treated as a
7898: floating-point number.
7899: @item
7900: There is a word @code{bin} but it does @i{not} set the number base!
7901: It is used to specify file types.
7902: @item
7903: ANS Forth requires the @code{.} of a double-precision number to be the
7904: final character in the string. Gforth allows the @code{.} to be
7905: anywhere after the first digit.
7906: @item
7907: The number conversion process does not check for overflow.
7908: @item
7909: In an ANS Forth program @code{base} is required to be decimal when
7910: converting floating-point numbers. In Gforth, number conversion to
7911: floating-point numbers always uses base &10, irrespective of the value
7912: of @code{base}.
7913: @end itemize
7914:
7915: You can read numbers into your programs with the words described in
7916: @ref{Input}.
7917:
7918: @node Interpret/Compile states, Interpreter Directives, Number Conversion, The Text Interpreter
7919: @subsection Interpret/Compile states
7920: @cindex Interpret/Compile states
7921:
7922: A standard program is not permitted to change @code{state}
7923: explicitly. However, it can change @code{state} implicitly, using the
7924: words @code{[} and @code{]}. When @code{[} is executed it switches
7925: @code{state} to interpret state, and therefore the text interpreter
7926: starts interpreting. When @code{]} is executed it switches @code{state}
7927: to compile state and therefore the text interpreter starts
7928: compiling. The most common usage for these words is for switching into
7929: interpret state and back from within a colon definition; this technique
7930: can be used to compile a literal (for an example, @pxref{Literals}) or
7931: for conditional compilation (for an example, @pxref{Interpreter
7932: Directives}).
7933:
7934:
7935: @c This is a bad example: It's non-standard, and it's not necessary.
7936: @c However, I can't think of a good example for switching into compile
7937: @c state when there is no current word (@code{state}-smart words are not a
7938: @c good reason). So maybe we should use an example for switching into
7939: @c interpret @code{state} in a colon def. - anton
7940: @c nac-> I agree. I started out by putting in the example, then realised
7941: @c that it was non-ANS, so wrote more words around it. I hope this
7942: @c re-written version is acceptable to you. I do want to keep the example
7943: @c as it is helpful for showing what is and what is not portable, particularly
7944: @c where it outlaws a style in common use.
7945:
7946: @c anton: it's more important to show what's portable. After we have done
7947: @c that, we can also show what's not. In any case, I have written a
7948: @c section Compiling Words which also deals with [ ].
7949:
7950: @c !! The following example does not work in Gforth 0.5.9 or later.
7951:
7952: @c @code{[} and @code{]} also give you the ability to switch into compile
7953: @c state and back, but we cannot think of any useful Standard application
7954: @c for this ability. Pre-ANS Forth textbooks have examples like this:
7955:
7956: @c @example
7957: @c : AA ." this is A" ;
7958: @c : BB ." this is B" ;
7959: @c : CC ." this is C" ;
7960:
7961: @c create table ] aa bb cc [
7962:
7963: @c : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7964: @c cells table + @@ execute ;
7965: @c @end example
7966:
7967: @c This example builds a jump table; @code{0 go} will display ``@code{this
7968: @c is A}''. Using @code{[} and @code{]} in this example is equivalent to
7969: @c defining @code{table} like this:
7970:
7971: @c @example
7972: @c create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
7973: @c @end example
7974:
7975: @c The problem with this code is that the definition of @code{table} is not
7976: @c portable -- it @i{compile}s execution tokens into code space. Whilst it
7977: @c @i{may} work on systems where code space and data space co-incide, the
7978: @c Standard only allows data space to be assigned for a @code{CREATE}d
7979: @c word. In addition, the Standard only allows @code{@@} to access data
7980: @c space, whilst this example is using it to access code space. The only
7981: @c portable, Standard way to build this table is to build it in data space,
7982: @c like this:
7983:
7984: @c @example
7985: @c create table ' aa , ' bb , ' cc ,
7986: @c @end example
7987:
7988: @c doc-state
7989:
7990:
7991: @node Interpreter Directives, , Interpret/Compile states, The Text Interpreter
7992: @subsection Interpreter Directives
7993: @cindex interpreter directives
7994: @cindex conditional compilation
7995:
7996: These words are usually used in interpret state; typically to control
7997: which parts of a source file are processed by the text
7998: interpreter. There are only a few ANS Forth Standard words, but Gforth
7999: supplements these with a rich set of immediate control structure words
8000: to compensate for the fact that the non-immediate versions can only be
8001: used in compile state (@pxref{Control Structures}). Typical usages:
8002:
8003: @example
8004: FALSE Constant HAVE-ASSEMBLER
8005: .
8006: .
8007: HAVE-ASSEMBLER [IF]
8008: : ASSEMBLER-FEATURE
8009: ...
8010: ;
8011: [ENDIF]
8012: .
8013: .
8014: : SEE
8015: ... \ general-purpose SEE code
8016: [ HAVE-ASSEMBLER [IF] ]
8017: ... \ assembler-specific SEE code
8018: [ [ENDIF] ]
8019: ;
8020: @end example
8021:
8022:
8023: doc-[IF]
8024: doc-[ELSE]
8025: doc-[THEN]
8026: doc-[ENDIF]
8027:
8028: doc-[IFDEF]
8029: doc-[IFUNDEF]
8030:
8031: doc-[?DO]
8032: doc-[DO]
8033: doc-[FOR]
8034: doc-[LOOP]
8035: doc-[+LOOP]
8036: doc-[NEXT]
8037:
8038: doc-[BEGIN]
8039: doc-[UNTIL]
8040: doc-[AGAIN]
8041: doc-[WHILE]
8042: doc-[REPEAT]
8043:
8044:
8045: @c -------------------------------------------------------------
8046: @node Word Lists, Environmental Queries, The Text Interpreter, Words
8047: @section Word Lists
8048: @cindex word lists
8049: @cindex header space
8050:
8051: A wordlist is a list of named words; you can add new words and look up
8052: words by name (and you can remove words in a restricted way with
8053: markers). Every named (and @code{reveal}ed) word is in one wordlist.
8054:
8055: @cindex search order stack
8056: The text interpreter searches the wordlists present in the search order
8057: (a stack of wordlists), from the top to the bottom. Within each
8058: wordlist, the search starts conceptually at the newest word; i.e., if
8059: two words in a wordlist have the same name, the newer word is found.
8060:
8061: @cindex compilation word list
8062: New words are added to the @dfn{compilation wordlist} (aka current
8063: wordlist).
8064:
8065: @cindex wid
8066: A word list is identified by a cell-sized word list identifier (@i{wid})
8067: in much the same way as a file is identified by a file handle. The
8068: numerical value of the wid has no (portable) meaning, and might change
8069: from session to session.
8070:
8071: The ANS Forth ``Search order'' word set is intended to provide a set of
8072: low-level tools that allow various different schemes to be
8073: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
8074: word. @file{compat/vocabulary.fs} provides an implementation in ANS
8075: Forth.
8076:
8077: @comment TODO: locals section refers to here, saying that every word list (aka
8078: @comment vocabulary) has its own methods for searching etc. Need to document that.
8079: @c anton: but better in a separate subsection on wordlist internals
8080:
8081: @comment TODO: document markers, reveal, tables, mappedwordlist
8082:
8083: @comment the gforthman- prefix is used to pick out the true definition of a
8084: @comment word from the source files, rather than some alias.
8085:
8086: doc-forth-wordlist
8087: doc-definitions
8088: doc-get-current
8089: doc-set-current
8090: doc-get-order
8091: doc---gforthman-set-order
8092: doc-wordlist
8093: doc-table
8094: doc->order
8095: doc-previous
8096: doc-also
8097: doc---gforthman-forth
8098: doc-only
8099: doc---gforthman-order
8100:
8101: doc-find
8102: doc-search-wordlist
8103:
8104: doc-words
8105: doc-vlist
8106: @c doc-words-deferred
8107:
8108: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
8109: doc-root
8110: doc-vocabulary
8111: doc-seal
8112: doc-vocs
8113: doc-current
8114: doc-context
8115:
8116:
8117: @menu
8118: * Vocabularies::
8119: * Why use word lists?::
8120: * Word list example::
8121: @end menu
8122:
8123: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
8124: @subsection Vocabularies
8125: @cindex Vocabularies, detailed explanation
8126:
8127: Here is an example of creating and using a new wordlist using ANS
8128: Forth words:
8129:
8130: @example
8131: wordlist constant my-new-words-wordlist
8132: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
8133:
8134: \ add it to the search order
8135: also my-new-words
8136:
8137: \ alternatively, add it to the search order and make it
8138: \ the compilation word list
8139: also my-new-words definitions
8140: \ type "order" to see the problem
8141: @end example
8142:
8143: The problem with this example is that @code{order} has no way to
8144: associate the name @code{my-new-words} with the wid of the word list (in
8145: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
8146: that has no associated name). There is no Standard way of associating a
8147: name with a wid.
8148:
8149: In Gforth, this example can be re-coded using @code{vocabulary}, which
8150: associates a name with a wid:
8151:
8152: @example
8153: vocabulary my-new-words
8154:
8155: \ add it to the search order
8156: also my-new-words
8157:
8158: \ alternatively, add it to the search order and make it
8159: \ the compilation word list
8160: my-new-words definitions
8161: \ type "order" to see that the problem is solved
8162: @end example
8163:
8164:
8165: @node Why use word lists?, Word list example, Vocabularies, Word Lists
8166: @subsection Why use word lists?
8167: @cindex word lists - why use them?
8168:
8169: Here are some reasons why people use wordlists:
8170:
8171: @itemize @bullet
8172:
8173: @c anton: Gforth's hashing implementation makes the search speed
8174: @c independent from the number of words. But it is linear with the number
8175: @c of wordlists that have to be searched, so in effect using more wordlists
8176: @c actually slows down compilation.
8177:
8178: @c @item
8179: @c To improve compilation speed by reducing the number of header space
8180: @c entries that must be searched. This is achieved by creating a new
8181: @c word list that contains all of the definitions that are used in the
8182: @c definition of a Forth system but which would not usually be used by
8183: @c programs running on that system. That word list would be on the search
8184: @c list when the Forth system was compiled but would be removed from the
8185: @c search list for normal operation. This can be a useful technique for
8186: @c low-performance systems (for example, 8-bit processors in embedded
8187: @c systems) but is unlikely to be necessary in high-performance desktop
8188: @c systems.
8189:
8190: @item
8191: To prevent a set of words from being used outside the context in which
8192: they are valid. Two classic examples of this are an integrated editor
8193: (all of the edit commands are defined in a separate word list; the
8194: search order is set to the editor word list when the editor is invoked;
8195: the old search order is restored when the editor is terminated) and an
8196: integrated assembler (the op-codes for the machine are defined in a
8197: separate word list which is used when a @code{CODE} word is defined).
8198:
8199: @item
8200: To organize the words of an application or library into a user-visible
8201: set (in @code{forth-wordlist} or some other common wordlist) and a set
8202: of helper words used just for the implementation (hidden in a separate
8203: wordlist). This keeps @code{words}' output smaller, separates
8204: implementation and interface, and reduces the chance of name conflicts
8205: within the common wordlist.
8206:
8207: @item
8208: To prevent a name-space clash between multiple definitions with the same
8209: name. For example, when building a cross-compiler you might have a word
8210: @code{IF} that generates conditional code for your target system. By
8211: placing this definition in a different word list you can control whether
8212: the host system's @code{IF} or the target system's @code{IF} get used in
8213: any particular context by controlling the order of the word lists on the
8214: search order stack.
8215:
8216: @end itemize
8217:
8218: The downsides of using wordlists are:
8219:
8220: @itemize
8221:
8222: @item
8223: Debugging becomes more cumbersome.
8224:
8225: @item
8226: Name conflicts worked around with wordlists are still there, and you
8227: have to arrange the search order carefully to get the desired results;
8228: if you forget to do that, you get hard-to-find errors (as in any case
8229: where you read the code differently from the compiler; @code{see} can
8230: help seeing which of several possible words the name resolves to in such
8231: cases). @code{See} displays just the name of the words, not what
8232: wordlist they belong to, so it might be misleading. Using unique names
8233: is a better approach to avoid name conflicts.
8234:
8235: @item
8236: You have to explicitly undo any changes to the search order. In many
8237: cases it would be more convenient if this happened implicitly. Gforth
8238: currently does not provide such a feature, but it may do so in the
8239: future.
8240: @end itemize
8241:
8242:
8243: @node Word list example, , Why use word lists?, Word Lists
8244: @subsection Word list example
8245: @cindex word lists - example
8246:
8247: The following example is from the
8248: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
8249: garbage collector} and uses wordlists to separate public words from
8250: helper words:
8251:
8252: @example
8253: get-current ( wid )
8254: vocabulary garbage-collector also garbage-collector definitions
8255: ... \ define helper words
8256: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
8257: ... \ define the public (i.e., API) words
8258: \ they can refer to the helper words
8259: previous \ restore original search order (helper words become invisible)
8260: @end example
8261:
8262: @c -------------------------------------------------------------
8263: @node Environmental Queries, Files, Word Lists, Words
8264: @section Environmental Queries
8265: @cindex environmental queries
8266:
8267: ANS Forth introduced the idea of ``environmental queries'' as a way
8268: for a program running on a system to determine certain characteristics of the system.
8269: The Standard specifies a number of strings that might be recognised by a system.
8270:
8271: The Standard requires that the header space used for environmental queries
8272: be distinct from the header space used for definitions.
8273:
8274: Typically, environmental queries are supported by creating a set of
8275: definitions in a word list that is @i{only} used during environmental
8276: queries; that is what Gforth does. There is no Standard way of adding
8277: definitions to the set of recognised environmental queries, but any
8278: implementation that supports the loading of optional word sets must have
8279: some mechanism for doing this (after loading the word set, the
8280: associated environmental query string must return @code{true}). In
8281: Gforth, the word list used to honour environmental queries can be
8282: manipulated just like any other word list.
8283:
8284:
8285: doc-environment?
8286: doc-environment-wordlist
8287:
8288: doc-gforth
8289: doc-os-class
8290:
8291:
8292: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
8293: returning two items on the stack, querying it using @code{environment?}
8294: will return an additional item; the @code{true} flag that shows that the
8295: string was recognised.
8296:
8297: @comment TODO Document the standard strings or note where they are documented herein
8298:
8299: Here are some examples of using environmental queries:
8300:
8301: @example
8302: s" address-unit-bits" environment? 0=
8303: [IF]
8304: cr .( environmental attribute address-units-bits unknown... ) cr
8305: [ELSE]
8306: drop \ ensure balanced stack effect
8307: [THEN]
8308:
8309: \ this might occur in the prelude of a standard program that uses THROW
8310: s" exception" environment? [IF]
8311: 0= [IF]
8312: : throw abort" exception thrown" ;
8313: [THEN]
8314: [ELSE] \ we don't know, so make sure
8315: : throw abort" exception thrown" ;
8316: [THEN]
8317:
8318: s" gforth" environment? [IF] .( Gforth version ) TYPE
8319: [ELSE] .( Not Gforth..) [THEN]
8320:
8321: \ a program using v*
8322: s" gforth" environment? [IF]
8323: s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
8324: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8325: >r swap 2swap swap 0e r> 0 ?DO
8326: dup f@ over + 2swap dup f@ f* f+ over + 2swap
8327: LOOP
8328: 2drop 2drop ;
8329: [THEN]
8330: [ELSE] \
8331: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8332: ...
8333: [THEN]
8334: @end example
8335:
8336: Here is an example of adding a definition to the environment word list:
8337:
8338: @example
8339: get-current environment-wordlist set-current
8340: true constant block
8341: true constant block-ext
8342: set-current
8343: @end example
8344:
8345: You can see what definitions are in the environment word list like this:
8346:
8347: @example
8348: environment-wordlist >order words previous
8349: @end example
8350:
8351:
8352: @c -------------------------------------------------------------
8353: @node Files, Blocks, Environmental Queries, Words
8354: @section Files
8355: @cindex files
8356: @cindex I/O - file-handling
8357:
8358: Gforth provides facilities for accessing files that are stored in the
8359: host operating system's file-system. Files that are processed by Gforth
8360: can be divided into two categories:
8361:
8362: @itemize @bullet
8363: @item
8364: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
8365: @item
8366: Files that are processed by some other program (@dfn{general files}).
8367: @end itemize
8368:
8369: @menu
8370: * Forth source files::
8371: * General files::
8372: * Search Paths::
8373: @end menu
8374:
8375: @c -------------------------------------------------------------
8376: @node Forth source files, General files, Files, Files
8377: @subsection Forth source files
8378: @cindex including files
8379: @cindex Forth source files
8380:
8381: The simplest way to interpret the contents of a file is to use one of
8382: these two formats:
8383:
8384: @example
8385: include mysource.fs
8386: s" mysource.fs" included
8387: @end example
8388:
8389: You usually want to include a file only if it is not included already
8390: (by, say, another source file). In that case, you can use one of these
8391: three formats:
8392:
8393: @example
8394: require mysource.fs
8395: needs mysource.fs
8396: s" mysource.fs" required
8397: @end example
8398:
8399: @cindex stack effect of included files
8400: @cindex including files, stack effect
8401: It is good practice to write your source files such that interpreting them
8402: does not change the stack. Source files designed in this way can be used with
8403: @code{required} and friends without complications. For example:
8404:
8405: @example
8406: 1024 require foo.fs drop
8407: @end example
8408:
8409: Here you want to pass the argument 1024 (e.g., a buffer size) to
8410: @file{foo.fs}. Interpreting @file{foo.fs} has the stack effect ( n -- n
8411: ), which allows its use with @code{require}. Of course with such
8412: parameters to required files, you have to ensure that the first
8413: @code{require} fits for all uses (i.e., @code{require} it early in the
8414: master load file).
8415:
8416: doc-include-file
8417: doc-included
8418: doc-included?
8419: doc-include
8420: doc-required
8421: doc-require
8422: doc-needs
8423: @c doc-init-included-files @c internal
8424: doc-sourcefilename
8425: doc-sourceline#
8426:
8427: A definition in ANS Forth for @code{required} is provided in
8428: @file{compat/required.fs}.
8429:
8430: @c -------------------------------------------------------------
8431: @node General files, Search Paths, Forth source files, Files
8432: @subsection General files
8433: @cindex general files
8434: @cindex file-handling
8435:
8436: Files are opened/created by name and type. The following file access
8437: methods (FAMs) are recognised:
8438:
8439: @cindex fam (file access method)
8440: doc-r/o
8441: doc-r/w
8442: doc-w/o
8443: doc-bin
8444:
8445:
8446: When a file is opened/created, it returns a file identifier,
8447: @i{wfileid} that is used for all other file commands. All file
8448: commands also return a status value, @i{wior}, that is 0 for a
8449: successful operation and an implementation-defined non-zero value in the
8450: case of an error.
8451:
8452:
8453: doc-open-file
8454: doc-create-file
8455:
8456: doc-close-file
8457: doc-delete-file
8458: doc-rename-file
8459: doc-read-file
8460: doc-read-line
8461: doc-write-file
8462: doc-write-line
8463: doc-emit-file
8464: doc-flush-file
8465:
8466: doc-file-status
8467: doc-file-position
8468: doc-reposition-file
8469: doc-file-size
8470: doc-resize-file
8471:
8472: doc-slurp-file
8473: doc-slurp-fid
8474:
8475: @c ---------------------------------------------------------
8476: @node Search Paths, , General files, Files
8477: @subsection Search Paths
8478: @cindex path for @code{included}
8479: @cindex file search path
8480: @cindex @code{include} search path
8481: @cindex search path for files
8482:
8483: If you specify an absolute filename (i.e., a filename starting with
8484: @file{/} or @file{~}, or with @file{:} in the second position (as in
8485: @samp{C:...})) for @code{included} and friends, that file is included
8486: just as you would expect.
8487:
8488: If the filename starts with @file{./}, this refers to the directory that
8489: the present file was @code{included} from. This allows files to include
8490: other files relative to their own position (irrespective of the current
8491: working directory or the absolute position). This feature is essential
8492: for libraries consisting of several files, where a file may include
8493: other files from the library. It corresponds to @code{#include "..."}
8494: in C. If the current input source is not a file, @file{.} refers to the
8495: directory of the innermost file being included, or, if there is no file
8496: being included, to the current working directory.
8497:
8498: For relative filenames (not starting with @file{./}), Gforth uses a
8499: search path similar to Forth's search order (@pxref{Word Lists}). It
8500: tries to find the given filename in the directories present in the path,
8501: and includes the first one it finds. There are separate search paths for
8502: Forth source files and general files. If the search path contains the
8503: directory @file{.}, this refers to the directory of the current file, or
8504: the working directory, as if the file had been specified with @file{./}.
8505:
8506: Use @file{~+} to refer to the current working directory (as in the
8507: @code{bash}).
8508:
8509: @c anton: fold the following subsubsections into this subsection?
8510:
8511: @menu
8512: * Source Search Paths::
8513: * General Search Paths::
8514: @end menu
8515:
8516: @c ---------------------------------------------------------
8517: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
8518: @subsubsection Source Search Paths
8519: @cindex search path control, source files
8520:
8521: The search path is initialized when you start Gforth (@pxref{Invoking
8522: Gforth}). You can display it and change it using @code{fpath} in
8523: combination with the general path handling words.
8524:
8525: doc-fpath
8526: @c the functionality of the following words is easily available through
8527: @c fpath and the general path words. The may go away.
8528: @c doc-.fpath
8529: @c doc-fpath+
8530: @c doc-fpath=
8531: @c doc-open-fpath-file
8532:
8533: @noindent
8534: Here is an example of using @code{fpath} and @code{require}:
8535:
8536: @example
8537: fpath path= /usr/lib/forth/|./
8538: require timer.fs
8539: @end example
8540:
8541:
8542: @c ---------------------------------------------------------
8543: @node General Search Paths, , Source Search Paths, Search Paths
8544: @subsubsection General Search Paths
8545: @cindex search path control, source files
8546:
8547: Your application may need to search files in several directories, like
8548: @code{included} does. To facilitate this, Gforth allows you to define
8549: and use your own search paths, by providing generic equivalents of the
8550: Forth search path words:
8551:
8552: doc-open-path-file
8553: doc-path-allot
8554: doc-clear-path
8555: doc-also-path
8556: doc-.path
8557: doc-path+
8558: doc-path=
8559:
8560: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
8561:
8562: Here's an example of creating an empty search path:
8563: @c
8564: @example
8565: create mypath 500 path-allot \ maximum length 500 chars (is checked)
8566: @end example
8567:
8568: @c -------------------------------------------------------------
8569: @node Blocks, Other I/O, Files, Words
8570: @section Blocks
8571: @cindex I/O - blocks
8572: @cindex blocks
8573:
8574: When you run Gforth on a modern desk-top computer, it runs under the
8575: control of an operating system which provides certain services. One of
8576: these services is @var{file services}, which allows Forth source code
8577: and data to be stored in files and read into Gforth (@pxref{Files}).
8578:
8579: Traditionally, Forth has been an important programming language on
8580: systems where it has interfaced directly to the underlying hardware with
8581: no intervening operating system. Forth provides a mechanism, called
8582: @dfn{blocks}, for accessing mass storage on such systems.
8583:
8584: A block is a 1024-byte data area, which can be used to hold data or
8585: Forth source code. No structure is imposed on the contents of the
8586: block. A block is identified by its number; blocks are numbered
8587: contiguously from 1 to an implementation-defined maximum.
8588:
8589: A typical system that used blocks but no operating system might use a
8590: single floppy-disk drive for mass storage, with the disks formatted to
8591: provide 256-byte sectors. Blocks would be implemented by assigning the
8592: first four sectors of the disk to block 1, the second four sectors to
8593: block 2 and so on, up to the limit of the capacity of the disk. The disk
8594: would not contain any file system information, just the set of blocks.
8595:
8596: @cindex blocks file
8597: On systems that do provide file services, blocks are typically
8598: implemented by storing a sequence of blocks within a single @dfn{blocks
8599: file}. The size of the blocks file will be an exact multiple of 1024
8600: bytes, corresponding to the number of blocks it contains. This is the
8601: mechanism that Gforth uses.
8602:
8603: @cindex @file{blocks.fb}
8604: Only one blocks file can be open at a time. If you use block words without
8605: having specified a blocks file, Gforth defaults to the blocks file
8606: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
8607: locate a blocks file (@pxref{Source Search Paths}).
8608:
8609: @cindex block buffers
8610: When you read and write blocks under program control, Gforth uses a
8611: number of @dfn{block buffers} as intermediate storage. These buffers are
8612: not used when you use @code{load} to interpret the contents of a block.
8613:
8614: The behaviour of the block buffers is analagous to that of a cache.
8615: Each block buffer has three states:
8616:
8617: @itemize @bullet
8618: @item
8619: Unassigned
8620: @item
8621: Assigned-clean
8622: @item
8623: Assigned-dirty
8624: @end itemize
8625:
8626: Initially, all block buffers are @i{unassigned}. In order to access a
8627: block, the block (specified by its block number) must be assigned to a
8628: block buffer.
8629:
8630: The assignment of a block to a block buffer is performed by @code{block}
8631: or @code{buffer}. Use @code{block} when you wish to modify the existing
8632: contents of a block. Use @code{buffer} when you don't care about the
8633: existing contents of the block@footnote{The ANS Forth definition of
8634: @code{buffer} is intended not to cause disk I/O; if the data associated
8635: with the particular block is already stored in a block buffer due to an
8636: earlier @code{block} command, @code{buffer} will return that block
8637: buffer and the existing contents of the block will be
8638: available. Otherwise, @code{buffer} will simply assign a new, empty
8639: block buffer for the block.}.
8640:
8641: Once a block has been assigned to a block buffer using @code{block} or
8642: @code{buffer}, that block buffer becomes the @i{current block
8643: buffer}. Data may only be manipulated (read or written) within the
8644: current block buffer.
8645:
8646: When the contents of the current block buffer has been modified it is
8647: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
8648: either abandon the changes (by doing nothing) or mark the block as
8649: changed (assigned-dirty), using @code{update}. Using @code{update} does
8650: not change the blocks file; it simply changes a block buffer's state to
8651: @i{assigned-dirty}. The block will be written implicitly when it's
8652: buffer is needed for another block, or explicitly by @code{flush} or
8653: @code{save-buffers}.
8654:
8655: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
8656: blocks file on disk. Leaving Gforth with @code{bye} also performs a
8657: @code{flush}.
8658:
8659: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
8660: algorithm to assign a block buffer to a block. That means that any
8661: particular block can only be assigned to one specific block buffer,
8662: called (for the particular operation) the @i{victim buffer}. If the
8663: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8664: the new block immediately. If it is @i{assigned-dirty} its current
8665: contents are written back to the blocks file on disk before it is
8666: allocated to the new block.
8667:
8668: Although no structure is imposed on the contents of a block, it is
8669: traditional to display the contents as 16 lines each of 64 characters. A
8670: block provides a single, continuous stream of input (for example, it
8671: acts as a single parse area) -- there are no end-of-line characters
8672: within a block, and no end-of-file character at the end of a
8673: block. There are two consequences of this:
8674:
8675: @itemize @bullet
8676: @item
8677: The last character of one line wraps straight into the first character
8678: of the following line
8679: @item
8680: The word @code{\} -- comment to end of line -- requires special
8681: treatment; in the context of a block it causes all characters until the
8682: end of the current 64-character ``line'' to be ignored.
8683: @end itemize
8684:
8685: In Gforth, when you use @code{block} with a non-existent block number,
8686: the current blocks file will be extended to the appropriate size and the
8687: block buffer will be initialised with spaces.
8688:
8689: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8690: for details) but doesn't encourage the use of blocks; the mechanism is
8691: only provided for backward compatibility -- ANS Forth requires blocks to
8692: be available when files are.
8693:
8694: Common techniques that are used when working with blocks include:
8695:
8696: @itemize @bullet
8697: @item
8698: A screen editor that allows you to edit blocks without leaving the Forth
8699: environment.
8700: @item
8701: Shadow screens; where every code block has an associated block
8702: containing comments (for example: code in odd block numbers, comments in
8703: even block numbers). Typically, the block editor provides a convenient
8704: mechanism to toggle between code and comments.
8705: @item
8706: Load blocks; a single block (typically block 1) contains a number of
8707: @code{thru} commands which @code{load} the whole of the application.
8708: @end itemize
8709:
8710: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8711: integrated into a Forth programming environment.
8712:
8713: @comment TODO what about errors on open-blocks?
8714:
8715: doc-open-blocks
8716: doc-use
8717: doc-block-offset
8718: doc-get-block-fid
8719: doc-block-position
8720:
8721: doc-list
8722: doc-scr
8723:
8724: doc---gforthman-block
8725: doc-buffer
8726:
8727: doc-empty-buffers
8728: doc-empty-buffer
8729: doc-update
8730: doc-updated?
8731: doc-save-buffers
8732: doc-save-buffer
8733: doc-flush
8734:
8735: doc-load
8736: doc-thru
8737: doc-+load
8738: doc-+thru
8739: doc---gforthman--->
8740: doc-block-included
8741:
8742:
8743: @c -------------------------------------------------------------
8744: @node Other I/O, Locals, Blocks, Words
8745: @section Other I/O
8746: @cindex I/O - keyboard and display
8747:
8748: @menu
8749: * Simple numeric output:: Predefined formats
8750: * Formatted numeric output:: Formatted (pictured) output
8751: * String Formats:: How Forth stores strings in memory
8752: * Displaying characters and strings:: Other stuff
8753: * Input:: Input
8754: @end menu
8755:
8756: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8757: @subsection Simple numeric output
8758: @cindex numeric output - simple/free-format
8759:
8760: The simplest output functions are those that display numbers from the
8761: data or floating-point stacks. Floating-point output is always displayed
8762: using base 10. Numbers displayed from the data stack use the value stored
8763: in @code{base}.
8764:
8765:
8766: doc-.
8767: doc-dec.
8768: doc-hex.
8769: doc-u.
8770: doc-.r
8771: doc-u.r
8772: doc-d.
8773: doc-ud.
8774: doc-d.r
8775: doc-ud.r
8776: doc-f.
8777: doc-fe.
8778: doc-fs.
8779:
8780:
8781: Examples of printing the number 1234.5678E23 in the different floating-point output
8782: formats are shown below:
8783:
8784: @example
8785: f. 123456779999999000000000000.
8786: fe. 123.456779999999E24
8787: fs. 1.23456779999999E26
8788: @end example
8789:
8790:
8791: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8792: @subsection Formatted numeric output
8793: @cindex formatted numeric output
8794: @cindex pictured numeric output
8795: @cindex numeric output - formatted
8796:
8797: Forth traditionally uses a technique called @dfn{pictured numeric
8798: output} for formatted printing of integers. In this technique, digits
8799: are extracted from the number (using the current output radix defined by
8800: @code{base}), converted to ASCII codes and appended to a string that is
8801: built in a scratch-pad area of memory (@pxref{core-idef,
8802: Implementation-defined options, Implementation-defined
8803: options}). Arbitrary characters can be appended to the string during the
8804: extraction process. The completed string is specified by an address
8805: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8806: under program control.
8807:
8808: All of the integer output words described in the previous section
8809: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
8810: numeric output.
8811:
8812: Three important things to remember about pictured numeric output:
8813:
8814: @itemize @bullet
8815: @item
8816: It always operates on double-precision numbers; to display a
8817: single-precision number, convert it first (for ways of doing this
8818: @pxref{Double precision}).
8819: @item
8820: It always treats the double-precision number as though it were
8821: unsigned. The examples below show ways of printing signed numbers.
8822: @item
8823: The string is built up from right to left; least significant digit first.
8824: @end itemize
8825:
8826:
8827: doc-<#
8828: doc-<<#
8829: doc-#
8830: doc-#s
8831: doc-hold
8832: doc-sign
8833: doc-#>
8834: doc-#>>
8835:
8836: doc-represent
8837:
8838:
8839: @noindent
8840: Here are some examples of using pictured numeric output:
8841:
8842: @example
8843: : my-u. ( u -- )
8844: \ Simplest use of pns.. behaves like Standard u.
8845: 0 \ convert to unsigned double
8846: <<# \ start conversion
8847: #s \ convert all digits
8848: #> \ complete conversion
8849: TYPE SPACE \ display, with trailing space
8850: #>> ; \ release hold area
8851:
8852: : cents-only ( u -- )
8853: 0 \ convert to unsigned double
8854: <<# \ start conversion
8855: # # \ convert two least-significant digits
8856: #> \ complete conversion, discard other digits
8857: TYPE SPACE \ display, with trailing space
8858: #>> ; \ release hold area
8859:
8860: : dollars-and-cents ( u -- )
8861: 0 \ convert to unsigned double
8862: <<# \ start conversion
8863: # # \ convert two least-significant digits
8864: [char] . hold \ insert decimal point
8865: #s \ convert remaining digits
8866: [char] $ hold \ append currency symbol
8867: #> \ complete conversion
8868: TYPE SPACE \ display, with trailing space
8869: #>> ; \ release hold area
8870:
8871: : my-. ( n -- )
8872: \ handling negatives.. behaves like Standard .
8873: s>d \ convert to signed double
8874: swap over dabs \ leave sign byte followed by unsigned double
8875: <<# \ start conversion
8876: #s \ convert all digits
8877: rot sign \ get at sign byte, append "-" if needed
8878: #> \ complete conversion
8879: TYPE SPACE \ display, with trailing space
8880: #>> ; \ release hold area
8881:
8882: : account. ( n -- )
8883: \ accountants don't like minus signs, they use parentheses
8884: \ for negative numbers
8885: s>d \ convert to signed double
8886: swap over dabs \ leave sign byte followed by unsigned double
8887: <<# \ start conversion
8888: 2 pick \ get copy of sign byte
8889: 0< IF [char] ) hold THEN \ right-most character of output
8890: #s \ convert all digits
8891: rot \ get at sign byte
8892: 0< IF [char] ( hold THEN
8893: #> \ complete conversion
8894: TYPE SPACE \ display, with trailing space
8895: #>> ; \ release hold area
8896:
8897: @end example
8898:
8899: Here are some examples of using these words:
8900:
8901: @example
8902: 1 my-u. 1
8903: hex -1 my-u. decimal FFFFFFFF
8904: 1 cents-only 01
8905: 1234 cents-only 34
8906: 2 dollars-and-cents $0.02
8907: 1234 dollars-and-cents $12.34
8908: 123 my-. 123
8909: -123 my. -123
8910: 123 account. 123
8911: -456 account. (456)
8912: @end example
8913:
8914:
8915: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8916: @subsection String Formats
8917: @cindex strings - see character strings
8918: @cindex character strings - formats
8919: @cindex I/O - see character strings
8920: @cindex counted strings
8921:
8922: @c anton: this does not really belong here; maybe the memory section,
8923: @c or the principles chapter
8924:
8925: Forth commonly uses two different methods for representing character
8926: strings:
8927:
8928: @itemize @bullet
8929: @item
8930: @cindex address of counted string
8931: @cindex counted string
8932: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8933: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8934: string and the string occupies the subsequent @i{n} char addresses in
8935: memory.
8936: @item
8937: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8938: of the string in characters, and @i{c-addr} is the address of the
8939: first byte of the string.
8940: @end itemize
8941:
8942: ANS Forth encourages the use of the second format when representing
8943: strings.
8944:
8945:
8946: doc-count
8947:
8948:
8949: For words that move, copy and search for strings see @ref{Memory
8950: Blocks}. For words that display characters and strings see
8951: @ref{Displaying characters and strings}.
8952:
8953: @node Displaying characters and strings, Input, String Formats, Other I/O
8954: @subsection Displaying characters and strings
8955: @cindex characters - compiling and displaying
8956: @cindex character strings - compiling and displaying
8957:
8958: This section starts with a glossary of Forth words and ends with a set
8959: of examples.
8960:
8961:
8962: doc-bl
8963: doc-space
8964: doc-spaces
8965: doc-emit
8966: doc-toupper
8967: doc-."
8968: doc-.(
8969: doc-.\"
8970: doc-type
8971: doc-typewhite
8972: doc-cr
8973: @cindex cursor control
8974: doc-at-xy
8975: doc-page
8976: doc-s"
8977: doc-s\"
8978: doc-c"
8979: doc-char
8980: doc-[char]
8981:
8982:
8983: @noindent
8984: As an example, consider the following text, stored in a file @file{test.fs}:
8985:
8986: @example
8987: .( text-1)
8988: : my-word
8989: ." text-2" cr
8990: .( text-3)
8991: ;
8992:
8993: ." text-4"
8994:
8995: : my-char
8996: [char] ALPHABET emit
8997: char emit
8998: ;
8999: @end example
9000:
9001: When you load this code into Gforth, the following output is generated:
9002:
9003: @example
9004: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
9005: @end example
9006:
9007: @itemize @bullet
9008: @item
9009: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
9010: is an immediate word; it behaves in the same way whether it is used inside
9011: or outside a colon definition.
9012: @item
9013: Message @code{text-4} is displayed because of Gforth's added interpretation
9014: semantics for @code{."}.
9015: @item
9016: Message @code{text-2} is @i{not} displayed, because the text interpreter
9017: performs the compilation semantics for @code{."} within the definition of
9018: @code{my-word}.
9019: @end itemize
9020:
9021: Here are some examples of executing @code{my-word} and @code{my-char}:
9022:
9023: @example
9024: @kbd{my-word @key{RET}} text-2
9025: ok
9026: @kbd{my-char fred @key{RET}} Af ok
9027: @kbd{my-char jim @key{RET}} Aj ok
9028: @end example
9029:
9030: @itemize @bullet
9031: @item
9032: Message @code{text-2} is displayed because of the run-time behaviour of
9033: @code{."}.
9034: @item
9035: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
9036: on the stack at run-time. @code{emit} always displays the character
9037: when @code{my-char} is executed.
9038: @item
9039: @code{char} parses a string at run-time and the second @code{emit} displays
9040: the first character of the string.
9041: @item
9042: If you type @code{see my-char} you can see that @code{[char]} discarded
9043: the text ``LPHABET'' and only compiled the display code for ``A'' into the
9044: definition of @code{my-char}.
9045: @end itemize
9046:
9047:
9048:
9049: @node Input, , Displaying characters and strings, Other I/O
9050: @subsection Input
9051: @cindex input
9052: @cindex I/O - see input
9053: @cindex parsing a string
9054:
9055: For ways of storing character strings in memory see @ref{String Formats}.
9056:
9057: @comment TODO examples for >number >float accept key key? pad parse word refill
9058: @comment then index them
9059:
9060:
9061: doc-key
9062: doc-key?
9063: doc-ekey
9064: doc-ekey?
9065: doc-ekey>char
9066: doc->number
9067: doc->float
9068: doc-accept
9069: doc-pad
9070: @c anton: these belong in the input stream section
9071: doc-parse
9072: doc-word
9073: doc-name
9074: doc-parse-word
9075: doc-\"-parse
9076: doc-sword
9077: doc-refill
9078: @comment obsolescent words..
9079: doc-convert
9080: doc-query
9081: doc-expect
9082: doc-span
9083:
9084:
9085: @c -------------------------------------------------------------
9086: @node Locals, Structures, Other I/O, Words
9087: @section Locals
9088: @cindex locals
9089:
9090: Local variables can make Forth programming more enjoyable and Forth
9091: programs easier to read. Unfortunately, the locals of ANS Forth are
9092: laden with restrictions. Therefore, we provide not only the ANS Forth
9093: locals wordset, but also our own, more powerful locals wordset (we
9094: implemented the ANS Forth locals wordset through our locals wordset).
9095:
9096: The ideas in this section have also been published in M. Anton Ertl,
9097: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
9098: Automatic Scoping of Local Variables}}, EuroForth '94.
9099:
9100: @menu
9101: * Gforth locals::
9102: * ANS Forth locals::
9103: @end menu
9104:
9105: @node Gforth locals, ANS Forth locals, Locals, Locals
9106: @subsection Gforth locals
9107: @cindex Gforth locals
9108: @cindex locals, Gforth style
9109:
9110: Locals can be defined with
9111:
9112: @example
9113: @{ local1 local2 ... -- comment @}
9114: @end example
9115: or
9116: @example
9117: @{ local1 local2 ... @}
9118: @end example
9119:
9120: E.g.,
9121: @example
9122: : max @{ n1 n2 -- n3 @}
9123: n1 n2 > if
9124: n1
9125: else
9126: n2
9127: endif ;
9128: @end example
9129:
9130: The similarity of locals definitions with stack comments is intended. A
9131: locals definition often replaces the stack comment of a word. The order
9132: of the locals corresponds to the order in a stack comment and everything
9133: after the @code{--} is really a comment.
9134:
9135: This similarity has one disadvantage: It is too easy to confuse locals
9136: declarations with stack comments, causing bugs and making them hard to
9137: find. However, this problem can be avoided by appropriate coding
9138: conventions: Do not use both notations in the same program. If you do,
9139: they should be distinguished using additional means, e.g. by position.
9140:
9141: @cindex types of locals
9142: @cindex locals types
9143: The name of the local may be preceded by a type specifier, e.g.,
9144: @code{F:} for a floating point value:
9145:
9146: @example
9147: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
9148: \ complex multiplication
9149: Ar Br f* Ai Bi f* f-
9150: Ar Bi f* Ai Br f* f+ ;
9151: @end example
9152:
9153: @cindex flavours of locals
9154: @cindex locals flavours
9155: @cindex value-flavoured locals
9156: @cindex variable-flavoured locals
9157: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9158: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9159: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9160: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9161: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9162: produces its address (which becomes invalid when the variable's scope is
9163: left). E.g., the standard word @code{emit} can be defined in terms of
9164: @code{type} like this:
9165:
9166: @example
9167: : emit @{ C^ char* -- @}
9168: char* 1 type ;
9169: @end example
9170:
9171: @cindex default type of locals
9172: @cindex locals, default type
9173: A local without type specifier is a @code{W:} local. Both flavours of
9174: locals are initialized with values from the data or FP stack.
9175:
9176: Currently there is no way to define locals with user-defined data
9177: structures, but we are working on it.
9178:
9179: Gforth allows defining locals everywhere in a colon definition. This
9180: poses the following questions:
9181:
9182: @menu
9183: * Where are locals visible by name?::
9184: * How long do locals live?::
9185: * Locals programming style::
9186: * Locals implementation::
9187: @end menu
9188:
9189: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9190: @subsubsection Where are locals visible by name?
9191: @cindex locals visibility
9192: @cindex visibility of locals
9193: @cindex scope of locals
9194:
9195: Basically, the answer is that locals are visible where you would expect
9196: it in block-structured languages, and sometimes a little longer. If you
9197: want to restrict the scope of a local, enclose its definition in
9198: @code{SCOPE}...@code{ENDSCOPE}.
9199:
9200:
9201: doc-scope
9202: doc-endscope
9203:
9204:
9205: These words behave like control structure words, so you can use them
9206: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9207: arbitrary ways.
9208:
9209: If you want a more exact answer to the visibility question, here's the
9210: basic principle: A local is visible in all places that can only be
9211: reached through the definition of the local@footnote{In compiler
9212: construction terminology, all places dominated by the definition of the
9213: local.}. In other words, it is not visible in places that can be reached
9214: without going through the definition of the local. E.g., locals defined
9215: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9216: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9217: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
9218:
9219: The reasoning behind this solution is: We want to have the locals
9220: visible as long as it is meaningful. The user can always make the
9221: visibility shorter by using explicit scoping. In a place that can
9222: only be reached through the definition of a local, the meaning of a
9223: local name is clear. In other places it is not: How is the local
9224: initialized at the control flow path that does not contain the
9225: definition? Which local is meant, if the same name is defined twice in
9226: two independent control flow paths?
9227:
9228: This should be enough detail for nearly all users, so you can skip the
9229: rest of this section. If you really must know all the gory details and
9230: options, read on.
9231:
9232: In order to implement this rule, the compiler has to know which places
9233: are unreachable. It knows this automatically after @code{AHEAD},
9234: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9235: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9236: compiler that the control flow never reaches that place. If
9237: @code{UNREACHABLE} is not used where it could, the only consequence is
9238: that the visibility of some locals is more limited than the rule above
9239: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9240: lie to the compiler), buggy code will be produced.
9241:
9242:
9243: doc-unreachable
9244:
9245:
9246: Another problem with this rule is that at @code{BEGIN}, the compiler
9247: does not know which locals will be visible on the incoming
9248: back-edge. All problems discussed in the following are due to this
9249: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9250: loops as examples; the discussion also applies to @code{?DO} and other
9251: loops). Perhaps the most insidious example is:
9252: @example
9253: AHEAD
9254: BEGIN
9255: x
9256: [ 1 CS-ROLL ] THEN
9257: @{ x @}
9258: ...
9259: UNTIL
9260: @end example
9261:
9262: This should be legal according to the visibility rule. The use of
9263: @code{x} can only be reached through the definition; but that appears
9264: textually below the use.
9265:
9266: From this example it is clear that the visibility rules cannot be fully
9267: implemented without major headaches. Our implementation treats common
9268: cases as advertised and the exceptions are treated in a safe way: The
9269: compiler makes a reasonable guess about the locals visible after a
9270: @code{BEGIN}; if it is too pessimistic, the
9271: user will get a spurious error about the local not being defined; if the
9272: compiler is too optimistic, it will notice this later and issue a
9273: warning. In the case above the compiler would complain about @code{x}
9274: being undefined at its use. You can see from the obscure examples in
9275: this section that it takes quite unusual control structures to get the
9276: compiler into trouble, and even then it will often do fine.
9277:
9278: If the @code{BEGIN} is reachable from above, the most optimistic guess
9279: is that all locals visible before the @code{BEGIN} will also be
9280: visible after the @code{BEGIN}. This guess is valid for all loops that
9281: are entered only through the @code{BEGIN}, in particular, for normal
9282: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9283: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9284: compiler. When the branch to the @code{BEGIN} is finally generated by
9285: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9286: warns the user if it was too optimistic:
9287: @example
9288: IF
9289: @{ x @}
9290: BEGIN
9291: \ x ?
9292: [ 1 cs-roll ] THEN
9293: ...
9294: UNTIL
9295: @end example
9296:
9297: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9298: optimistically assumes that it lives until the @code{THEN}. It notices
9299: this difference when it compiles the @code{UNTIL} and issues a
9300: warning. The user can avoid the warning, and make sure that @code{x}
9301: is not used in the wrong area by using explicit scoping:
9302: @example
9303: IF
9304: SCOPE
9305: @{ x @}
9306: ENDSCOPE
9307: BEGIN
9308: [ 1 cs-roll ] THEN
9309: ...
9310: UNTIL
9311: @end example
9312:
9313: Since the guess is optimistic, there will be no spurious error messages
9314: about undefined locals.
9315:
9316: If the @code{BEGIN} is not reachable from above (e.g., after
9317: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9318: optimistic guess, as the locals visible after the @code{BEGIN} may be
9319: defined later. Therefore, the compiler assumes that no locals are
9320: visible after the @code{BEGIN}. However, the user can use
9321: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9322: visible at the BEGIN as at the point where the top control-flow stack
9323: item was created.
9324:
9325:
9326: doc-assume-live
9327:
9328:
9329: @noindent
9330: E.g.,
9331: @example
9332: @{ x @}
9333: AHEAD
9334: ASSUME-LIVE
9335: BEGIN
9336: x
9337: [ 1 CS-ROLL ] THEN
9338: ...
9339: UNTIL
9340: @end example
9341:
9342: Other cases where the locals are defined before the @code{BEGIN} can be
9343: handled by inserting an appropriate @code{CS-ROLL} before the
9344: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9345: behind the @code{ASSUME-LIVE}).
9346:
9347: Cases where locals are defined after the @code{BEGIN} (but should be
9348: visible immediately after the @code{BEGIN}) can only be handled by
9349: rearranging the loop. E.g., the ``most insidious'' example above can be
9350: arranged into:
9351: @example
9352: BEGIN
9353: @{ x @}
9354: ... 0=
9355: WHILE
9356: x
9357: REPEAT
9358: @end example
9359:
9360: @node How long do locals live?, Locals programming style, Where are locals visible by name?, Gforth locals
9361: @subsubsection How long do locals live?
9362: @cindex locals lifetime
9363: @cindex lifetime of locals
9364:
9365: The right answer for the lifetime question would be: A local lives at
9366: least as long as it can be accessed. For a value-flavoured local this
9367: means: until the end of its visibility. However, a variable-flavoured
9368: local could be accessed through its address far beyond its visibility
9369: scope. Ultimately, this would mean that such locals would have to be
9370: garbage collected. Since this entails un-Forth-like implementation
9371: complexities, I adopted the same cowardly solution as some other
9372: languages (e.g., C): The local lives only as long as it is visible;
9373: afterwards its address is invalid (and programs that access it
9374: afterwards are erroneous).
9375:
9376: @node Locals programming style, Locals implementation, How long do locals live?, Gforth locals
9377: @subsubsection Locals programming style
9378: @cindex locals programming style
9379: @cindex programming style, locals
9380:
9381: The freedom to define locals anywhere has the potential to change
9382: programming styles dramatically. In particular, the need to use the
9383: return stack for intermediate storage vanishes. Moreover, all stack
9384: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9385: determined arguments) can be eliminated: If the stack items are in the
9386: wrong order, just write a locals definition for all of them; then
9387: write the items in the order you want.
9388:
9389: This seems a little far-fetched and eliminating stack manipulations is
9390: unlikely to become a conscious programming objective. Still, the number
9391: of stack manipulations will be reduced dramatically if local variables
9392: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
9393: a traditional implementation of @code{max}).
9394:
9395: This shows one potential benefit of locals: making Forth programs more
9396: readable. Of course, this benefit will only be realized if the
9397: programmers continue to honour the principle of factoring instead of
9398: using the added latitude to make the words longer.
9399:
9400: @cindex single-assignment style for locals
9401: Using @code{TO} can and should be avoided. Without @code{TO},
9402: every value-flavoured local has only a single assignment and many
9403: advantages of functional languages apply to Forth. I.e., programs are
9404: easier to analyse, to optimize and to read: It is clear from the
9405: definition what the local stands for, it does not turn into something
9406: different later.
9407:
9408: E.g., a definition using @code{TO} might look like this:
9409: @example
9410: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9411: u1 u2 min 0
9412: ?do
9413: addr1 c@@ addr2 c@@ -
9414: ?dup-if
9415: unloop exit
9416: then
9417: addr1 char+ TO addr1
9418: addr2 char+ TO addr2
9419: loop
9420: u1 u2 - ;
9421: @end example
9422: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9423: every loop iteration. @code{strcmp} is a typical example of the
9424: readability problems of using @code{TO}. When you start reading
9425: @code{strcmp}, you think that @code{addr1} refers to the start of the
9426: string. Only near the end of the loop you realize that it is something
9427: else.
9428:
9429: This can be avoided by defining two locals at the start of the loop that
9430: are initialized with the right value for the current iteration.
9431: @example
9432: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9433: addr1 addr2
9434: u1 u2 min 0
9435: ?do @{ s1 s2 @}
9436: s1 c@@ s2 c@@ -
9437: ?dup-if
9438: unloop exit
9439: then
9440: s1 char+ s2 char+
9441: loop
9442: 2drop
9443: u1 u2 - ;
9444: @end example
9445: Here it is clear from the start that @code{s1} has a different value
9446: in every loop iteration.
9447:
9448: @node Locals implementation, , Locals programming style, Gforth locals
9449: @subsubsection Locals implementation
9450: @cindex locals implementation
9451: @cindex implementation of locals
9452:
9453: @cindex locals stack
9454: Gforth uses an extra locals stack. The most compelling reason for
9455: this is that the return stack is not float-aligned; using an extra stack
9456: also eliminates the problems and restrictions of using the return stack
9457: as locals stack. Like the other stacks, the locals stack grows toward
9458: lower addresses. A few primitives allow an efficient implementation:
9459:
9460:
9461: doc-@local#
9462: doc-f@local#
9463: doc-laddr#
9464: doc-lp+!#
9465: doc-lp!
9466: doc->l
9467: doc-f>l
9468:
9469:
9470: In addition to these primitives, some specializations of these
9471: primitives for commonly occurring inline arguments are provided for
9472: efficiency reasons, e.g., @code{@@local0} as specialization of
9473: @code{@@local#} for the inline argument 0. The following compiling words
9474: compile the right specialized version, or the general version, as
9475: appropriate:
9476:
9477:
9478: @c doc-compile-@local
9479: @c doc-compile-f@local
9480: doc-compile-lp+!
9481:
9482:
9483: Combinations of conditional branches and @code{lp+!#} like
9484: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9485: is taken) are provided for efficiency and correctness in loops.
9486:
9487: A special area in the dictionary space is reserved for keeping the
9488: local variable names. @code{@{} switches the dictionary pointer to this
9489: area and @code{@}} switches it back and generates the locals
9490: initializing code. @code{W:} etc.@ are normal defining words. This
9491: special area is cleared at the start of every colon definition.
9492:
9493: @cindex word list for defining locals
9494: A special feature of Gforth's dictionary is used to implement the
9495: definition of locals without type specifiers: every word list (aka
9496: vocabulary) has its own methods for searching
9497: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9498: with a special search method: When it is searched for a word, it
9499: actually creates that word using @code{W:}. @code{@{} changes the search
9500: order to first search the word list containing @code{@}}, @code{W:} etc.,
9501: and then the word list for defining locals without type specifiers.
9502:
9503: The lifetime rules support a stack discipline within a colon
9504: definition: The lifetime of a local is either nested with other locals
9505: lifetimes or it does not overlap them.
9506:
9507: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9508: pointer manipulation is generated. Between control structure words
9509: locals definitions can push locals onto the locals stack. @code{AGAIN}
9510: is the simplest of the other three control flow words. It has to
9511: restore the locals stack depth of the corresponding @code{BEGIN}
9512: before branching. The code looks like this:
9513: @format
9514: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9515: @code{branch} <begin>
9516: @end format
9517:
9518: @code{UNTIL} is a little more complicated: If it branches back, it
9519: must adjust the stack just like @code{AGAIN}. But if it falls through,
9520: the locals stack must not be changed. The compiler generates the
9521: following code:
9522: @format
9523: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9524: @end format
9525: The locals stack pointer is only adjusted if the branch is taken.
9526:
9527: @code{THEN} can produce somewhat inefficient code:
9528: @format
9529: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9530: <orig target>:
9531: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9532: @end format
9533: The second @code{lp+!#} adjusts the locals stack pointer from the
9534: level at the @i{orig} point to the level after the @code{THEN}. The
9535: first @code{lp+!#} adjusts the locals stack pointer from the current
9536: level to the level at the orig point, so the complete effect is an
9537: adjustment from the current level to the right level after the
9538: @code{THEN}.
9539:
9540: @cindex locals information on the control-flow stack
9541: @cindex control-flow stack items, locals information
9542: In a conventional Forth implementation a dest control-flow stack entry
9543: is just the target address and an orig entry is just the address to be
9544: patched. Our locals implementation adds a word list to every orig or dest
9545: item. It is the list of locals visible (or assumed visible) at the point
9546: described by the entry. Our implementation also adds a tag to identify
9547: the kind of entry, in particular to differentiate between live and dead
9548: (reachable and unreachable) orig entries.
9549:
9550: A few unusual operations have to be performed on locals word lists:
9551:
9552:
9553: doc-common-list
9554: doc-sub-list?
9555: doc-list-size
9556:
9557:
9558: Several features of our locals word list implementation make these
9559: operations easy to implement: The locals word lists are organised as
9560: linked lists; the tails of these lists are shared, if the lists
9561: contain some of the same locals; and the address of a name is greater
9562: than the address of the names behind it in the list.
9563:
9564: Another important implementation detail is the variable
9565: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9566: determine if they can be reached directly or only through the branch
9567: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9568: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9569: definition, by @code{BEGIN} and usually by @code{THEN}.
9570:
9571: Counted loops are similar to other loops in most respects, but
9572: @code{LEAVE} requires special attention: It performs basically the same
9573: service as @code{AHEAD}, but it does not create a control-flow stack
9574: entry. Therefore the information has to be stored elsewhere;
9575: traditionally, the information was stored in the target fields of the
9576: branches created by the @code{LEAVE}s, by organizing these fields into a
9577: linked list. Unfortunately, this clever trick does not provide enough
9578: space for storing our extended control flow information. Therefore, we
9579: introduce another stack, the leave stack. It contains the control-flow
9580: stack entries for all unresolved @code{LEAVE}s.
9581:
9582: Local names are kept until the end of the colon definition, even if
9583: they are no longer visible in any control-flow path. In a few cases
9584: this may lead to increased space needs for the locals name area, but
9585: usually less than reclaiming this space would cost in code size.
9586:
9587:
9588: @node ANS Forth locals, , Gforth locals, Locals
9589: @subsection ANS Forth locals
9590: @cindex locals, ANS Forth style
9591:
9592: The ANS Forth locals wordset does not define a syntax for locals, but
9593: words that make it possible to define various syntaxes. One of the
9594: possible syntaxes is a subset of the syntax we used in the Gforth locals
9595: wordset, i.e.:
9596:
9597: @example
9598: @{ local1 local2 ... -- comment @}
9599: @end example
9600: @noindent
9601: or
9602: @example
9603: @{ local1 local2 ... @}
9604: @end example
9605:
9606: The order of the locals corresponds to the order in a stack comment. The
9607: restrictions are:
9608:
9609: @itemize @bullet
9610: @item
9611: Locals can only be cell-sized values (no type specifiers are allowed).
9612: @item
9613: Locals can be defined only outside control structures.
9614: @item
9615: Locals can interfere with explicit usage of the return stack. For the
9616: exact (and long) rules, see the standard. If you don't use return stack
9617: accessing words in a definition using locals, you will be all right. The
9618: purpose of this rule is to make locals implementation on the return
9619: stack easier.
9620: @item
9621: The whole definition must be in one line.
9622: @end itemize
9623:
9624: Locals defined in ANS Forth behave like @code{VALUE}s
9625: (@pxref{Values}). I.e., they are initialized from the stack. Using their
9626: name produces their value. Their value can be changed using @code{TO}.
9627:
9628: Since the syntax above is supported by Gforth directly, you need not do
9629: anything to use it. If you want to port a program using this syntax to
9630: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9631: syntax on the other system.
9632:
9633: Note that a syntax shown in the standard, section A.13 looks
9634: similar, but is quite different in having the order of locals
9635: reversed. Beware!
9636:
9637: The ANS Forth locals wordset itself consists of one word:
9638:
9639: doc-(local)
9640:
9641: The ANS Forth locals extension wordset defines a syntax using
9642: @code{locals|}, but it is so awful that we strongly recommend not to use
9643: it. We have implemented this syntax to make porting to Gforth easy, but
9644: do not document it here. The problem with this syntax is that the locals
9645: are defined in an order reversed with respect to the standard stack
9646: comment notation, making programs harder to read, and easier to misread
9647: and miswrite. The only merit of this syntax is that it is easy to
9648: implement using the ANS Forth locals wordset.
9649:
9650:
9651: @c ----------------------------------------------------------
9652: @node Structures, Object-oriented Forth, Locals, Words
9653: @section Structures
9654: @cindex structures
9655: @cindex records
9656:
9657: This section presents the structure package that comes with Gforth. A
9658: version of the package implemented in ANS Forth is available in
9659: @file{compat/struct.fs}. This package was inspired by a posting on
9660: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9661: possibly John Hayes). A version of this section has been published in
9662: M. Anton Ertl,
9663: @uref{http://www.complang.tuwien.ac.at/forth/objects/structs.html, Yet
9664: Another Forth Structures Package}, Forth Dimensions 19(3), pages
9665: 13--16. Marcel Hendrix provided helpful comments.
9666:
9667: @menu
9668: * Why explicit structure support?::
9669: * Structure Usage::
9670: * Structure Naming Convention::
9671: * Structure Implementation::
9672: * Structure Glossary::
9673: @end menu
9674:
9675: @node Why explicit structure support?, Structure Usage, Structures, Structures
9676: @subsection Why explicit structure support?
9677:
9678: @cindex address arithmetic for structures
9679: @cindex structures using address arithmetic
9680: If we want to use a structure containing several fields, we could simply
9681: reserve memory for it, and access the fields using address arithmetic
9682: (@pxref{Address arithmetic}). As an example, consider a structure with
9683: the following fields
9684:
9685: @table @code
9686: @item a
9687: is a float
9688: @item b
9689: is a cell
9690: @item c
9691: is a float
9692: @end table
9693:
9694: Given the (float-aligned) base address of the structure we get the
9695: address of the field
9696:
9697: @table @code
9698: @item a
9699: without doing anything further.
9700: @item b
9701: with @code{float+}
9702: @item c
9703: with @code{float+ cell+ faligned}
9704: @end table
9705:
9706: It is easy to see that this can become quite tiring.
9707:
9708: Moreover, it is not very readable, because seeing a
9709: @code{cell+} tells us neither which kind of structure is
9710: accessed nor what field is accessed; we have to somehow infer the kind
9711: of structure, and then look up in the documentation, which field of
9712: that structure corresponds to that offset.
9713:
9714: Finally, this kind of address arithmetic also causes maintenance
9715: troubles: If you add or delete a field somewhere in the middle of the
9716: structure, you have to find and change all computations for the fields
9717: afterwards.
9718:
9719: So, instead of using @code{cell+} and friends directly, how
9720: about storing the offsets in constants:
9721:
9722: @example
9723: 0 constant a-offset
9724: 0 float+ constant b-offset
9725: 0 float+ cell+ faligned c-offset
9726: @end example
9727:
9728: Now we can get the address of field @code{x} with @code{x-offset
9729: +}. This is much better in all respects. Of course, you still
9730: have to change all later offset definitions if you add a field. You can
9731: fix this by declaring the offsets in the following way:
9732:
9733: @example
9734: 0 constant a-offset
9735: a-offset float+ constant b-offset
9736: b-offset cell+ faligned constant c-offset
9737: @end example
9738:
9739: Since we always use the offsets with @code{+}, we could use a defining
9740: word @code{cfield} that includes the @code{+} in the action of the
9741: defined word:
9742:
9743: @example
9744: : cfield ( n "name" -- )
9745: create ,
9746: does> ( name execution: addr1 -- addr2 )
9747: @@ + ;
9748:
9749: 0 cfield a
9750: 0 a float+ cfield b
9751: 0 b cell+ faligned cfield c
9752: @end example
9753:
9754: Instead of @code{x-offset +}, we now simply write @code{x}.
9755:
9756: The structure field words now can be used quite nicely. However,
9757: their definition is still a bit cumbersome: We have to repeat the
9758: name, the information about size and alignment is distributed before
9759: and after the field definitions etc. The structure package presented
9760: here addresses these problems.
9761:
9762: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9763: @subsection Structure Usage
9764: @cindex structure usage
9765:
9766: @cindex @code{field} usage
9767: @cindex @code{struct} usage
9768: @cindex @code{end-struct} usage
9769: You can define a structure for a (data-less) linked list with:
9770: @example
9771: struct
9772: cell% field list-next
9773: end-struct list%
9774: @end example
9775:
9776: With the address of the list node on the stack, you can compute the
9777: address of the field that contains the address of the next node with
9778: @code{list-next}. E.g., you can determine the length of a list
9779: with:
9780:
9781: @example
9782: : list-length ( list -- n )
9783: \ "list" is a pointer to the first element of a linked list
9784: \ "n" is the length of the list
9785: 0 BEGIN ( list1 n1 )
9786: over
9787: WHILE ( list1 n1 )
9788: 1+ swap list-next @@ swap
9789: REPEAT
9790: nip ;
9791: @end example
9792:
9793: You can reserve memory for a list node in the dictionary with
9794: @code{list% %allot}, which leaves the address of the list node on the
9795: stack. For the equivalent allocation on the heap you can use @code{list%
9796: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9797: use @code{list% %allocate}). You can get the the size of a list
9798: node with @code{list% %size} and its alignment with @code{list%
9799: %alignment}.
9800:
9801: Note that in ANS Forth the body of a @code{create}d word is
9802: @code{aligned} but not necessarily @code{faligned};
9803: therefore, if you do a:
9804:
9805: @example
9806: create @emph{name} foo% %allot drop
9807: @end example
9808:
9809: @noindent
9810: then the memory alloted for @code{foo%} is guaranteed to start at the
9811: body of @code{@emph{name}} only if @code{foo%} contains only character,
9812: cell and double fields. Therefore, if your structure contains floats,
9813: better use
9814:
9815: @example
9816: foo% %allot constant @emph{name}
9817: @end example
9818:
9819: @cindex structures containing structures
9820: You can include a structure @code{foo%} as a field of
9821: another structure, like this:
9822: @example
9823: struct
9824: ...
9825: foo% field ...
9826: ...
9827: end-struct ...
9828: @end example
9829:
9830: @cindex structure extension
9831: @cindex extended records
9832: Instead of starting with an empty structure, you can extend an
9833: existing structure. E.g., a plain linked list without data, as defined
9834: above, is hardly useful; You can extend it to a linked list of integers,
9835: like this:@footnote{This feature is also known as @emph{extended
9836: records}. It is the main innovation in the Oberon language; in other
9837: words, adding this feature to Modula-2 led Wirth to create a new
9838: language, write a new compiler etc. Adding this feature to Forth just
9839: required a few lines of code.}
9840:
9841: @example
9842: list%
9843: cell% field intlist-int
9844: end-struct intlist%
9845: @end example
9846:
9847: @code{intlist%} is a structure with two fields:
9848: @code{list-next} and @code{intlist-int}.
9849:
9850: @cindex structures containing arrays
9851: You can specify an array type containing @emph{n} elements of
9852: type @code{foo%} like this:
9853:
9854: @example
9855: foo% @emph{n} *
9856: @end example
9857:
9858: You can use this array type in any place where you can use a normal
9859: type, e.g., when defining a @code{field}, or with
9860: @code{%allot}.
9861:
9862: @cindex first field optimization
9863: The first field is at the base address of a structure and the word for
9864: this field (e.g., @code{list-next}) actually does not change the address
9865: on the stack. You may be tempted to leave it away in the interest of
9866: run-time and space efficiency. This is not necessary, because the
9867: structure package optimizes this case: If you compile a first-field
9868: words, no code is generated. So, in the interest of readability and
9869: maintainability you should include the word for the field when accessing
9870: the field.
9871:
9872:
9873: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
9874: @subsection Structure Naming Convention
9875: @cindex structure naming convention
9876:
9877: The field names that come to (my) mind are often quite generic, and,
9878: if used, would cause frequent name clashes. E.g., many structures
9879: probably contain a @code{counter} field. The structure names
9880: that come to (my) mind are often also the logical choice for the names
9881: of words that create such a structure.
9882:
9883: Therefore, I have adopted the following naming conventions:
9884:
9885: @itemize @bullet
9886: @cindex field naming convention
9887: @item
9888: The names of fields are of the form
9889: @code{@emph{struct}-@emph{field}}, where
9890: @code{@emph{struct}} is the basic name of the structure, and
9891: @code{@emph{field}} is the basic name of the field. You can
9892: think of field words as converting the (address of the)
9893: structure into the (address of the) field.
9894:
9895: @cindex structure naming convention
9896: @item
9897: The names of structures are of the form
9898: @code{@emph{struct}%}, where
9899: @code{@emph{struct}} is the basic name of the structure.
9900: @end itemize
9901:
9902: This naming convention does not work that well for fields of extended
9903: structures; e.g., the integer list structure has a field
9904: @code{intlist-int}, but has @code{list-next}, not
9905: @code{intlist-next}.
9906:
9907: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
9908: @subsection Structure Implementation
9909: @cindex structure implementation
9910: @cindex implementation of structures
9911:
9912: The central idea in the implementation is to pass the data about the
9913: structure being built on the stack, not in some global
9914: variable. Everything else falls into place naturally once this design
9915: decision is made.
9916:
9917: The type description on the stack is of the form @emph{align
9918: size}. Keeping the size on the top-of-stack makes dealing with arrays
9919: very simple.
9920:
9921: @code{field} is a defining word that uses @code{Create}
9922: and @code{DOES>}. The body of the field contains the offset
9923: of the field, and the normal @code{DOES>} action is simply:
9924:
9925: @example
9926: @@ +
9927: @end example
9928:
9929: @noindent
9930: i.e., add the offset to the address, giving the stack effect
9931: @i{addr1 -- addr2} for a field.
9932:
9933: @cindex first field optimization, implementation
9934: This simple structure is slightly complicated by the optimization
9935: for fields with offset 0, which requires a different
9936: @code{DOES>}-part (because we cannot rely on there being
9937: something on the stack if such a field is invoked during
9938: compilation). Therefore, we put the different @code{DOES>}-parts
9939: in separate words, and decide which one to invoke based on the
9940: offset. For a zero offset, the field is basically a noop; it is
9941: immediate, and therefore no code is generated when it is compiled.
9942:
9943: @node Structure Glossary, , Structure Implementation, Structures
9944: @subsection Structure Glossary
9945: @cindex structure glossary
9946:
9947:
9948: doc-%align
9949: doc-%alignment
9950: doc-%alloc
9951: doc-%allocate
9952: doc-%allot
9953: doc-cell%
9954: doc-char%
9955: doc-dfloat%
9956: doc-double%
9957: doc-end-struct
9958: doc-field
9959: doc-float%
9960: doc-naligned
9961: doc-sfloat%
9962: doc-%size
9963: doc-struct
9964:
9965:
9966: @c -------------------------------------------------------------
9967: @node Object-oriented Forth, Programming Tools, Structures, Words
9968: @section Object-oriented Forth
9969:
9970: Gforth comes with three packages for object-oriented programming:
9971: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
9972: is preloaded, so you have to @code{include} them before use. The most
9973: important differences between these packages (and others) are discussed
9974: in @ref{Comparison with other object models}. All packages are written
9975: in ANS Forth and can be used with any other ANS Forth.
9976:
9977: @menu
9978: * Why object-oriented programming?::
9979: * Object-Oriented Terminology::
9980: * Objects::
9981: * OOF::
9982: * Mini-OOF::
9983: * Comparison with other object models::
9984: @end menu
9985:
9986: @c ----------------------------------------------------------------
9987: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
9988: @subsection Why object-oriented programming?
9989: @cindex object-oriented programming motivation
9990: @cindex motivation for object-oriented programming
9991:
9992: Often we have to deal with several data structures (@emph{objects}),
9993: that have to be treated similarly in some respects, but differently in
9994: others. Graphical objects are the textbook example: circles, triangles,
9995: dinosaurs, icons, and others, and we may want to add more during program
9996: development. We want to apply some operations to any graphical object,
9997: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
9998: has to do something different for every kind of object.
9999: @comment TODO add some other operations eg perimeter, area
10000: @comment and tie in to concrete examples later..
10001:
10002: We could implement @code{draw} as a big @code{CASE}
10003: control structure that executes the appropriate code depending on the
10004: kind of object to be drawn. This would be not be very elegant, and,
10005: moreover, we would have to change @code{draw} every time we add
10006: a new kind of graphical object (say, a spaceship).
10007:
10008: What we would rather do is: When defining spaceships, we would tell
10009: the system: ``Here's how you @code{draw} a spaceship; you figure
10010: out the rest''.
10011:
10012: This is the problem that all systems solve that (rightfully) call
10013: themselves object-oriented; the object-oriented packages presented here
10014: solve this problem (and not much else).
10015: @comment TODO ?list properties of oo systems.. oo vs o-based?
10016:
10017: @c ------------------------------------------------------------------------
10018: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
10019: @subsection Object-Oriented Terminology
10020: @cindex object-oriented terminology
10021: @cindex terminology for object-oriented programming
10022:
10023: This section is mainly for reference, so you don't have to understand
10024: all of it right away. The terminology is mainly Smalltalk-inspired. In
10025: short:
10026:
10027: @table @emph
10028: @cindex class
10029: @item class
10030: a data structure definition with some extras.
10031:
10032: @cindex object
10033: @item object
10034: an instance of the data structure described by the class definition.
10035:
10036: @cindex instance variables
10037: @item instance variables
10038: fields of the data structure.
10039:
10040: @cindex selector
10041: @cindex method selector
10042: @cindex virtual function
10043: @item selector
10044: (or @emph{method selector}) a word (e.g.,
10045: @code{draw}) that performs an operation on a variety of data
10046: structures (classes). A selector describes @emph{what} operation to
10047: perform. In C++ terminology: a (pure) virtual function.
10048:
10049: @cindex method
10050: @item method
10051: the concrete definition that performs the operation
10052: described by the selector for a specific class. A method specifies
10053: @emph{how} the operation is performed for a specific class.
10054:
10055: @cindex selector invocation
10056: @cindex message send
10057: @cindex invoking a selector
10058: @item selector invocation
10059: a call of a selector. One argument of the call (the TOS (top-of-stack))
10060: is used for determining which method is used. In Smalltalk terminology:
10061: a message (consisting of the selector and the other arguments) is sent
10062: to the object.
10063:
10064: @cindex receiving object
10065: @item receiving object
10066: the object used for determining the method executed by a selector
10067: invocation. In the @file{objects.fs} model, it is the object that is on
10068: the TOS when the selector is invoked. (@emph{Receiving} comes from
10069: the Smalltalk @emph{message} terminology.)
10070:
10071: @cindex child class
10072: @cindex parent class
10073: @cindex inheritance
10074: @item child class
10075: a class that has (@emph{inherits}) all properties (instance variables,
10076: selectors, methods) from a @emph{parent class}. In Smalltalk
10077: terminology: The subclass inherits from the superclass. In C++
10078: terminology: The derived class inherits from the base class.
10079:
10080: @end table
10081:
10082: @c If you wonder about the message sending terminology, it comes from
10083: @c a time when each object had it's own task and objects communicated via
10084: @c message passing; eventually the Smalltalk developers realized that
10085: @c they can do most things through simple (indirect) calls. They kept the
10086: @c terminology.
10087:
10088: @c --------------------------------------------------------------
10089: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
10090: @subsection The @file{objects.fs} model
10091: @cindex objects
10092: @cindex object-oriented programming
10093:
10094: @cindex @file{objects.fs}
10095: @cindex @file{oof.fs}
10096:
10097: This section describes the @file{objects.fs} package. This material also
10098: has been published in M. Anton Ertl,
10099: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
10100: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
10101: 37--43.
10102: @c McKewan's and Zsoter's packages
10103:
10104: This section assumes that you have read @ref{Structures}.
10105:
10106: The techniques on which this model is based have been used to implement
10107: the parser generator, Gray, and have also been used in Gforth for
10108: implementing the various flavours of word lists (hashed or not,
10109: case-sensitive or not, special-purpose word lists for locals etc.).
10110:
10111:
10112: @menu
10113: * Properties of the Objects model::
10114: * Basic Objects Usage::
10115: * The Objects base class::
10116: * Creating objects::
10117: * Object-Oriented Programming Style::
10118: * Class Binding::
10119: * Method conveniences::
10120: * Classes and Scoping::
10121: * Dividing classes::
10122: * Object Interfaces::
10123: * Objects Implementation::
10124: * Objects Glossary::
10125: @end menu
10126:
10127: Marcel Hendrix provided helpful comments on this section.
10128:
10129: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
10130: @subsubsection Properties of the @file{objects.fs} model
10131: @cindex @file{objects.fs} properties
10132:
10133: @itemize @bullet
10134: @item
10135: It is straightforward to pass objects on the stack. Passing
10136: selectors on the stack is a little less convenient, but possible.
10137:
10138: @item
10139: Objects are just data structures in memory, and are referenced by their
10140: address. You can create words for objects with normal defining words
10141: like @code{constant}. Likewise, there is no difference between instance
10142: variables that contain objects and those that contain other data.
10143:
10144: @item
10145: Late binding is efficient and easy to use.
10146:
10147: @item
10148: It avoids parsing, and thus avoids problems with state-smartness
10149: and reduced extensibility; for convenience there are a few parsing
10150: words, but they have non-parsing counterparts. There are also a few
10151: defining words that parse. This is hard to avoid, because all standard
10152: defining words parse (except @code{:noname}); however, such
10153: words are not as bad as many other parsing words, because they are not
10154: state-smart.
10155:
10156: @item
10157: It does not try to incorporate everything. It does a few things and does
10158: them well (IMO). In particular, this model was not designed to support
10159: information hiding (although it has features that may help); you can use
10160: a separate package for achieving this.
10161:
10162: @item
10163: It is layered; you don't have to learn and use all features to use this
10164: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10165: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
10166: are optional and independent of each other.
10167:
10168: @item
10169: An implementation in ANS Forth is available.
10170:
10171: @end itemize
10172:
10173:
10174: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10175: @subsubsection Basic @file{objects.fs} Usage
10176: @cindex basic objects usage
10177: @cindex objects, basic usage
10178:
10179: You can define a class for graphical objects like this:
10180:
10181: @cindex @code{class} usage
10182: @cindex @code{end-class} usage
10183: @cindex @code{selector} usage
10184: @example
10185: object class \ "object" is the parent class
10186: selector draw ( x y graphical -- )
10187: end-class graphical
10188: @end example
10189:
10190: This code defines a class @code{graphical} with an
10191: operation @code{draw}. We can perform the operation
10192: @code{draw} on any @code{graphical} object, e.g.:
10193:
10194: @example
10195: 100 100 t-rex draw
10196: @end example
10197:
10198: @noindent
10199: where @code{t-rex} is a word (say, a constant) that produces a
10200: graphical object.
10201:
10202: @comment TODO add a 2nd operation eg perimeter.. and use for
10203: @comment a concrete example
10204:
10205: @cindex abstract class
10206: How do we create a graphical object? With the present definitions,
10207: we cannot create a useful graphical object. The class
10208: @code{graphical} describes graphical objects in general, but not
10209: any concrete graphical object type (C++ users would call it an
10210: @emph{abstract class}); e.g., there is no method for the selector
10211: @code{draw} in the class @code{graphical}.
10212:
10213: For concrete graphical objects, we define child classes of the
10214: class @code{graphical}, e.g.:
10215:
10216: @cindex @code{overrides} usage
10217: @cindex @code{field} usage in class definition
10218: @example
10219: graphical class \ "graphical" is the parent class
10220: cell% field circle-radius
10221:
10222: :noname ( x y circle -- )
10223: circle-radius @@ draw-circle ;
10224: overrides draw
10225:
10226: :noname ( n-radius circle -- )
10227: circle-radius ! ;
10228: overrides construct
10229:
10230: end-class circle
10231: @end example
10232:
10233: Here we define a class @code{circle} as a child of @code{graphical},
10234: with field @code{circle-radius} (which behaves just like a field
10235: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10236: for the selectors @code{draw} and @code{construct} (@code{construct} is
10237: defined in @code{object}, the parent class of @code{graphical}).
10238:
10239: Now we can create a circle on the heap (i.e.,
10240: @code{allocate}d memory) with:
10241:
10242: @cindex @code{heap-new} usage
10243: @example
10244: 50 circle heap-new constant my-circle
10245: @end example
10246:
10247: @noindent
10248: @code{heap-new} invokes @code{construct}, thus
10249: initializing the field @code{circle-radius} with 50. We can draw
10250: this new circle at (100,100) with:
10251:
10252: @example
10253: 100 100 my-circle draw
10254: @end example
10255:
10256: @cindex selector invocation, restrictions
10257: @cindex class definition, restrictions
10258: Note: You can only invoke a selector if the object on the TOS
10259: (the receiving object) belongs to the class where the selector was
10260: defined or one of its descendents; e.g., you can invoke
10261: @code{draw} only for objects belonging to @code{graphical}
10262: or its descendents (e.g., @code{circle}). Immediately before
10263: @code{end-class}, the search order has to be the same as
10264: immediately after @code{class}.
10265:
10266: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10267: @subsubsection The @file{object.fs} base class
10268: @cindex @code{object} class
10269:
10270: When you define a class, you have to specify a parent class. So how do
10271: you start defining classes? There is one class available from the start:
10272: @code{object}. It is ancestor for all classes and so is the
10273: only class that has no parent. It has two selectors: @code{construct}
10274: and @code{print}.
10275:
10276: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10277: @subsubsection Creating objects
10278: @cindex creating objects
10279: @cindex object creation
10280: @cindex object allocation options
10281:
10282: @cindex @code{heap-new} discussion
10283: @cindex @code{dict-new} discussion
10284: @cindex @code{construct} discussion
10285: You can create and initialize an object of a class on the heap with
10286: @code{heap-new} ( ... class -- object ) and in the dictionary
10287: (allocation with @code{allot}) with @code{dict-new} (
10288: ... class -- object ). Both words invoke @code{construct}, which
10289: consumes the stack items indicated by "..." above.
10290:
10291: @cindex @code{init-object} discussion
10292: @cindex @code{class-inst-size} discussion
10293: If you want to allocate memory for an object yourself, you can get its
10294: alignment and size with @code{class-inst-size 2@@} ( class --
10295: align size ). Once you have memory for an object, you can initialize
10296: it with @code{init-object} ( ... class object -- );
10297: @code{construct} does only a part of the necessary work.
10298:
10299: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10300: @subsubsection Object-Oriented Programming Style
10301: @cindex object-oriented programming style
10302: @cindex programming style, object-oriented
10303:
10304: This section is not exhaustive.
10305:
10306: @cindex stack effects of selectors
10307: @cindex selectors and stack effects
10308: In general, it is a good idea to ensure that all methods for the
10309: same selector have the same stack effect: when you invoke a selector,
10310: you often have no idea which method will be invoked, so, unless all
10311: methods have the same stack effect, you will not know the stack effect
10312: of the selector invocation.
10313:
10314: One exception to this rule is methods for the selector
10315: @code{construct}. We know which method is invoked, because we
10316: specify the class to be constructed at the same place. Actually, I
10317: defined @code{construct} as a selector only to give the users a
10318: convenient way to specify initialization. The way it is used, a
10319: mechanism different from selector invocation would be more natural
10320: (but probably would take more code and more space to explain).
10321:
10322: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10323: @subsubsection Class Binding
10324: @cindex class binding
10325: @cindex early binding
10326:
10327: @cindex late binding
10328: Normal selector invocations determine the method at run-time depending
10329: on the class of the receiving object. This run-time selection is called
10330: @i{late binding}.
10331:
10332: Sometimes it's preferable to invoke a different method. For example,
10333: you might want to use the simple method for @code{print}ing
10334: @code{object}s instead of the possibly long-winded @code{print} method
10335: of the receiver class. You can achieve this by replacing the invocation
10336: of @code{print} with:
10337:
10338: @cindex @code{[bind]} usage
10339: @example
10340: [bind] object print
10341: @end example
10342:
10343: @noindent
10344: in compiled code or:
10345:
10346: @cindex @code{bind} usage
10347: @example
10348: bind object print
10349: @end example
10350:
10351: @cindex class binding, alternative to
10352: @noindent
10353: in interpreted code. Alternatively, you can define the method with a
10354: name (e.g., @code{print-object}), and then invoke it through the
10355: name. Class binding is just a (often more convenient) way to achieve
10356: the same effect; it avoids name clutter and allows you to invoke
10357: methods directly without naming them first.
10358:
10359: @cindex superclass binding
10360: @cindex parent class binding
10361: A frequent use of class binding is this: When we define a method
10362: for a selector, we often want the method to do what the selector does
10363: in the parent class, and a little more. There is a special word for
10364: this purpose: @code{[parent]}; @code{[parent]
10365: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10366: selector}}, where @code{@emph{parent}} is the parent
10367: class of the current class. E.g., a method definition might look like:
10368:
10369: @cindex @code{[parent]} usage
10370: @example
10371: :noname
10372: dup [parent] foo \ do parent's foo on the receiving object
10373: ... \ do some more
10374: ; overrides foo
10375: @end example
10376:
10377: @cindex class binding as optimization
10378: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10379: March 1997), Andrew McKewan presents class binding as an optimization
10380: technique. I recommend not using it for this purpose unless you are in
10381: an emergency. Late binding is pretty fast with this model anyway, so the
10382: benefit of using class binding is small; the cost of using class binding
10383: where it is not appropriate is reduced maintainability.
10384:
10385: While we are at programming style questions: You should bind
10386: selectors only to ancestor classes of the receiving object. E.g., say,
10387: you know that the receiving object is of class @code{foo} or its
10388: descendents; then you should bind only to @code{foo} and its
10389: ancestors.
10390:
10391: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10392: @subsubsection Method conveniences
10393: @cindex method conveniences
10394:
10395: In a method you usually access the receiving object pretty often. If
10396: you define the method as a plain colon definition (e.g., with
10397: @code{:noname}), you may have to do a lot of stack
10398: gymnastics. To avoid this, you can define the method with @code{m:
10399: ... ;m}. E.g., you could define the method for
10400: @code{draw}ing a @code{circle} with
10401:
10402: @cindex @code{this} usage
10403: @cindex @code{m:} usage
10404: @cindex @code{;m} usage
10405: @example
10406: m: ( x y circle -- )
10407: ( x y ) this circle-radius @@ draw-circle ;m
10408: @end example
10409:
10410: @cindex @code{exit} in @code{m: ... ;m}
10411: @cindex @code{exitm} discussion
10412: @cindex @code{catch} in @code{m: ... ;m}
10413: When this method is executed, the receiver object is removed from the
10414: stack; you can access it with @code{this} (admittedly, in this
10415: example the use of @code{m: ... ;m} offers no advantage). Note
10416: that I specify the stack effect for the whole method (i.e. including
10417: the receiver object), not just for the code between @code{m:}
10418: and @code{;m}. You cannot use @code{exit} in
10419: @code{m:...;m}; instead, use
10420: @code{exitm}.@footnote{Moreover, for any word that calls
10421: @code{catch} and was defined before loading
10422: @code{objects.fs}, you have to redefine it like I redefined
10423: @code{catch}: @code{: catch this >r catch r> to-this ;}}
10424:
10425: @cindex @code{inst-var} usage
10426: You will frequently use sequences of the form @code{this
10427: @emph{field}} (in the example above: @code{this
10428: circle-radius}). If you use the field only in this way, you can
10429: define it with @code{inst-var} and eliminate the
10430: @code{this} before the field name. E.g., the @code{circle}
10431: class above could also be defined with:
10432:
10433: @example
10434: graphical class
10435: cell% inst-var radius
10436:
10437: m: ( x y circle -- )
10438: radius @@ draw-circle ;m
10439: overrides draw
10440:
10441: m: ( n-radius circle -- )
10442: radius ! ;m
10443: overrides construct
10444:
10445: end-class circle
10446: @end example
10447:
10448: @code{radius} can only be used in @code{circle} and its
10449: descendent classes and inside @code{m:...;m}.
10450:
10451: @cindex @code{inst-value} usage
10452: You can also define fields with @code{inst-value}, which is
10453: to @code{inst-var} what @code{value} is to
10454: @code{variable}. You can change the value of such a field with
10455: @code{[to-inst]}. E.g., we could also define the class
10456: @code{circle} like this:
10457:
10458: @example
10459: graphical class
10460: inst-value radius
10461:
10462: m: ( x y circle -- )
10463: radius draw-circle ;m
10464: overrides draw
10465:
10466: m: ( n-radius circle -- )
10467: [to-inst] radius ;m
10468: overrides construct
10469:
10470: end-class circle
10471: @end example
10472:
10473: @c !! :m is easy to confuse with m:. Another name would be better.
10474:
10475: @c Finally, you can define named methods with @code{:m}. One use of this
10476: @c feature is the definition of words that occur only in one class and are
10477: @c not intended to be overridden, but which still need method context
10478: @c (e.g., for accessing @code{inst-var}s). Another use is for methods that
10479: @c would be bound frequently, if defined anonymously.
10480:
10481:
10482: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
10483: @subsubsection Classes and Scoping
10484: @cindex classes and scoping
10485: @cindex scoping and classes
10486:
10487: Inheritance is frequent, unlike structure extension. This exacerbates
10488: the problem with the field name convention (@pxref{Structure Naming
10489: Convention}): One always has to remember in which class the field was
10490: originally defined; changing a part of the class structure would require
10491: changes for renaming in otherwise unaffected code.
10492:
10493: @cindex @code{inst-var} visibility
10494: @cindex @code{inst-value} visibility
10495: To solve this problem, I added a scoping mechanism (which was not in my
10496: original charter): A field defined with @code{inst-var} (or
10497: @code{inst-value}) is visible only in the class where it is defined and in
10498: the descendent classes of this class. Using such fields only makes
10499: sense in @code{m:}-defined methods in these classes anyway.
10500:
10501: This scoping mechanism allows us to use the unadorned field name,
10502: because name clashes with unrelated words become much less likely.
10503:
10504: @cindex @code{protected} discussion
10505: @cindex @code{private} discussion
10506: Once we have this mechanism, we can also use it for controlling the
10507: visibility of other words: All words defined after
10508: @code{protected} are visible only in the current class and its
10509: descendents. @code{public} restores the compilation
10510: (i.e. @code{current}) word list that was in effect before. If you
10511: have several @code{protected}s without an intervening
10512: @code{public} or @code{set-current}, @code{public}
10513: will restore the compilation word list in effect before the first of
10514: these @code{protected}s.
10515:
10516: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10517: @subsubsection Dividing classes
10518: @cindex Dividing classes
10519: @cindex @code{methods}...@code{end-methods}
10520:
10521: You may want to do the definition of methods separate from the
10522: definition of the class, its selectors, fields, and instance variables,
10523: i.e., separate the implementation from the definition. You can do this
10524: in the following way:
10525:
10526: @example
10527: graphical class
10528: inst-value radius
10529: end-class circle
10530:
10531: ... \ do some other stuff
10532:
10533: circle methods \ now we are ready
10534:
10535: m: ( x y circle -- )
10536: radius draw-circle ;m
10537: overrides draw
10538:
10539: m: ( n-radius circle -- )
10540: [to-inst] radius ;m
10541: overrides construct
10542:
10543: end-methods
10544: @end example
10545:
10546: You can use several @code{methods}...@code{end-methods} sections. The
10547: only things you can do to the class in these sections are: defining
10548: methods, and overriding the class's selectors. You must not define new
10549: selectors or fields.
10550:
10551: Note that you often have to override a selector before using it. In
10552: particular, you usually have to override @code{construct} with a new
10553: method before you can invoke @code{heap-new} and friends. E.g., you
10554: must not create a circle before the @code{overrides construct} sequence
10555: in the example above.
10556:
10557: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
10558: @subsubsection Object Interfaces
10559: @cindex object interfaces
10560: @cindex interfaces for objects
10561:
10562: In this model you can only call selectors defined in the class of the
10563: receiving objects or in one of its ancestors. If you call a selector
10564: with a receiving object that is not in one of these classes, the
10565: result is undefined; if you are lucky, the program crashes
10566: immediately.
10567:
10568: @cindex selectors common to hardly-related classes
10569: Now consider the case when you want to have a selector (or several)
10570: available in two classes: You would have to add the selector to a
10571: common ancestor class, in the worst case to @code{object}. You
10572: may not want to do this, e.g., because someone else is responsible for
10573: this ancestor class.
10574:
10575: The solution for this problem is interfaces. An interface is a
10576: collection of selectors. If a class implements an interface, the
10577: selectors become available to the class and its descendents. A class
10578: can implement an unlimited number of interfaces. For the problem
10579: discussed above, we would define an interface for the selector(s), and
10580: both classes would implement the interface.
10581:
10582: As an example, consider an interface @code{storage} for
10583: writing objects to disk and getting them back, and a class
10584: @code{foo} that implements it. The code would look like this:
10585:
10586: @cindex @code{interface} usage
10587: @cindex @code{end-interface} usage
10588: @cindex @code{implementation} usage
10589: @example
10590: interface
10591: selector write ( file object -- )
10592: selector read1 ( file object -- )
10593: end-interface storage
10594:
10595: bar class
10596: storage implementation
10597:
10598: ... overrides write
10599: ... overrides read1
10600: ...
10601: end-class foo
10602: @end example
10603:
10604: @noindent
10605: (I would add a word @code{read} @i{( file -- object )} that uses
10606: @code{read1} internally, but that's beyond the point illustrated
10607: here.)
10608:
10609: Note that you cannot use @code{protected} in an interface; and
10610: of course you cannot define fields.
10611:
10612: In the Neon model, all selectors are available for all classes;
10613: therefore it does not need interfaces. The price you pay in this model
10614: is slower late binding, and therefore, added complexity to avoid late
10615: binding.
10616:
10617: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10618: @subsubsection @file{objects.fs} Implementation
10619: @cindex @file{objects.fs} implementation
10620:
10621: @cindex @code{object-map} discussion
10622: An object is a piece of memory, like one of the data structures
10623: described with @code{struct...end-struct}. It has a field
10624: @code{object-map} that points to the method map for the object's
10625: class.
10626:
10627: @cindex method map
10628: @cindex virtual function table
10629: The @emph{method map}@footnote{This is Self terminology; in C++
10630: terminology: virtual function table.} is an array that contains the
10631: execution tokens (@i{xt}s) of the methods for the object's class. Each
10632: selector contains an offset into a method map.
10633:
10634: @cindex @code{selector} implementation, class
10635: @code{selector} is a defining word that uses
10636: @code{CREATE} and @code{DOES>}. The body of the
10637: selector contains the offset; the @code{DOES>} action for a
10638: class selector is, basically:
10639:
10640: @example
10641: ( object addr ) @@ over object-map @@ + @@ execute
10642: @end example
10643:
10644: Since @code{object-map} is the first field of the object, it
10645: does not generate any code. As you can see, calling a selector has a
10646: small, constant cost.
10647:
10648: @cindex @code{current-interface} discussion
10649: @cindex class implementation and representation
10650: A class is basically a @code{struct} combined with a method
10651: map. During the class definition the alignment and size of the class
10652: are passed on the stack, just as with @code{struct}s, so
10653: @code{field} can also be used for defining class
10654: fields. However, passing more items on the stack would be
10655: inconvenient, so @code{class} builds a data structure in memory,
10656: which is accessed through the variable
10657: @code{current-interface}. After its definition is complete, the
10658: class is represented on the stack by a pointer (e.g., as parameter for
10659: a child class definition).
10660:
10661: A new class starts off with the alignment and size of its parent,
10662: and a copy of the parent's method map. Defining new fields extends the
10663: size and alignment; likewise, defining new selectors extends the
10664: method map. @code{overrides} just stores a new @i{xt} in the method
10665: map at the offset given by the selector.
10666:
10667: @cindex class binding, implementation
10668: Class binding just gets the @i{xt} at the offset given by the selector
10669: from the class's method map and @code{compile,}s (in the case of
10670: @code{[bind]}) it.
10671:
10672: @cindex @code{this} implementation
10673: @cindex @code{catch} and @code{this}
10674: @cindex @code{this} and @code{catch}
10675: I implemented @code{this} as a @code{value}. At the
10676: start of an @code{m:...;m} method the old @code{this} is
10677: stored to the return stack and restored at the end; and the object on
10678: the TOS is stored @code{TO this}. This technique has one
10679: disadvantage: If the user does not leave the method via
10680: @code{;m}, but via @code{throw} or @code{exit},
10681: @code{this} is not restored (and @code{exit} may
10682: crash). To deal with the @code{throw} problem, I have redefined
10683: @code{catch} to save and restore @code{this}; the same
10684: should be done with any word that can catch an exception. As for
10685: @code{exit}, I simply forbid it (as a replacement, there is
10686: @code{exitm}).
10687:
10688: @cindex @code{inst-var} implementation
10689: @code{inst-var} is just the same as @code{field}, with
10690: a different @code{DOES>} action:
10691: @example
10692: @@ this +
10693: @end example
10694: Similar for @code{inst-value}.
10695:
10696: @cindex class scoping implementation
10697: Each class also has a word list that contains the words defined with
10698: @code{inst-var} and @code{inst-value}, and its protected
10699: words. It also has a pointer to its parent. @code{class} pushes
10700: the word lists of the class and all its ancestors onto the search order stack,
10701: and @code{end-class} drops them.
10702:
10703: @cindex interface implementation
10704: An interface is like a class without fields, parent and protected
10705: words; i.e., it just has a method map. If a class implements an
10706: interface, its method map contains a pointer to the method map of the
10707: interface. The positive offsets in the map are reserved for class
10708: methods, therefore interface map pointers have negative
10709: offsets. Interfaces have offsets that are unique throughout the
10710: system, unlike class selectors, whose offsets are only unique for the
10711: classes where the selector is available (invokable).
10712:
10713: This structure means that interface selectors have to perform one
10714: indirection more than class selectors to find their method. Their body
10715: contains the interface map pointer offset in the class method map, and
10716: the method offset in the interface method map. The
10717: @code{does>} action for an interface selector is, basically:
10718:
10719: @example
10720: ( object selector-body )
10721: 2dup selector-interface @@ ( object selector-body object interface-offset )
10722: swap object-map @@ + @@ ( object selector-body map )
10723: swap selector-offset @@ + @@ execute
10724: @end example
10725:
10726: where @code{object-map} and @code{selector-offset} are
10727: first fields and generate no code.
10728:
10729: As a concrete example, consider the following code:
10730:
10731: @example
10732: interface
10733: selector if1sel1
10734: selector if1sel2
10735: end-interface if1
10736:
10737: object class
10738: if1 implementation
10739: selector cl1sel1
10740: cell% inst-var cl1iv1
10741:
10742: ' m1 overrides construct
10743: ' m2 overrides if1sel1
10744: ' m3 overrides if1sel2
10745: ' m4 overrides cl1sel2
10746: end-class cl1
10747:
10748: create obj1 object dict-new drop
10749: create obj2 cl1 dict-new drop
10750: @end example
10751:
10752: The data structure created by this code (including the data structure
10753: for @code{object}) is shown in the
10754: @uref{objects-implementation.eps,figure}, assuming a cell size of 4.
10755: @comment TODO add this diagram..
10756:
10757: @node Objects Glossary, , Objects Implementation, Objects
10758: @subsubsection @file{objects.fs} Glossary
10759: @cindex @file{objects.fs} Glossary
10760:
10761:
10762: doc---objects-bind
10763: doc---objects-<bind>
10764: doc---objects-bind'
10765: doc---objects-[bind]
10766: doc---objects-class
10767: doc---objects-class->map
10768: doc---objects-class-inst-size
10769: doc---objects-class-override!
10770: doc---objects-class-previous
10771: doc---objects-class>order
10772: doc---objects-construct
10773: doc---objects-current'
10774: doc---objects-[current]
10775: doc---objects-current-interface
10776: doc---objects-dict-new
10777: doc---objects-end-class
10778: doc---objects-end-class-noname
10779: doc---objects-end-interface
10780: doc---objects-end-interface-noname
10781: doc---objects-end-methods
10782: doc---objects-exitm
10783: doc---objects-heap-new
10784: doc---objects-implementation
10785: doc---objects-init-object
10786: doc---objects-inst-value
10787: doc---objects-inst-var
10788: doc---objects-interface
10789: doc---objects-m:
10790: doc---objects-:m
10791: doc---objects-;m
10792: doc---objects-method
10793: doc---objects-methods
10794: doc---objects-object
10795: doc---objects-overrides
10796: doc---objects-[parent]
10797: doc---objects-print
10798: doc---objects-protected
10799: doc---objects-public
10800: doc---objects-selector
10801: doc---objects-this
10802: doc---objects-<to-inst>
10803: doc---objects-[to-inst]
10804: doc---objects-to-this
10805: doc---objects-xt-new
10806:
10807:
10808: @c -------------------------------------------------------------
10809: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10810: @subsection The @file{oof.fs} model
10811: @cindex oof
10812: @cindex object-oriented programming
10813:
10814: @cindex @file{objects.fs}
10815: @cindex @file{oof.fs}
10816:
10817: This section describes the @file{oof.fs} package.
10818:
10819: The package described in this section has been used in bigFORTH since 1991, and
10820: used for two large applications: a chromatographic system used to
10821: create new medicaments, and a graphic user interface library (MINOS).
10822:
10823: You can find a description (in German) of @file{oof.fs} in @cite{Object
10824: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10825: 10(2), 1994.
10826:
10827: @menu
10828: * Properties of the OOF model::
10829: * Basic OOF Usage::
10830: * The OOF base class::
10831: * Class Declaration::
10832: * Class Implementation::
10833: @end menu
10834:
10835: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
10836: @subsubsection Properties of the @file{oof.fs} model
10837: @cindex @file{oof.fs} properties
10838:
10839: @itemize @bullet
10840: @item
10841: This model combines object oriented programming with information
10842: hiding. It helps you writing large application, where scoping is
10843: necessary, because it provides class-oriented scoping.
10844:
10845: @item
10846: Named objects, object pointers, and object arrays can be created,
10847: selector invocation uses the ``object selector'' syntax. Selector invocation
10848: to objects and/or selectors on the stack is a bit less convenient, but
10849: possible.
10850:
10851: @item
10852: Selector invocation and instance variable usage of the active object is
10853: straightforward, since both make use of the active object.
10854:
10855: @item
10856: Late binding is efficient and easy to use.
10857:
10858: @item
10859: State-smart objects parse selectors. However, extensibility is provided
10860: using a (parsing) selector @code{postpone} and a selector @code{'}.
10861:
10862: @item
10863: An implementation in ANS Forth is available.
10864:
10865: @end itemize
10866:
10867:
10868: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
10869: @subsubsection Basic @file{oof.fs} Usage
10870: @cindex @file{oof.fs} usage
10871:
10872: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
10873:
10874: You can define a class for graphical objects like this:
10875:
10876: @cindex @code{class} usage
10877: @cindex @code{class;} usage
10878: @cindex @code{method} usage
10879: @example
10880: object class graphical \ "object" is the parent class
10881: method draw ( x y graphical -- )
10882: class;
10883: @end example
10884:
10885: This code defines a class @code{graphical} with an
10886: operation @code{draw}. We can perform the operation
10887: @code{draw} on any @code{graphical} object, e.g.:
10888:
10889: @example
10890: 100 100 t-rex draw
10891: @end example
10892:
10893: @noindent
10894: where @code{t-rex} is an object or object pointer, created with e.g.
10895: @code{graphical : t-rex}.
10896:
10897: @cindex abstract class
10898: How do we create a graphical object? With the present definitions,
10899: we cannot create a useful graphical object. The class
10900: @code{graphical} describes graphical objects in general, but not
10901: any concrete graphical object type (C++ users would call it an
10902: @emph{abstract class}); e.g., there is no method for the selector
10903: @code{draw} in the class @code{graphical}.
10904:
10905: For concrete graphical objects, we define child classes of the
10906: class @code{graphical}, e.g.:
10907:
10908: @example
10909: graphical class circle \ "graphical" is the parent class
10910: cell var circle-radius
10911: how:
10912: : draw ( x y -- )
10913: circle-radius @@ draw-circle ;
10914:
10915: : init ( n-radius -- (
10916: circle-radius ! ;
10917: class;
10918: @end example
10919:
10920: Here we define a class @code{circle} as a child of @code{graphical},
10921: with a field @code{circle-radius}; it defines new methods for the
10922: selectors @code{draw} and @code{init} (@code{init} is defined in
10923: @code{object}, the parent class of @code{graphical}).
10924:
10925: Now we can create a circle in the dictionary with:
10926:
10927: @example
10928: 50 circle : my-circle
10929: @end example
10930:
10931: @noindent
10932: @code{:} invokes @code{init}, thus initializing the field
10933: @code{circle-radius} with 50. We can draw this new circle at (100,100)
10934: with:
10935:
10936: @example
10937: 100 100 my-circle draw
10938: @end example
10939:
10940: @cindex selector invocation, restrictions
10941: @cindex class definition, restrictions
10942: Note: You can only invoke a selector if the receiving object belongs to
10943: the class where the selector was defined or one of its descendents;
10944: e.g., you can invoke @code{draw} only for objects belonging to
10945: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
10946: mechanism will check if you try to invoke a selector that is not
10947: defined in this class hierarchy, so you'll get an error at compilation
10948: time.
10949:
10950:
10951: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
10952: @subsubsection The @file{oof.fs} base class
10953: @cindex @file{oof.fs} base class
10954:
10955: When you define a class, you have to specify a parent class. So how do
10956: you start defining classes? There is one class available from the start:
10957: @code{object}. You have to use it as ancestor for all classes. It is the
10958: only class that has no parent. Classes are also objects, except that
10959: they don't have instance variables; class manipulation such as
10960: inheritance or changing definitions of a class is handled through
10961: selectors of the class @code{object}.
10962:
10963: @code{object} provides a number of selectors:
10964:
10965: @itemize @bullet
10966: @item
10967: @code{class} for subclassing, @code{definitions} to add definitions
10968: later on, and @code{class?} to get type informations (is the class a
10969: subclass of the class passed on the stack?).
10970:
10971: doc---object-class
10972: doc---object-definitions
10973: doc---object-class?
10974:
10975:
10976: @item
10977: @code{init} and @code{dispose} as constructor and destructor of the
10978: object. @code{init} is invocated after the object's memory is allocated,
10979: while @code{dispose} also handles deallocation. Thus if you redefine
10980: @code{dispose}, you have to call the parent's dispose with @code{super
10981: dispose}, too.
10982:
10983: doc---object-init
10984: doc---object-dispose
10985:
10986:
10987: @item
10988: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
10989: @code{[]} to create named and unnamed objects and object arrays or
10990: object pointers.
10991:
10992: doc---object-new
10993: doc---object-new[]
10994: doc---object-:
10995: doc---object-ptr
10996: doc---object-asptr
10997: doc---object-[]
10998:
10999:
11000: @item
11001: @code{::} and @code{super} for explicit scoping. You should use explicit
11002: scoping only for super classes or classes with the same set of instance
11003: variables. Explicitly-scoped selectors use early binding.
11004:
11005: doc---object-::
11006: doc---object-super
11007:
11008:
11009: @item
11010: @code{self} to get the address of the object
11011:
11012: doc---object-self
11013:
11014:
11015: @item
11016: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
11017: pointers and instance defers.
11018:
11019: doc---object-bind
11020: doc---object-bound
11021: doc---object-link
11022: doc---object-is
11023:
11024:
11025: @item
11026: @code{'} to obtain selector tokens, @code{send} to invocate selectors
11027: form the stack, and @code{postpone} to generate selector invocation code.
11028:
11029: doc---object-'
11030: doc---object-postpone
11031:
11032:
11033: @item
11034: @code{with} and @code{endwith} to select the active object from the
11035: stack, and enable its scope. Using @code{with} and @code{endwith}
11036: also allows you to create code using selector @code{postpone} without being
11037: trapped by the state-smart objects.
11038:
11039: doc---object-with
11040: doc---object-endwith
11041:
11042:
11043: @end itemize
11044:
11045: @node Class Declaration, Class Implementation, The OOF base class, OOF
11046: @subsubsection Class Declaration
11047: @cindex class declaration
11048:
11049: @itemize @bullet
11050: @item
11051: Instance variables
11052:
11053: doc---oof-var
11054:
11055:
11056: @item
11057: Object pointers
11058:
11059: doc---oof-ptr
11060: doc---oof-asptr
11061:
11062:
11063: @item
11064: Instance defers
11065:
11066: doc---oof-defer
11067:
11068:
11069: @item
11070: Method selectors
11071:
11072: doc---oof-early
11073: doc---oof-method
11074:
11075:
11076: @item
11077: Class-wide variables
11078:
11079: doc---oof-static
11080:
11081:
11082: @item
11083: End declaration
11084:
11085: doc---oof-how:
11086: doc---oof-class;
11087:
11088:
11089: @end itemize
11090:
11091: @c -------------------------------------------------------------
11092: @node Class Implementation, , Class Declaration, OOF
11093: @subsubsection Class Implementation
11094: @cindex class implementation
11095:
11096: @c -------------------------------------------------------------
11097: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
11098: @subsection The @file{mini-oof.fs} model
11099: @cindex mini-oof
11100:
11101: Gforth's third object oriented Forth package is a 12-liner. It uses a
11102: mixture of the @file{objects.fs} and the @file{oof.fs} syntax,
11103: and reduces to the bare minimum of features. This is based on a posting
11104: of Bernd Paysan in comp.lang.forth.
11105:
11106: @menu
11107: * Basic Mini-OOF Usage::
11108: * Mini-OOF Example::
11109: * Mini-OOF Implementation::
11110: @end menu
11111:
11112: @c -------------------------------------------------------------
11113: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
11114: @subsubsection Basic @file{mini-oof.fs} Usage
11115: @cindex mini-oof usage
11116:
11117: There is a base class (@code{class}, which allocates one cell for the
11118: object pointer) plus seven other words: to define a method, a variable,
11119: a class; to end a class, to resolve binding, to allocate an object and
11120: to compile a class method.
11121: @comment TODO better description of the last one
11122:
11123:
11124: doc-object
11125: doc-method
11126: doc-var
11127: doc-class
11128: doc-end-class
11129: doc-defines
11130: doc-new
11131: doc-::
11132:
11133:
11134:
11135: @c -------------------------------------------------------------
11136: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
11137: @subsubsection Mini-OOF Example
11138: @cindex mini-oof example
11139:
11140: A short example shows how to use this package. This example, in slightly
11141: extended form, is supplied as @file{moof-exm.fs}
11142: @comment TODO could flesh this out with some comments from the Forthwrite article
11143:
11144: @example
11145: object class
11146: method init
11147: method draw
11148: end-class graphical
11149: @end example
11150:
11151: This code defines a class @code{graphical} with an
11152: operation @code{draw}. We can perform the operation
11153: @code{draw} on any @code{graphical} object, e.g.:
11154:
11155: @example
11156: 100 100 t-rex draw
11157: @end example
11158:
11159: where @code{t-rex} is an object or object pointer, created with e.g.
11160: @code{graphical new Constant t-rex}.
11161:
11162: For concrete graphical objects, we define child classes of the
11163: class @code{graphical}, e.g.:
11164:
11165: @example
11166: graphical class
11167: cell var circle-radius
11168: end-class circle \ "graphical" is the parent class
11169:
11170: :noname ( x y -- )
11171: circle-radius @@ draw-circle ; circle defines draw
11172: :noname ( r -- )
11173: circle-radius ! ; circle defines init
11174: @end example
11175:
11176: There is no implicit init method, so we have to define one. The creation
11177: code of the object now has to call init explicitely.
11178:
11179: @example
11180: circle new Constant my-circle
11181: 50 my-circle init
11182: @end example
11183:
11184: It is also possible to add a function to create named objects with
11185: automatic call of @code{init}, given that all objects have @code{init}
11186: on the same place:
11187:
11188: @example
11189: : new: ( .. o "name" -- )
11190: new dup Constant init ;
11191: 80 circle new: large-circle
11192: @end example
11193:
11194: We can draw this new circle at (100,100) with:
11195:
11196: @example
11197: 100 100 my-circle draw
11198: @end example
11199:
11200: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
11201: @subsubsection @file{mini-oof.fs} Implementation
11202:
11203: Object-oriented systems with late binding typically use a
11204: ``vtable''-approach: the first variable in each object is a pointer to a
11205: table, which contains the methods as function pointers. The vtable
11206: may also contain other information.
11207:
11208: So first, let's declare selectors:
11209:
11210: @example
11211: : method ( m v "name" -- m' v ) Create over , swap cell+ swap
11212: DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
11213: @end example
11214:
11215: During selector declaration, the number of selectors and instance
11216: variables is on the stack (in address units). @code{method} creates one
11217: selector and increments the selector number. To execute a selector, it
11218: takes the object, fetches the vtable pointer, adds the offset, and
11219: executes the method @i{xt} stored there. Each selector takes the object
11220: it is invoked with as top of stack parameter; it passes the parameters
11221: (including the object) unchanged to the appropriate method which should
11222: consume that object.
11223:
11224: Now, we also have to declare instance variables
11225:
11226: @example
11227: : var ( m v size "name" -- m v' ) Create over , +
11228: DOES> ( o -- addr ) @@ + ;
11229: @end example
11230:
11231: As before, a word is created with the current offset. Instance
11232: variables can have different sizes (cells, floats, doubles, chars), so
11233: all we do is take the size and add it to the offset. If your machine
11234: has alignment restrictions, put the proper @code{aligned} or
11235: @code{faligned} before the variable, to adjust the variable
11236: offset. That's why it is on the top of stack.
11237:
11238: We need a starting point (the base object) and some syntactic sugar:
11239:
11240: @example
11241: Create object 1 cells , 2 cells ,
11242: : class ( class -- class selectors vars ) dup 2@@ ;
11243: @end example
11244:
11245: For inheritance, the vtable of the parent object has to be
11246: copied when a new, derived class is declared. This gives all the
11247: methods of the parent class, which can be overridden, though.
11248:
11249: @example
11250: : end-class ( class selectors vars "name" -- )
11251: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
11252: cell+ dup cell+ r> rot @@ 2 cells /string move ;
11253: @end example
11254:
11255: The first line creates the vtable, initialized with
11256: @code{noop}s. The second line is the inheritance mechanism, it
11257: copies the xts from the parent vtable.
11258:
11259: We still have no way to define new methods, let's do that now:
11260:
11261: @example
11262: : defines ( xt class "name" -- ) ' >body @@ + ! ;
11263: @end example
11264:
11265: To allocate a new object, we need a word, too:
11266:
11267: @example
11268: : new ( class -- o ) here over @@ allot swap over ! ;
11269: @end example
11270:
11271: Sometimes derived classes want to access the method of the
11272: parent object. There are two ways to achieve this with Mini-OOF:
11273: first, you could use named words, and second, you could look up the
11274: vtable of the parent object.
11275:
11276: @example
11277: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
11278: @end example
11279:
11280:
11281: Nothing can be more confusing than a good example, so here is
11282: one. First let's declare a text object (called
11283: @code{button}), that stores text and position:
11284:
11285: @example
11286: object class
11287: cell var text
11288: cell var len
11289: cell var x
11290: cell var y
11291: method init
11292: method draw
11293: end-class button
11294: @end example
11295:
11296: @noindent
11297: Now, implement the two methods, @code{draw} and @code{init}:
11298:
11299: @example
11300: :noname ( o -- )
11301: >r r@@ x @@ r@@ y @@ at-xy r@@ text @@ r> len @@ type ;
11302: button defines draw
11303: :noname ( addr u o -- )
11304: >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
11305: button defines init
11306: @end example
11307:
11308: @noindent
11309: To demonstrate inheritance, we define a class @code{bold-button}, with no
11310: new data and no new selectors:
11311:
11312: @example
11313: button class
11314: end-class bold-button
11315:
11316: : bold 27 emit ." [1m" ;
11317: : normal 27 emit ." [0m" ;
11318: @end example
11319:
11320: @noindent
11321: The class @code{bold-button} has a different draw method to
11322: @code{button}, but the new method is defined in terms of the draw method
11323: for @code{button}:
11324:
11325: @example
11326: :noname bold [ button :: draw ] normal ; bold-button defines draw
11327: @end example
11328:
11329: @noindent
11330: Finally, create two objects and apply selectors:
11331:
11332: @example
11333: button new Constant foo
11334: s" thin foo" foo init
11335: page
11336: foo draw
11337: bold-button new Constant bar
11338: s" fat bar" bar init
11339: 1 bar y !
11340: bar draw
11341: @end example
11342:
11343:
11344: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11345: @subsection Comparison with other object models
11346: @cindex comparison of object models
11347: @cindex object models, comparison
11348:
11349: Many object-oriented Forth extensions have been proposed (@cite{A survey
11350: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11351: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11352: relation of the object models described here to two well-known and two
11353: closely-related (by the use of method maps) models. Andras Zsoter
11354: helped us with this section.
11355:
11356: @cindex Neon model
11357: The most popular model currently seems to be the Neon model (see
11358: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11359: 1997) by Andrew McKewan) but this model has a number of limitations
11360: @footnote{A longer version of this critique can be
11361: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11362: Dimensions, May 1997) by Anton Ertl.}:
11363:
11364: @itemize @bullet
11365: @item
11366: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11367: to pass objects on the stack.
11368:
11369: @item
11370: It requires that the selector parses the input stream (at
11371: compile time); this leads to reduced extensibility and to bugs that are
11372: hard to find.
11373:
11374: @item
11375: It allows using every selector on every object; this eliminates the
11376: need for interfaces, but makes it harder to create efficient
11377: implementations.
11378: @end itemize
11379:
11380: @cindex Pountain's object-oriented model
11381: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11382: Press, London, 1987) by Dick Pountain. However, it is not really about
11383: object-oriented programming, because it hardly deals with late
11384: binding. Instead, it focuses on features like information hiding and
11385: overloading that are characteristic of modular languages like Ada (83).
11386:
11387: @cindex Zsoter's object-oriented model
11388: In @uref{http://www.forth.org/oopf.html, Does late binding have to be
11389: slow?} (Forth Dimensions 18(1) 1996, pages 31-35) Andras Zsoter
11390: describes a model that makes heavy use of an active object (like
11391: @code{this} in @file{objects.fs}): The active object is not only used
11392: for accessing all fields, but also specifies the receiving object of
11393: every selector invocation; you have to change the active object
11394: explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
11395: changes more or less implicitly at @code{m: ... ;m}. Such a change at
11396: the method entry point is unnecessary with Zsoter's model, because the
11397: receiving object is the active object already. On the other hand, the
11398: explicit change is absolutely necessary in that model, because otherwise
11399: no one could ever change the active object. An ANS Forth implementation
11400: of this model is available through
11401: @uref{http://www.forth.org/oopf.html}.
11402:
11403: @cindex @file{oof.fs}, differences to other models
11404: The @file{oof.fs} model combines information hiding and overloading
11405: resolution (by keeping names in various word lists) with object-oriented
11406: programming. It sets the active object implicitly on method entry, but
11407: also allows explicit changing (with @code{>o...o>} or with
11408: @code{with...endwith}). It uses parsing and state-smart objects and
11409: classes for resolving overloading and for early binding: the object or
11410: class parses the selector and determines the method from this. If the
11411: selector is not parsed by an object or class, it performs a call to the
11412: selector for the active object (late binding), like Zsoter's model.
11413: Fields are always accessed through the active object. The big
11414: disadvantage of this model is the parsing and the state-smartness, which
11415: reduces extensibility and increases the opportunities for subtle bugs;
11416: essentially, you are only safe if you never tick or @code{postpone} an
11417: object or class (Bernd disagrees, but I (Anton) am not convinced).
11418:
11419: @cindex @file{mini-oof.fs}, differences to other models
11420: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11421: version of the @file{objects.fs} model, but syntactically it is a
11422: mixture of the @file{objects.fs} and @file{oof.fs} models.
11423:
11424:
11425: @c -------------------------------------------------------------
11426: @node Programming Tools, Assembler and Code Words, Object-oriented Forth, Words
11427: @section Programming Tools
11428: @cindex programming tools
11429:
11430: @c !! move this and assembler down below OO stuff.
11431:
11432: @menu
11433: * Examining::
11434: * Forgetting words::
11435: * Debugging:: Simple and quick.
11436: * Assertions:: Making your programs self-checking.
11437: * Singlestep Debugger:: Executing your program word by word.
11438: @end menu
11439:
11440: @node Examining, Forgetting words, Programming Tools, Programming Tools
11441: @subsection Examining data and code
11442: @cindex examining data and code
11443: @cindex data examination
11444: @cindex code examination
11445:
11446: The following words inspect the stack non-destructively:
11447:
11448: doc-.s
11449: doc-f.s
11450:
11451: There is a word @code{.r} but it does @i{not} display the return stack!
11452: It is used for formatted numeric output (@pxref{Simple numeric output}).
11453:
11454: doc-depth
11455: doc-fdepth
11456: doc-clearstack
11457:
11458: The following words inspect memory.
11459:
11460: doc-?
11461: doc-dump
11462:
11463: And finally, @code{see} allows to inspect code:
11464:
11465: doc-see
11466: doc-xt-see
11467:
11468: @node Forgetting words, Debugging, Examining, Programming Tools
11469: @subsection Forgetting words
11470: @cindex words, forgetting
11471: @cindex forgeting words
11472:
11473: @c anton: other, maybe better places for this subsection: Defining Words;
11474: @c Dictionary allocation. At least a reference should be there.
11475:
11476: Forth allows you to forget words (and everything that was alloted in the
11477: dictonary after them) in a LIFO manner.
11478:
11479: doc-marker
11480:
11481: The most common use of this feature is during progam development: when
11482: you change a source file, forget all the words it defined and load it
11483: again (since you also forget everything defined after the source file
11484: was loaded, you have to reload that, too). Note that effects like
11485: storing to variables and destroyed system words are not undone when you
11486: forget words. With a system like Gforth, that is fast enough at
11487: starting up and compiling, I find it more convenient to exit and restart
11488: Gforth, as this gives me a clean slate.
11489:
11490: Here's an example of using @code{marker} at the start of a source file
11491: that you are debugging; it ensures that you only ever have one copy of
11492: the file's definitions compiled at any time:
11493:
11494: @example
11495: [IFDEF] my-code
11496: my-code
11497: [ENDIF]
11498:
11499: marker my-code
11500: init-included-files
11501:
11502: \ .. definitions start here
11503: \ .
11504: \ .
11505: \ end
11506: @end example
11507:
11508:
11509: @node Debugging, Assertions, Forgetting words, Programming Tools
11510: @subsection Debugging
11511: @cindex debugging
11512:
11513: Languages with a slow edit/compile/link/test development loop tend to
11514: require sophisticated tracing/stepping debuggers to facilate debugging.
11515:
11516: A much better (faster) way in fast-compiling languages is to add
11517: printing code at well-selected places, let the program run, look at
11518: the output, see where things went wrong, add more printing code, etc.,
11519: until the bug is found.
11520:
11521: The simple debugging aids provided in @file{debugs.fs}
11522: are meant to support this style of debugging.
11523:
11524: The word @code{~~} prints debugging information (by default the source
11525: location and the stack contents). It is easy to insert. If you use Emacs
11526: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
11527: query-replace them with nothing). The deferred words
11528: @code{printdebugdata} and @code{.debugline} control the output of
11529: @code{~~}. The default source location output format works well with
11530: Emacs' compilation mode, so you can step through the program at the
11531: source level using @kbd{C-x `} (the advantage over a stepping debugger
11532: is that you can step in any direction and you know where the crash has
11533: happened or where the strange data has occurred).
11534:
11535: doc-~~
11536: doc-printdebugdata
11537: doc-.debugline
11538:
11539: @cindex filenames in @code{~~} output
11540: @code{~~} (and assertions) will usually print the wrong file name if a
11541: marker is executed in the same file after their occurance. They will
11542: print @samp{*somewhere*} as file name if a marker is executed in the
11543: same file before their occurance.
11544:
11545:
11546: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
11547: @subsection Assertions
11548: @cindex assertions
11549:
11550: It is a good idea to make your programs self-checking, especially if you
11551: make an assumption that may become invalid during maintenance (for
11552: example, that a certain field of a data structure is never zero). Gforth
11553: supports @dfn{assertions} for this purpose. They are used like this:
11554:
11555: @example
11556: assert( @i{flag} )
11557: @end example
11558:
11559: The code between @code{assert(} and @code{)} should compute a flag, that
11560: should be true if everything is alright and false otherwise. It should
11561: not change anything else on the stack. The overall stack effect of the
11562: assertion is @code{( -- )}. E.g.
11563:
11564: @example
11565: assert( 1 1 + 2 = ) \ what we learn in school
11566: assert( dup 0<> ) \ assert that the top of stack is not zero
11567: assert( false ) \ this code should not be reached
11568: @end example
11569:
11570: The need for assertions is different at different times. During
11571: debugging, we want more checking, in production we sometimes care more
11572: for speed. Therefore, assertions can be turned off, i.e., the assertion
11573: becomes a comment. Depending on the importance of an assertion and the
11574: time it takes to check it, you may want to turn off some assertions and
11575: keep others turned on. Gforth provides several levels of assertions for
11576: this purpose:
11577:
11578:
11579: doc-assert0(
11580: doc-assert1(
11581: doc-assert2(
11582: doc-assert3(
11583: doc-assert(
11584: doc-)
11585:
11586:
11587: The variable @code{assert-level} specifies the highest assertions that
11588: are turned on. I.e., at the default @code{assert-level} of one,
11589: @code{assert0(} and @code{assert1(} assertions perform checking, while
11590: @code{assert2(} and @code{assert3(} assertions are treated as comments.
11591:
11592: The value of @code{assert-level} is evaluated at compile-time, not at
11593: run-time. Therefore you cannot turn assertions on or off at run-time;
11594: you have to set the @code{assert-level} appropriately before compiling a
11595: piece of code. You can compile different pieces of code at different
11596: @code{assert-level}s (e.g., a trusted library at level 1 and
11597: newly-written code at level 3).
11598:
11599:
11600: doc-assert-level
11601:
11602:
11603: If an assertion fails, a message compatible with Emacs' compilation mode
11604: is produced and the execution is aborted (currently with @code{ABORT"}.
11605: If there is interest, we will introduce a special throw code. But if you
11606: intend to @code{catch} a specific condition, using @code{throw} is
11607: probably more appropriate than an assertion).
11608:
11609: @cindex filenames in assertion output
11610: Assertions (and @code{~~}) will usually print the wrong file name if a
11611: marker is executed in the same file after their occurance. They will
11612: print @samp{*somewhere*} as file name if a marker is executed in the
11613: same file before their occurance.
11614:
11615: Definitions in ANS Forth for these assertion words are provided
11616: in @file{compat/assert.fs}.
11617:
11618:
11619: @node Singlestep Debugger, , Assertions, Programming Tools
11620: @subsection Singlestep Debugger
11621: @cindex singlestep Debugger
11622: @cindex debugging Singlestep
11623:
11624: When you create a new word there's often the need to check whether it
11625: behaves correctly or not. You can do this by typing @code{dbg
11626: badword}. A debug session might look like this:
11627:
11628: @example
11629: : badword 0 DO i . LOOP ; ok
11630: 2 dbg badword
11631: : badword
11632: Scanning code...
11633:
11634: Nesting debugger ready!
11635:
11636: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
11637: 400D4740 8049F68 DO -> [ 0 ]
11638: 400D4744 804A0C8 i -> [ 1 ] 00000
11639: 400D4748 400C5E60 . -> 0 [ 0 ]
11640: 400D474C 8049D0C LOOP -> [ 0 ]
11641: 400D4744 804A0C8 i -> [ 1 ] 00001
11642: 400D4748 400C5E60 . -> 1 [ 0 ]
11643: 400D474C 8049D0C LOOP -> [ 0 ]
11644: 400D4758 804B384 ; -> ok
11645: @end example
11646:
11647: Each line displayed is one step. You always have to hit return to
11648: execute the next word that is displayed. If you don't want to execute
11649: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
11650: an overview what keys are available:
11651:
11652: @table @i
11653:
11654: @item @key{RET}
11655: Next; Execute the next word.
11656:
11657: @item n
11658: Nest; Single step through next word.
11659:
11660: @item u
11661: Unnest; Stop debugging and execute rest of word. If we got to this word
11662: with nest, continue debugging with the calling word.
11663:
11664: @item d
11665: Done; Stop debugging and execute rest.
11666:
11667: @item s
11668: Stop; Abort immediately.
11669:
11670: @end table
11671:
11672: Debugging large application with this mechanism is very difficult, because
11673: you have to nest very deeply into the program before the interesting part
11674: begins. This takes a lot of time.
11675:
11676: To do it more directly put a @code{BREAK:} command into your source code.
11677: When program execution reaches @code{BREAK:} the single step debugger is
11678: invoked and you have all the features described above.
11679:
11680: If you have more than one part to debug it is useful to know where the
11681: program has stopped at the moment. You can do this by the
11682: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
11683: string is typed out when the ``breakpoint'' is reached.
11684:
11685:
11686: doc-dbg
11687: doc-break:
11688: doc-break"
11689:
11690:
11691:
11692: @c -------------------------------------------------------------
11693: @node Assembler and Code Words, Threading Words, Programming Tools, Words
11694: @section Assembler and Code Words
11695: @cindex assembler
11696: @cindex code words
11697:
11698: @menu
11699: * Code and ;code::
11700: * Common Assembler:: Assembler Syntax
11701: * Common Disassembler::
11702: * 386 Assembler:: Deviations and special cases
11703: * Alpha Assembler:: Deviations and special cases
11704: * MIPS assembler:: Deviations and special cases
11705: * Other assemblers:: How to write them
11706: @end menu
11707:
11708: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
11709: @subsection @code{Code} and @code{;code}
11710:
11711: Gforth provides some words for defining primitives (words written in
11712: machine code), and for defining the machine-code equivalent of
11713: @code{DOES>}-based defining words. However, the machine-independent
11714: nature of Gforth poses a few problems: First of all, Gforth runs on
11715: several architectures, so it can provide no standard assembler. What's
11716: worse is that the register allocation not only depends on the processor,
11717: but also on the @code{gcc} version and options used.
11718:
11719: The words that Gforth offers encapsulate some system dependences (e.g.,
11720: the header structure), so a system-independent assembler may be used in
11721: Gforth. If you do not have an assembler, you can compile machine code
11722: directly with @code{,} and @code{c,}@footnote{This isn't portable,
11723: because these words emit stuff in @i{data} space; it works because
11724: Gforth has unified code/data spaces. Assembler isn't likely to be
11725: portable anyway.}.
11726:
11727:
11728: doc-assembler
11729: doc-init-asm
11730: doc-code
11731: doc-end-code
11732: doc-;code
11733: doc-flush-icache
11734:
11735:
11736: If @code{flush-icache} does not work correctly, @code{code} words
11737: etc. will not work (reliably), either.
11738:
11739: The typical usage of these @code{code} words can be shown most easily by
11740: analogy to the equivalent high-level defining words:
11741:
11742: @example
11743: : foo code foo
11744: <high-level Forth words> <assembler>
11745: ; end-code
11746:
11747: : bar : bar
11748: <high-level Forth words> <high-level Forth words>
11749: CREATE CREATE
11750: <high-level Forth words> <high-level Forth words>
11751: DOES> ;code
11752: <high-level Forth words> <assembler>
11753: ; end-code
11754: @end example
11755:
11756: @c anton: the following stuff is also in "Common Assembler", in less detail.
11757:
11758: @cindex registers of the inner interpreter
11759: In the assembly code you will want to refer to the inner interpreter's
11760: registers (e.g., the data stack pointer) and you may want to use other
11761: registers for temporary storage. Unfortunately, the register allocation
11762: is installation-dependent.
11763:
11764: In particular, @code{ip} (Forth instruction pointer) and @code{rp}
11765: (return stack pointer) may be in different places in @code{gforth} and
11766: @code{gforth-fast}, or different installations. This means that you
11767: cannot write a @code{NEXT} routine that works reliably on both versions
11768: or different installations; so for doing @code{NEXT}, I recommend
11769: jumping to @code{' noop >code-address}, which contains nothing but a
11770: @code{NEXT}.
11771:
11772: For general accesses to the inner interpreter's registers, the easiest
11773: solution is to use explicit register declarations (@pxref{Explicit Reg
11774: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) for
11775: all of the inner interpreter's registers: You have to compile Gforth
11776: with @code{-DFORCE_REG} (configure option @code{--enable-force-reg}) and
11777: the appropriate declarations must be present in the @code{machine.h}
11778: file (see @code{mips.h} for an example; you can find a full list of all
11779: declarable register symbols with @code{grep register engine.c}). If you
11780: give explicit registers to all variables that are declared at the
11781: beginning of @code{engine()}, you should be able to use the other
11782: caller-saved registers for temporary storage. Alternatively, you can use
11783: the @code{gcc} option @code{-ffixed-REG} (@pxref{Code Gen Options, ,
11784: Options for Code Generation Conventions, gcc.info, GNU C Manual}) to
11785: reserve a register (however, this restriction on register allocation may
11786: slow Gforth significantly).
11787:
11788: If this solution is not viable (e.g., because @code{gcc} does not allow
11789: you to explicitly declare all the registers you need), you have to find
11790: out by looking at the code where the inner interpreter's registers
11791: reside and which registers can be used for temporary storage. You can
11792: get an assembly listing of the engine's code with @code{make engine.s}.
11793:
11794: In any case, it is good practice to abstract your assembly code from the
11795: actual register allocation. E.g., if the data stack pointer resides in
11796: register @code{$17}, create an alias for this register called @code{sp},
11797: and use that in your assembly code.
11798:
11799: @cindex code words, portable
11800: Another option for implementing normal and defining words efficiently
11801: is to add the desired functionality to the source of Gforth. For normal
11802: words you just have to edit @file{primitives} (@pxref{Automatic
11803: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
11804: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
11805: @file{prims2x.fs}, and possibly @file{cross.fs}.
11806:
11807: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
11808: @subsection Common Assembler
11809:
11810: The assemblers in Gforth generally use a postfix syntax, i.e., the
11811: instruction name follows the operands.
11812:
11813: The operands are passed in the usual order (the same that is used in the
11814: manual of the architecture). Since they all are Forth words, they have
11815: to be separated by spaces; you can also use Forth words to compute the
11816: operands.
11817:
11818: The instruction names usually end with a @code{,}. This makes it easier
11819: to visually separate instructions if you put several of them on one
11820: line; it also avoids shadowing other Forth words (e.g., @code{and}).
11821:
11822: Registers are usually specified by number; e.g., (decimal) @code{11}
11823: specifies registers R11 and F11 on the Alpha architecture (which one,
11824: depends on the instruction). The usual names are also available, e.g.,
11825: @code{s2} for R11 on Alpha.
11826:
11827: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
11828: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
11829: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
11830: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
11831: conditions are specified in a way specific to each assembler.
11832:
11833: Note that the register assignments of the Gforth engine can change
11834: between Gforth versions, or even between different compilations of the
11835: same Gforth version (e.g., if you use a different GCC version). So if
11836: you want to refer to Gforth's registers (e.g., the stack pointer or
11837: TOS), I recommend defining your own words for refering to these
11838: registers, and using them later on; then you can easily adapt to a
11839: changed register assignment. The stability of the register assignment
11840: is usually better if you build Gforth with @code{--enable-force-reg}.
11841:
11842: The most common use of these registers is to dispatch to the next word
11843: (the @code{next} routine). A portable way to do this is to jump to
11844: @code{' noop >code-address} (of course, this is less efficient than
11845: integrating the @code{next} code and scheduling it well).
11846:
11847: Another difference between Gforth version is that the top of stack is
11848: kept in memory in @code{gforth} and, on most platforms, in a register in
11849: @code{gforth-fast}.
11850:
11851: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
11852: @subsection Common Disassembler
11853:
11854: You can disassemble a @code{code} word with @code{see}
11855: (@pxref{Debugging}). You can disassemble a section of memory with
11856:
11857: doc-disasm
11858:
11859: The disassembler generally produces output that can be fed into the
11860: assembler (i.e., same syntax, etc.). It also includes additional
11861: information in comments. In particular, the address of the instruction
11862: is given in a comment before the instruction.
11863:
11864: @code{See} may display more or less than the actual code of the word,
11865: because the recognition of the end of the code is unreliable. You can
11866: use @code{disasm} if it did not display enough. It may display more, if
11867: the code word is not immediately followed by a named word. If you have
11868: something else there, you can follow the word with @code{align last @ ,}
11869: to ensure that the end is recognized.
11870:
11871: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
11872: @subsection 386 Assembler
11873:
11874: The 386 assembler included in Gforth was written by Bernd Paysan, it's
11875: available under GPL, and originally part of bigFORTH.
11876:
11877: The 386 disassembler included in Gforth was written by Andrew McKewan
11878: and is in the public domain.
11879:
11880: The disassembler displays code in an Intel-like prefix syntax.
11881:
11882: The assembler uses a postfix syntax with reversed parameters.
11883:
11884: The assembler includes all instruction of the Athlon, i.e. 486 core
11885: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
11886: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
11887: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
11888:
11889: There are several prefixes to switch between different operation sizes,
11890: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
11891: double-word accesses. Addressing modes can be switched with @code{.wa}
11892: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
11893: need a prefix for byte register names (@code{AL} et al).
11894:
11895: For floating point operations, the prefixes are @code{.fs} (IEEE
11896: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
11897: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
11898:
11899: The MMX opcodes don't have size prefixes, they are spelled out like in
11900: the Intel assembler. Instead of move from and to memory, there are
11901: PLDQ/PLDD and PSTQ/PSTD.
11902:
11903: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
11904: ax. Immediate values are indicated by postfixing them with @code{#},
11905: e.g., @code{3 #}. Here are some examples of addressing modes in various
11906: syntaxes:
11907:
11908: @example
11909: Gforth Intel (NASM) AT&T (gas) Name
11910: .w ax ax %ax register (16 bit)
11911: ax eax %eax register (32 bit)
11912: 3 # offset 3 $3 immediate
11913: 1000 #) byte ptr 1000 1000 displacement
11914: bx ) [ebx] (%ebx) base
11915: 100 di d) 100[edi] 100(%edi) base+displacement
11916: 20 ax *4 i#) 20[eax*4] 20(,%eax,4) (index*scale)+displacement
11917: di ax *4 i) [edi][eax*4] (%edi,%eax,4) base+(index*scale)
11918: 4 bx cx di) 4[ebx][ecx] 4(%ebx,%ecx) base+index+displacement
11919: 12 sp ax *2 di) 12[esp][eax*2] 12(%esp,%eax,2) base+(index*scale)+displacement
11920: @end example
11921:
11922: You can use @code{L)} and @code{LI)} instead of @code{D)} and
11923: @code{DI)} to enforce 32-bit displacement fields (useful for
11924: later patching).
11925:
11926: Some example of instructions are:
11927:
11928: @example
11929: ax bx mov \ move ebx,eax
11930: 3 # ax mov \ mov eax,3
11931: 100 di ) ax mov \ mov eax,100[edi]
11932: 4 bx cx di) ax mov \ mov eax,4[ebx][ecx]
11933: .w ax bx mov \ mov bx,ax
11934: @end example
11935:
11936: The following forms are supported for binary instructions:
11937:
11938: @example
11939: <reg> <reg> <inst>
11940: <n> # <reg> <inst>
11941: <mem> <reg> <inst>
11942: <reg> <mem> <inst>
11943: @end example
11944:
11945: Immediate to memory is not supported. The shift/rotate syntax is:
11946:
11947: @example
11948: <reg/mem> 1 # shl \ shortens to shift without immediate
11949: <reg/mem> 4 # shl
11950: <reg/mem> cl shl
11951: @end example
11952:
11953: Precede string instructions (@code{movs} etc.) with @code{.b} to get
11954: the byte version.
11955:
11956: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
11957: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
11958: pc < >= <= >}. (Note that most of these words shadow some Forth words
11959: when @code{assembler} is in front of @code{forth} in the search path,
11960: e.g., in @code{code} words). Currently the control structure words use
11961: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
11962: to shuffle them (you can also use @code{swap} etc.).
11963:
11964: Here is an example of a @code{code} word (assumes that the stack pointer
11965: is in esi and the TOS is in ebx):
11966:
11967: @example
11968: code my+ ( n1 n2 -- n )
11969: 4 si D) bx add
11970: 4 # si add
11971: Next
11972: end-code
11973: @end example
11974:
11975: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
11976: @subsection Alpha Assembler
11977:
11978: The Alpha assembler and disassembler were originally written by Bernd
11979: Thallner.
11980:
11981: The register names @code{a0}--@code{a5} are not available to avoid
11982: shadowing hex numbers.
11983:
11984: Immediate forms of arithmetic instructions are distinguished by a
11985: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
11986: does not count as arithmetic instruction).
11987:
11988: You have to specify all operands to an instruction, even those that
11989: other assemblers consider optional, e.g., the destination register for
11990: @code{br,}, or the destination register and hint for @code{jmp,}.
11991:
11992: You can specify conditions for @code{if,} by removing the first @code{b}
11993: and the trailing @code{,} from a branch with a corresponding name; e.g.,
11994:
11995: @example
11996: 11 fgt if, \ if F11>0e
11997: ...
11998: endif,
11999: @end example
12000:
12001: @code{fbgt,} gives @code{fgt}.
12002:
12003: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
12004: @subsection MIPS assembler
12005:
12006: The MIPS assembler was originally written by Christian Pirker.
12007:
12008: Currently the assembler and disassembler only cover the MIPS-I
12009: architecture (R3000), and don't support FP instructions.
12010:
12011: The register names @code{$a0}--@code{$a3} are not available to avoid
12012: shadowing hex numbers.
12013:
12014: Because there is no way to distinguish registers from immediate values,
12015: you have to explicitly use the immediate forms of instructions, i.e.,
12016: @code{addiu,}, not just @code{addu,} (@command{as} does this
12017: implicitly).
12018:
12019: If the architecture manual specifies several formats for the instruction
12020: (e.g., for @code{jalr,}), you usually have to use the one with more
12021: arguments (i.e., two for @code{jalr,}). When in doubt, see
12022: @code{arch/mips/testasm.fs} for an example of correct use.
12023:
12024: Branches and jumps in the MIPS architecture have a delay slot. You have
12025: to fill it yourself (the simplest way is to use @code{nop,}), the
12026: assembler does not do it for you (unlike @command{as}). Even
12027: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
12028: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
12029: and @code{then,} just specify branch targets, they are not affected.
12030:
12031: Note that you must not put branches, jumps, or @code{li,} into the delay
12032: slot: @code{li,} may expand to several instructions, and control flow
12033: instructions may not be put into the branch delay slot in any case.
12034:
12035: For branches the argument specifying the target is a relative address;
12036: You have to add the address of the delay slot to get the absolute
12037: address.
12038:
12039: The MIPS architecture also has load delay slots and restrictions on
12040: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
12041: yourself to satisfy these restrictions, the assembler does not do it for
12042: you.
12043:
12044: You can specify the conditions for @code{if,} etc. by taking a
12045: conditional branch and leaving away the @code{b} at the start and the
12046: @code{,} at the end. E.g.,
12047:
12048: @example
12049: 4 5 eq if,
12050: ... \ do something if $4 equals $5
12051: then,
12052: @end example
12053:
12054: @node Other assemblers, , MIPS assembler, Assembler and Code Words
12055: @subsection Other assemblers
12056:
12057: If you want to contribute another assembler/disassembler, please contact
12058: us (@email{anton@@mips.complang.tuwien.ac.at}) to check if we have such
12059: an assembler already. If you are writing them from scratch, please use
12060: a similar syntax style as the one we use (i.e., postfix, commas at the
12061: end of the instruction names, @pxref{Common Assembler}); make the output
12062: of the disassembler be valid input for the assembler, and keep the style
12063: similar to the style we used.
12064:
12065: Hints on implementation: The most important part is to have a good test
12066: suite that contains all instructions. Once you have that, the rest is
12067: easy. For actual coding you can take a look at
12068: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
12069: the assembler and disassembler, avoiding redundancy and some potential
12070: bugs. You can also look at that file (and @pxref{Advanced does> usage
12071: example}) to get ideas how to factor a disassembler.
12072:
12073: Start with the disassembler, because it's easier to reuse data from the
12074: disassembler for the assembler than the other way round.
12075:
12076: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
12077: how simple it can be.
12078:
12079: @c -------------------------------------------------------------
12080: @node Threading Words, Passing Commands to the OS, Assembler and Code Words, Words
12081: @section Threading Words
12082: @cindex threading words
12083:
12084: @cindex code address
12085: These words provide access to code addresses and other threading stuff
12086: in Gforth (and, possibly, other interpretive Forths). It more or less
12087: abstracts away the differences between direct and indirect threading
12088: (and, for direct threading, the machine dependences). However, at
12089: present this wordset is still incomplete. It is also pretty low-level;
12090: some day it will hopefully be made unnecessary by an internals wordset
12091: that abstracts implementation details away completely.
12092:
12093: The terminology used here stems from indirect threaded Forth systems; in
12094: such a system, the XT of a word is represented by the CFA (code field
12095: address) of a word; the CFA points to a cell that contains the code
12096: address. The code address is the address of some machine code that
12097: performs the run-time action of invoking the word (e.g., the
12098: @code{dovar:} routine pushes the address of the body of the word (a
12099: variable) on the stack
12100: ).
12101:
12102: @cindex code address
12103: @cindex code field address
12104: In an indirect threaded Forth, you can get the code address of @i{name}
12105: with @code{' @i{name} @@}; in Gforth you can get it with @code{' @i{name}
12106: >code-address}, independent of the threading method.
12107:
12108: doc-threading-method
12109: doc->code-address
12110: doc-code-address!
12111:
12112: @cindex @code{does>}-handler
12113: @cindex @code{does>}-code
12114: For a word defined with @code{DOES>}, the code address usually points to
12115: a jump instruction (the @dfn{does-handler}) that jumps to the dodoes
12116: routine (in Gforth on some platforms, it can also point to the dodoes
12117: routine itself). What you are typically interested in, though, is
12118: whether a word is a @code{DOES>}-defined word, and what Forth code it
12119: executes; @code{>does-code} tells you that.
12120:
12121: doc->does-code
12122:
12123: To create a @code{DOES>}-defined word with the following basic words,
12124: you have to set up a @code{DOES>}-handler with @code{does-handler!};
12125: @code{/does-handler} aus behind you have to place your executable Forth
12126: code. Finally you have to create a word and modify its behaviour with
12127: @code{does-handler!}.
12128:
12129: doc-does-code!
12130: doc-does-handler!
12131: doc-/does-handler
12132:
12133: The code addresses produced by various defining words are produced by
12134: the following words:
12135:
12136: doc-docol:
12137: doc-docon:
12138: doc-dovar:
12139: doc-douser:
12140: doc-dodefer:
12141: doc-dofield:
12142:
12143: @cindex definer
12144: The following two words generalize @code{>code-address},
12145: @code{>does-code}, @code{code-address!}, and @code{does-code!}:
12146:
12147: doc->definer
12148: doc-definer!
12149:
12150: @c -------------------------------------------------------------
12151: @node Passing Commands to the OS, Keeping track of Time, Threading Words, Words
12152: @section Passing Commands to the Operating System
12153: @cindex operating system - passing commands
12154: @cindex shell commands
12155:
12156: Gforth allows you to pass an arbitrary string to the host operating
12157: system shell (if such a thing exists) for execution.
12158:
12159:
12160: doc-sh
12161: doc-system
12162: doc-$?
12163: doc-getenv
12164:
12165:
12166: @c -------------------------------------------------------------
12167: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
12168: @section Keeping track of Time
12169: @cindex time-related words
12170:
12171: doc-ms
12172: doc-time&date
12173: doc-utime
12174: doc-cputime
12175:
12176:
12177: @c -------------------------------------------------------------
12178: @node Miscellaneous Words, , Keeping track of Time, Words
12179: @section Miscellaneous Words
12180: @cindex miscellaneous words
12181:
12182: @comment TODO find homes for these
12183:
12184: These section lists the ANS Forth words that are not documented
12185: elsewhere in this manual. Ultimately, they all need proper homes.
12186:
12187: doc-quit
12188:
12189: The following ANS Forth words are not currently supported by Gforth
12190: (@pxref{ANS conformance}):
12191:
12192: @code{EDITOR}
12193: @code{EMIT?}
12194: @code{FORGET}
12195:
12196: @c ******************************************************************
12197: @node Error messages, Tools, Words, Top
12198: @chapter Error messages
12199: @cindex error messages
12200: @cindex backtrace
12201:
12202: A typical Gforth error message looks like this:
12203:
12204: @example
12205: in file included from \evaluated string/:-1
12206: in file included from ./yyy.fs:1
12207: ./xxx.fs:4: Invalid memory address
12208: bar
12209: ^^^
12210: Backtrace:
12211: $400E664C @@
12212: $400E6664 foo
12213: @end example
12214:
12215: The message identifying the error is @code{Invalid memory address}. The
12216: error happened when text-interpreting line 4 of the file
12217: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
12218: word on the line where the error happened, is pointed out (with
12219: @code{^^^}).
12220:
12221: The file containing the error was included in line 1 of @file{./yyy.fs},
12222: and @file{yyy.fs} was included from a non-file (in this case, by giving
12223: @file{yyy.fs} as command-line parameter to Gforth).
12224:
12225: At the end of the error message you find a return stack dump that can be
12226: interpreted as a backtrace (possibly empty). On top you find the top of
12227: the return stack when the @code{throw} happened, and at the bottom you
12228: find the return stack entry just above the return stack of the topmost
12229: text interpreter.
12230:
12231: To the right of most return stack entries you see a guess for the word
12232: that pushed that return stack entry as its return address. This gives a
12233: backtrace. In our case we see that @code{bar} called @code{foo}, and
12234: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
12235: address} exception).
12236:
12237: Note that the backtrace is not perfect: We don't know which return stack
12238: entries are return addresses (so we may get false positives); and in
12239: some cases (e.g., for @code{abort"}) we cannot determine from the return
12240: address the word that pushed the return address, so for some return
12241: addresses you see no names in the return stack dump.
12242:
12243: @cindex @code{catch} and backtraces
12244: The return stack dump represents the return stack at the time when a
12245: specific @code{throw} was executed. In programs that make use of
12246: @code{catch}, it is not necessarily clear which @code{throw} should be
12247: used for the return stack dump (e.g., consider one @code{throw} that
12248: indicates an error, which is caught, and during recovery another error
12249: happens; which @code{throw} should be used for the stack dump?). Gforth
12250: presents the return stack dump for the first @code{throw} after the last
12251: executed (not returned-to) @code{catch}; this works well in the usual
12252: case.
12253:
12254: @cindex @code{gforth-fast} and backtraces
12255: @cindex @code{gforth-fast}, difference from @code{gforth}
12256: @cindex backtraces with @code{gforth-fast}
12257: @cindex return stack dump with @code{gforth-fast}
12258: @code{Gforth} is able to do a return stack dump for throws generated
12259: from primitives (e.g., invalid memory address, stack empty etc.);
12260: @code{gforth-fast} is only able to do a return stack dump from a
12261: directly called @code{throw} (including @code{abort} etc.). Given an
12262: exception caused by a primitive in @code{gforth-fast}, you will
12263: typically see no return stack dump at all; however, if the exception is
12264: caught by @code{catch} (e.g., for restoring some state), and then
12265: @code{throw}n again, the return stack dump will be for the first such
12266: @code{throw}.
12267:
12268: @c ******************************************************************
12269: @node Tools, ANS conformance, Error messages, Top
12270: @chapter Tools
12271:
12272: @menu
12273: * ANS Report:: Report the words used, sorted by wordset.
12274: @end menu
12275:
12276: See also @ref{Emacs and Gforth}.
12277:
12278: @node ANS Report, , Tools, Tools
12279: @section @file{ans-report.fs}: Report the words used, sorted by wordset
12280: @cindex @file{ans-report.fs}
12281: @cindex report the words used in your program
12282: @cindex words used in your program
12283:
12284: If you want to label a Forth program as ANS Forth Program, you must
12285: document which wordsets the program uses; for extension wordsets, it is
12286: helpful to list the words the program requires from these wordsets
12287: (because Forth systems are allowed to provide only some words of them).
12288:
12289: The @file{ans-report.fs} tool makes it easy for you to determine which
12290: words from which wordset and which non-ANS words your application
12291: uses. You simply have to include @file{ans-report.fs} before loading the
12292: program you want to check. After loading your program, you can get the
12293: report with @code{print-ans-report}. A typical use is to run this as
12294: batch job like this:
12295: @example
12296: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
12297: @end example
12298:
12299: The output looks like this (for @file{compat/control.fs}):
12300: @example
12301: The program uses the following words
12302: from CORE :
12303: : POSTPONE THEN ; immediate ?dup IF 0=
12304: from BLOCK-EXT :
12305: \
12306: from FILE :
12307: (
12308: @end example
12309:
12310: @subsection Caveats
12311:
12312: Note that @file{ans-report.fs} just checks which words are used, not whether
12313: they are used in an ANS Forth conforming way!
12314:
12315: Some words are defined in several wordsets in the
12316: standard. @file{ans-report.fs} reports them for only one of the
12317: wordsets, and not necessarily the one you expect. It depends on usage
12318: which wordset is the right one to specify. E.g., if you only use the
12319: compilation semantics of @code{S"}, it is a Core word; if you also use
12320: its interpretation semantics, it is a File word.
12321:
12322: @c ******************************************************************
12323: @node ANS conformance, Standard vs Extensions, Tools, Top
12324: @chapter ANS conformance
12325: @cindex ANS conformance of Gforth
12326:
12327: To the best of our knowledge, Gforth is an
12328:
12329: ANS Forth System
12330: @itemize @bullet
12331: @item providing the Core Extensions word set
12332: @item providing the Block word set
12333: @item providing the Block Extensions word set
12334: @item providing the Double-Number word set
12335: @item providing the Double-Number Extensions word set
12336: @item providing the Exception word set
12337: @item providing the Exception Extensions word set
12338: @item providing the Facility word set
12339: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
12340: @item providing the File Access word set
12341: @item providing the File Access Extensions word set
12342: @item providing the Floating-Point word set
12343: @item providing the Floating-Point Extensions word set
12344: @item providing the Locals word set
12345: @item providing the Locals Extensions word set
12346: @item providing the Memory-Allocation word set
12347: @item providing the Memory-Allocation Extensions word set (that one's easy)
12348: @item providing the Programming-Tools word set
12349: @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
12350: @item providing the Search-Order word set
12351: @item providing the Search-Order Extensions word set
12352: @item providing the String word set
12353: @item providing the String Extensions word set (another easy one)
12354: @end itemize
12355:
12356: @cindex system documentation
12357: In addition, ANS Forth systems are required to document certain
12358: implementation choices. This chapter tries to meet these
12359: requirements. In many cases it gives a way to ask the system for the
12360: information instead of providing the information directly, in
12361: particular, if the information depends on the processor, the operating
12362: system or the installation options chosen, or if they are likely to
12363: change during the maintenance of Gforth.
12364:
12365: @comment The framework for the rest has been taken from pfe.
12366:
12367: @menu
12368: * The Core Words::
12369: * The optional Block word set::
12370: * The optional Double Number word set::
12371: * The optional Exception word set::
12372: * The optional Facility word set::
12373: * The optional File-Access word set::
12374: * The optional Floating-Point word set::
12375: * The optional Locals word set::
12376: * The optional Memory-Allocation word set::
12377: * The optional Programming-Tools word set::
12378: * The optional Search-Order word set::
12379: @end menu
12380:
12381:
12382: @c =====================================================================
12383: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
12384: @comment node-name, next, previous, up
12385: @section The Core Words
12386: @c =====================================================================
12387: @cindex core words, system documentation
12388: @cindex system documentation, core words
12389:
12390: @menu
12391: * core-idef:: Implementation Defined Options
12392: * core-ambcond:: Ambiguous Conditions
12393: * core-other:: Other System Documentation
12394: @end menu
12395:
12396: @c ---------------------------------------------------------------------
12397: @node core-idef, core-ambcond, The Core Words, The Core Words
12398: @subsection Implementation Defined Options
12399: @c ---------------------------------------------------------------------
12400: @cindex core words, implementation-defined options
12401: @cindex implementation-defined options, core words
12402:
12403:
12404: @table @i
12405: @item (Cell) aligned addresses:
12406: @cindex cell-aligned addresses
12407: @cindex aligned addresses
12408: processor-dependent. Gforth's alignment words perform natural alignment
12409: (e.g., an address aligned for a datum of size 8 is divisible by
12410: 8). Unaligned accesses usually result in a @code{-23 THROW}.
12411:
12412: @item @code{EMIT} and non-graphic characters:
12413: @cindex @code{EMIT} and non-graphic characters
12414: @cindex non-graphic characters and @code{EMIT}
12415: The character is output using the C library function (actually, macro)
12416: @code{putc}.
12417:
12418: @item character editing of @code{ACCEPT} and @code{EXPECT}:
12419: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
12420: @cindex editing in @code{ACCEPT} and @code{EXPECT}
12421: @cindex @code{ACCEPT}, editing
12422: @cindex @code{EXPECT}, editing
12423: This is modeled on the GNU readline library (@pxref{Readline
12424: Interaction, , Command Line Editing, readline, The GNU Readline
12425: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
12426: producing a full word completion every time you type it (instead of
12427: producing the common prefix of all completions). @xref{Command-line editing}.
12428:
12429: @item character set:
12430: @cindex character set
12431: The character set of your computer and display device. Gforth is
12432: 8-bit-clean (but some other component in your system may make trouble).
12433:
12434: @item Character-aligned address requirements:
12435: @cindex character-aligned address requirements
12436: installation-dependent. Currently a character is represented by a C
12437: @code{unsigned char}; in the future we might switch to @code{wchar_t}
12438: (Comments on that requested).
12439:
12440: @item character-set extensions and matching of names:
12441: @cindex character-set extensions and matching of names
12442: @cindex case-sensitivity for name lookup
12443: @cindex name lookup, case-sensitivity
12444: @cindex locale and case-sensitivity
12445: Any character except the ASCII NUL character can be used in a
12446: name. Matching is case-insensitive (except in @code{TABLE}s). The
12447: matching is performed using the C library function @code{strncasecmp}, whose
12448: function is probably influenced by the locale. E.g., the @code{C} locale
12449: does not know about accents and umlauts, so they are matched
12450: case-sensitively in that locale. For portability reasons it is best to
12451: write programs such that they work in the @code{C} locale. Then one can
12452: use libraries written by a Polish programmer (who might use words
12453: containing ISO Latin-2 encoded characters) and by a French programmer
12454: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
12455: funny results for some of the words (which ones, depends on the font you
12456: are using)). Also, the locale you prefer may not be available in other
12457: operating systems. Hopefully, Unicode will solve these problems one day.
12458:
12459: @item conditions under which control characters match a space delimiter:
12460: @cindex space delimiters
12461: @cindex control characters as delimiters
12462: If @code{WORD} is called with the space character as a delimiter, all
12463: white-space characters (as identified by the C macro @code{isspace()})
12464: are delimiters. @code{PARSE}, on the other hand, treats space like other
12465: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
12466: like @code{PARSE} otherwise. @code{Name}, which is used by the outer
12467: interpreter (aka text interpreter) by default, treats all white-space
12468: characters as delimiters.
12469:
12470: @item format of the control-flow stack:
12471: @cindex control-flow stack, format
12472: The data stack is used as control-flow stack. The size of a control-flow
12473: stack item in cells is given by the constant @code{cs-item-size}. At the
12474: time of this writing, an item consists of a (pointer to a) locals list
12475: (third), an address in the code (second), and a tag for identifying the
12476: item (TOS). The following tags are used: @code{defstart},
12477: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
12478: @code{scopestart}.
12479:
12480: @item conversion of digits > 35
12481: @cindex digits > 35
12482: The characters @code{[\]^_'} are the digits with the decimal value
12483: 36@minus{}41. There is no way to input many of the larger digits.
12484:
12485: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
12486: @cindex @code{EXPECT}, display after end of input
12487: @cindex @code{ACCEPT}, display after end of input
12488: The cursor is moved to the end of the entered string. If the input is
12489: terminated using the @kbd{Return} key, a space is typed.
12490:
12491: @item exception abort sequence of @code{ABORT"}:
12492: @cindex exception abort sequence of @code{ABORT"}
12493: @cindex @code{ABORT"}, exception abort sequence
12494: The error string is stored into the variable @code{"error} and a
12495: @code{-2 throw} is performed.
12496:
12497: @item input line terminator:
12498: @cindex input line terminator
12499: @cindex line terminator on input
12500: @cindex newline character on input
12501: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
12502: lines. One of these characters is typically produced when you type the
12503: @kbd{Enter} or @kbd{Return} key.
12504:
12505: @item maximum size of a counted string:
12506: @cindex maximum size of a counted string
12507: @cindex counted string, maximum size
12508: @code{s" /counted-string" environment? drop .}. Currently 255 characters
12509: on all platforms, but this may change.
12510:
12511: @item maximum size of a parsed string:
12512: @cindex maximum size of a parsed string
12513: @cindex parsed string, maximum size
12514: Given by the constant @code{/line}. Currently 255 characters.
12515:
12516: @item maximum size of a definition name, in characters:
12517: @cindex maximum size of a definition name, in characters
12518: @cindex name, maximum length
12519: 31
12520:
12521: @item maximum string length for @code{ENVIRONMENT?}, in characters:
12522: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
12523: @cindex @code{ENVIRONMENT?} string length, maximum
12524: 31
12525:
12526: @item method of selecting the user input device:
12527: @cindex user input device, method of selecting
12528: The user input device is the standard input. There is currently no way to
12529: change it from within Gforth. However, the input can typically be
12530: redirected in the command line that starts Gforth.
12531:
12532: @item method of selecting the user output device:
12533: @cindex user output device, method of selecting
12534: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
12535: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
12536: output when the user output device is a terminal, otherwise the output
12537: is buffered.
12538:
12539: @item methods of dictionary compilation:
12540: What are we expected to document here?
12541:
12542: @item number of bits in one address unit:
12543: @cindex number of bits in one address unit
12544: @cindex address unit, size in bits
12545: @code{s" address-units-bits" environment? drop .}. 8 in all current
12546: platforms.
12547:
12548: @item number representation and arithmetic:
12549: @cindex number representation and arithmetic
12550: Processor-dependent. Binary two's complement on all current platforms.
12551:
12552: @item ranges for integer types:
12553: @cindex ranges for integer types
12554: @cindex integer types, ranges
12555: Installation-dependent. Make environmental queries for @code{MAX-N},
12556: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
12557: unsigned (and positive) types is 0. The lower bound for signed types on
12558: two's complement and one's complement machines machines can be computed
12559: by adding 1 to the upper bound.
12560:
12561: @item read-only data space regions:
12562: @cindex read-only data space regions
12563: @cindex data-space, read-only regions
12564: The whole Forth data space is writable.
12565:
12566: @item size of buffer at @code{WORD}:
12567: @cindex size of buffer at @code{WORD}
12568: @cindex @code{WORD} buffer size
12569: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12570: shared with the pictured numeric output string. If overwriting
12571: @code{PAD} is acceptable, it is as large as the remaining dictionary
12572: space, although only as much can be sensibly used as fits in a counted
12573: string.
12574:
12575: @item size of one cell in address units:
12576: @cindex cell size
12577: @code{1 cells .}.
12578:
12579: @item size of one character in address units:
12580: @cindex char size
12581: @code{1 chars .}. 1 on all current platforms.
12582:
12583: @item size of the keyboard terminal buffer:
12584: @cindex size of the keyboard terminal buffer
12585: @cindex terminal buffer, size
12586: Varies. You can determine the size at a specific time using @code{lp@@
12587: tib - .}. It is shared with the locals stack and TIBs of files that
12588: include the current file. You can change the amount of space for TIBs
12589: and locals stack at Gforth startup with the command line option
12590: @code{-l}.
12591:
12592: @item size of the pictured numeric output buffer:
12593: @cindex size of the pictured numeric output buffer
12594: @cindex pictured numeric output buffer, size
12595: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12596: shared with @code{WORD}.
12597:
12598: @item size of the scratch area returned by @code{PAD}:
12599: @cindex size of the scratch area returned by @code{PAD}
12600: @cindex @code{PAD} size
12601: The remainder of dictionary space. @code{unused pad here - - .}.
12602:
12603: @item system case-sensitivity characteristics:
12604: @cindex case-sensitivity characteristics
12605: Dictionary searches are case-insensitive (except in
12606: @code{TABLE}s). However, as explained above under @i{character-set
12607: extensions}, the matching for non-ASCII characters is determined by the
12608: locale you are using. In the default @code{C} locale all non-ASCII
12609: characters are matched case-sensitively.
12610:
12611: @item system prompt:
12612: @cindex system prompt
12613: @cindex prompt
12614: @code{ ok} in interpret state, @code{ compiled} in compile state.
12615:
12616: @item division rounding:
12617: @cindex division rounding
12618: installation dependent. @code{s" floored" environment? drop .}. We leave
12619: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
12620: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
12621:
12622: @item values of @code{STATE} when true:
12623: @cindex @code{STATE} values
12624: -1.
12625:
12626: @item values returned after arithmetic overflow:
12627: On two's complement machines, arithmetic is performed modulo
12628: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12629: arithmetic (with appropriate mapping for signed types). Division by zero
12630: typically results in a @code{-55 throw} (Floating-point unidentified
12631: fault) or @code{-10 throw} (divide by zero).
12632:
12633: @item whether the current definition can be found after @t{DOES>}:
12634: @cindex @t{DOES>}, visibility of current definition
12635: No.
12636:
12637: @end table
12638:
12639: @c ---------------------------------------------------------------------
12640: @node core-ambcond, core-other, core-idef, The Core Words
12641: @subsection Ambiguous conditions
12642: @c ---------------------------------------------------------------------
12643: @cindex core words, ambiguous conditions
12644: @cindex ambiguous conditions, core words
12645:
12646: @table @i
12647:
12648: @item a name is neither a word nor a number:
12649: @cindex name not found
12650: @cindex undefined word
12651: @code{-13 throw} (Undefined word).
12652:
12653: @item a definition name exceeds the maximum length allowed:
12654: @cindex word name too long
12655: @code{-19 throw} (Word name too long)
12656:
12657: @item addressing a region not inside the various data spaces of the forth system:
12658: @cindex Invalid memory address
12659: The stacks, code space and header space are accessible. Machine code space is
12660: typically readable. Accessing other addresses gives results dependent on
12661: the operating system. On decent systems: @code{-9 throw} (Invalid memory
12662: address).
12663:
12664: @item argument type incompatible with parameter:
12665: @cindex argument type mismatch
12666: This is usually not caught. Some words perform checks, e.g., the control
12667: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
12668: mismatch).
12669:
12670: @item attempting to obtain the execution token of a word with undefined execution semantics:
12671: @cindex Interpreting a compile-only word, for @code{'} etc.
12672: @cindex execution token of words with undefined execution semantics
12673: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
12674: get an execution token for @code{compile-only-error} (which performs a
12675: @code{-14 throw} when executed).
12676:
12677: @item dividing by zero:
12678: @cindex dividing by zero
12679: @cindex floating point unidentified fault, integer division
12680: On some platforms, this produces a @code{-10 throw} (Division by
12681: zero); on other systems, this typically results in a @code{-55 throw}
12682: (Floating-point unidentified fault).
12683:
12684: @item insufficient data stack or return stack space:
12685: @cindex insufficient data stack or return stack space
12686: @cindex stack overflow
12687: @cindex address alignment exception, stack overflow
12688: @cindex Invalid memory address, stack overflow
12689: Depending on the operating system, the installation, and the invocation
12690: of Gforth, this is either checked by the memory management hardware, or
12691: it is not checked. If it is checked, you typically get a @code{-3 throw}
12692: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
12693: throw} (Invalid memory address) (depending on the platform and how you
12694: achieved the overflow) as soon as the overflow happens. If it is not
12695: checked, overflows typically result in mysterious illegal memory
12696: accesses, producing @code{-9 throw} (Invalid memory address) or
12697: @code{-23 throw} (Address alignment exception); they might also destroy
12698: the internal data structure of @code{ALLOCATE} and friends, resulting in
12699: various errors in these words.
12700:
12701: @item insufficient space for loop control parameters:
12702: @cindex insufficient space for loop control parameters
12703: Like other return stack overflows.
12704:
12705: @item insufficient space in the dictionary:
12706: @cindex insufficient space in the dictionary
12707: @cindex dictionary overflow
12708: If you try to allot (either directly with @code{allot}, or indirectly
12709: with @code{,}, @code{create} etc.) more memory than available in the
12710: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
12711: to access memory beyond the end of the dictionary, the results are
12712: similar to stack overflows.
12713:
12714: @item interpreting a word with undefined interpretation semantics:
12715: @cindex interpreting a word with undefined interpretation semantics
12716: @cindex Interpreting a compile-only word
12717: For some words, we have defined interpretation semantics. For the
12718: others: @code{-14 throw} (Interpreting a compile-only word).
12719:
12720: @item modifying the contents of the input buffer or a string literal:
12721: @cindex modifying the contents of the input buffer or a string literal
12722: These are located in writable memory and can be modified.
12723:
12724: @item overflow of the pictured numeric output string:
12725: @cindex overflow of the pictured numeric output string
12726: @cindex pictured numeric output string, overflow
12727: @code{-17 throw} (Pictured numeric ouput string overflow).
12728:
12729: @item parsed string overflow:
12730: @cindex parsed string overflow
12731: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
12732:
12733: @item producing a result out of range:
12734: @cindex result out of range
12735: On two's complement machines, arithmetic is performed modulo
12736: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12737: arithmetic (with appropriate mapping for signed types). Division by zero
12738: typically results in a @code{-10 throw} (divide by zero) or @code{-55
12739: throw} (floating point unidentified fault). @code{convert} and
12740: @code{>number} currently overflow silently.
12741:
12742: @item reading from an empty data or return stack:
12743: @cindex stack empty
12744: @cindex stack underflow
12745: @cindex return stack underflow
12746: The data stack is checked by the outer (aka text) interpreter after
12747: every word executed. If it has underflowed, a @code{-4 throw} (Stack
12748: underflow) is performed. Apart from that, stacks may be checked or not,
12749: depending on operating system, installation, and invocation. If they are
12750: caught by a check, they typically result in @code{-4 throw} (Stack
12751: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
12752: (Invalid memory address), depending on the platform and which stack
12753: underflows and by how much. Note that even if the system uses checking
12754: (through the MMU), your program may have to underflow by a significant
12755: number of stack items to trigger the reaction (the reason for this is
12756: that the MMU, and therefore the checking, works with a page-size
12757: granularity). If there is no checking, the symptoms resulting from an
12758: underflow are similar to those from an overflow. Unbalanced return
12759: stack errors can result in a variety of symptoms, including @code{-9 throw}
12760: (Invalid memory address) and Illegal Instruction (typically @code{-260
12761: throw}).
12762:
12763: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
12764: @cindex unexpected end of the input buffer
12765: @cindex zero-length string as a name
12766: @cindex Attempt to use zero-length string as a name
12767: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
12768: use zero-length string as a name). Words like @code{'} probably will not
12769: find what they search. Note that it is possible to create zero-length
12770: names with @code{nextname} (should it not?).
12771:
12772: @item @code{>IN} greater than input buffer:
12773: @cindex @code{>IN} greater than input buffer
12774: The next invocation of a parsing word returns a string with length 0.
12775:
12776: @item @code{RECURSE} appears after @code{DOES>}:
12777: @cindex @code{RECURSE} appears after @code{DOES>}
12778: Compiles a recursive call to the defining word, not to the defined word.
12779:
12780: @item argument input source different than current input source for @code{RESTORE-INPUT}:
12781: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
12782: @cindex argument type mismatch, @code{RESTORE-INPUT}
12783: @cindex @code{RESTORE-INPUT}, Argument type mismatch
12784: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
12785: the end of the file was reached), its source-id may be
12786: reused. Therefore, restoring an input source specification referencing a
12787: closed file may lead to unpredictable results instead of a @code{-12
12788: THROW}.
12789:
12790: In the future, Gforth may be able to restore input source specifications
12791: from other than the current input source.
12792:
12793: @item data space containing definitions gets de-allocated:
12794: @cindex data space containing definitions gets de-allocated
12795: Deallocation with @code{allot} is not checked. This typically results in
12796: memory access faults or execution of illegal instructions.
12797:
12798: @item data space read/write with incorrect alignment:
12799: @cindex data space read/write with incorrect alignment
12800: @cindex alignment faults
12801: @cindex address alignment exception
12802: Processor-dependent. Typically results in a @code{-23 throw} (Address
12803: alignment exception). Under Linux-Intel on a 486 or later processor with
12804: alignment turned on, incorrect alignment results in a @code{-9 throw}
12805: (Invalid memory address). There are reportedly some processors with
12806: alignment restrictions that do not report violations.
12807:
12808: @item data space pointer not properly aligned, @code{,}, @code{C,}:
12809: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
12810: Like other alignment errors.
12811:
12812: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
12813: Like other stack underflows.
12814:
12815: @item loop control parameters not available:
12816: @cindex loop control parameters not available
12817: Not checked. The counted loop words simply assume that the top of return
12818: stack items are loop control parameters and behave accordingly.
12819:
12820: @item most recent definition does not have a name (@code{IMMEDIATE}):
12821: @cindex most recent definition does not have a name (@code{IMMEDIATE})
12822: @cindex last word was headerless
12823: @code{abort" last word was headerless"}.
12824:
12825: @item name not defined by @code{VALUE} used by @code{TO}:
12826: @cindex name not defined by @code{VALUE} used by @code{TO}
12827: @cindex @code{TO} on non-@code{VALUE}s
12828: @cindex Invalid name argument, @code{TO}
12829: @code{-32 throw} (Invalid name argument) (unless name is a local or was
12830: defined by @code{CONSTANT}; in the latter case it just changes the constant).
12831:
12832: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
12833: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
12834: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
12835: @code{-13 throw} (Undefined word)
12836:
12837: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
12838: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
12839: Gforth behaves as if they were of the same type. I.e., you can predict
12840: the behaviour by interpreting all parameters as, e.g., signed.
12841:
12842: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
12843: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
12844: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
12845: compilation semantics of @code{TO}.
12846:
12847: @item String longer than a counted string returned by @code{WORD}:
12848: @cindex string longer than a counted string returned by @code{WORD}
12849: @cindex @code{WORD}, string overflow
12850: Not checked. The string will be ok, but the count will, of course,
12851: contain only the least significant bits of the length.
12852:
12853: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
12854: @cindex @code{LSHIFT}, large shift counts
12855: @cindex @code{RSHIFT}, large shift counts
12856: Processor-dependent. Typical behaviours are returning 0 and using only
12857: the low bits of the shift count.
12858:
12859: @item word not defined via @code{CREATE}:
12860: @cindex @code{>BODY} of non-@code{CREATE}d words
12861: @code{>BODY} produces the PFA of the word no matter how it was defined.
12862:
12863: @cindex @code{DOES>} of non-@code{CREATE}d words
12864: @code{DOES>} changes the execution semantics of the last defined word no
12865: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
12866: @code{CREATE , DOES>}.
12867:
12868: @item words improperly used outside @code{<#} and @code{#>}:
12869: Not checked. As usual, you can expect memory faults.
12870:
12871: @end table
12872:
12873:
12874: @c ---------------------------------------------------------------------
12875: @node core-other, , core-ambcond, The Core Words
12876: @subsection Other system documentation
12877: @c ---------------------------------------------------------------------
12878: @cindex other system documentation, core words
12879: @cindex core words, other system documentation
12880:
12881: @table @i
12882: @item nonstandard words using @code{PAD}:
12883: @cindex @code{PAD} use by nonstandard words
12884: None.
12885:
12886: @item operator's terminal facilities available:
12887: @cindex operator's terminal facilities available
12888: After processing the OS's command line, Gforth goes into interactive mode,
12889: and you can give commands to Gforth interactively. The actual facilities
12890: available depend on how you invoke Gforth.
12891:
12892: @item program data space available:
12893: @cindex program data space available
12894: @cindex data space available
12895: @code{UNUSED .} gives the remaining dictionary space. The total
12896: dictionary space can be specified with the @code{-m} switch
12897: (@pxref{Invoking Gforth}) when Gforth starts up.
12898:
12899: @item return stack space available:
12900: @cindex return stack space available
12901: You can compute the total return stack space in cells with
12902: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
12903: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
12904:
12905: @item stack space available:
12906: @cindex stack space available
12907: You can compute the total data stack space in cells with
12908: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
12909: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
12910:
12911: @item system dictionary space required, in address units:
12912: @cindex system dictionary space required, in address units
12913: Type @code{here forthstart - .} after startup. At the time of this
12914: writing, this gives 80080 (bytes) on a 32-bit system.
12915: @end table
12916:
12917:
12918: @c =====================================================================
12919: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
12920: @section The optional Block word set
12921: @c =====================================================================
12922: @cindex system documentation, block words
12923: @cindex block words, system documentation
12924:
12925: @menu
12926: * block-idef:: Implementation Defined Options
12927: * block-ambcond:: Ambiguous Conditions
12928: * block-other:: Other System Documentation
12929: @end menu
12930:
12931:
12932: @c ---------------------------------------------------------------------
12933: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
12934: @subsection Implementation Defined Options
12935: @c ---------------------------------------------------------------------
12936: @cindex implementation-defined options, block words
12937: @cindex block words, implementation-defined options
12938:
12939: @table @i
12940: @item the format for display by @code{LIST}:
12941: @cindex @code{LIST} display format
12942: First the screen number is displayed, then 16 lines of 64 characters,
12943: each line preceded by the line number.
12944:
12945: @item the length of a line affected by @code{\}:
12946: @cindex length of a line affected by @code{\}
12947: @cindex @code{\}, line length in blocks
12948: 64 characters.
12949: @end table
12950:
12951:
12952: @c ---------------------------------------------------------------------
12953: @node block-ambcond, block-other, block-idef, The optional Block word set
12954: @subsection Ambiguous conditions
12955: @c ---------------------------------------------------------------------
12956: @cindex block words, ambiguous conditions
12957: @cindex ambiguous conditions, block words
12958:
12959: @table @i
12960: @item correct block read was not possible:
12961: @cindex block read not possible
12962: Typically results in a @code{throw} of some OS-derived value (between
12963: -512 and -2048). If the blocks file was just not long enough, blanks are
12964: supplied for the missing portion.
12965:
12966: @item I/O exception in block transfer:
12967: @cindex I/O exception in block transfer
12968: @cindex block transfer, I/O exception
12969: Typically results in a @code{throw} of some OS-derived value (between
12970: -512 and -2048).
12971:
12972: @item invalid block number:
12973: @cindex invalid block number
12974: @cindex block number invalid
12975: @code{-35 throw} (Invalid block number)
12976:
12977: @item a program directly alters the contents of @code{BLK}:
12978: @cindex @code{BLK}, altering @code{BLK}
12979: The input stream is switched to that other block, at the same
12980: position. If the storing to @code{BLK} happens when interpreting
12981: non-block input, the system will get quite confused when the block ends.
12982:
12983: @item no current block buffer for @code{UPDATE}:
12984: @cindex @code{UPDATE}, no current block buffer
12985: @code{UPDATE} has no effect.
12986:
12987: @end table
12988:
12989: @c ---------------------------------------------------------------------
12990: @node block-other, , block-ambcond, The optional Block word set
12991: @subsection Other system documentation
12992: @c ---------------------------------------------------------------------
12993: @cindex other system documentation, block words
12994: @cindex block words, other system documentation
12995:
12996: @table @i
12997: @item any restrictions a multiprogramming system places on the use of buffer addresses:
12998: No restrictions (yet).
12999:
13000: @item the number of blocks available for source and data:
13001: depends on your disk space.
13002:
13003: @end table
13004:
13005:
13006: @c =====================================================================
13007: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
13008: @section The optional Double Number word set
13009: @c =====================================================================
13010: @cindex system documentation, double words
13011: @cindex double words, system documentation
13012:
13013: @menu
13014: * double-ambcond:: Ambiguous Conditions
13015: @end menu
13016:
13017:
13018: @c ---------------------------------------------------------------------
13019: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
13020: @subsection Ambiguous conditions
13021: @c ---------------------------------------------------------------------
13022: @cindex double words, ambiguous conditions
13023: @cindex ambiguous conditions, double words
13024:
13025: @table @i
13026: @item @i{d} outside of range of @i{n} in @code{D>S}:
13027: @cindex @code{D>S}, @i{d} out of range of @i{n}
13028: The least significant cell of @i{d} is produced.
13029:
13030: @end table
13031:
13032:
13033: @c =====================================================================
13034: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
13035: @section The optional Exception word set
13036: @c =====================================================================
13037: @cindex system documentation, exception words
13038: @cindex exception words, system documentation
13039:
13040: @menu
13041: * exception-idef:: Implementation Defined Options
13042: @end menu
13043:
13044:
13045: @c ---------------------------------------------------------------------
13046: @node exception-idef, , The optional Exception word set, The optional Exception word set
13047: @subsection Implementation Defined Options
13048: @c ---------------------------------------------------------------------
13049: @cindex implementation-defined options, exception words
13050: @cindex exception words, implementation-defined options
13051:
13052: @table @i
13053: @item @code{THROW}-codes used in the system:
13054: @cindex @code{THROW}-codes used in the system
13055: The codes -256@minus{}-511 are used for reporting signals. The mapping
13056: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
13057: codes -512@minus{}-2047 are used for OS errors (for file and memory
13058: allocation operations). The mapping from OS error numbers to throw codes
13059: is -512@minus{}@code{errno}. One side effect of this mapping is that
13060: undefined OS errors produce a message with a strange number; e.g.,
13061: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
13062: @end table
13063:
13064: @c =====================================================================
13065: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
13066: @section The optional Facility word set
13067: @c =====================================================================
13068: @cindex system documentation, facility words
13069: @cindex facility words, system documentation
13070:
13071: @menu
13072: * facility-idef:: Implementation Defined Options
13073: * facility-ambcond:: Ambiguous Conditions
13074: @end menu
13075:
13076:
13077: @c ---------------------------------------------------------------------
13078: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
13079: @subsection Implementation Defined Options
13080: @c ---------------------------------------------------------------------
13081: @cindex implementation-defined options, facility words
13082: @cindex facility words, implementation-defined options
13083:
13084: @table @i
13085: @item encoding of keyboard events (@code{EKEY}):
13086: @cindex keyboard events, encoding in @code{EKEY}
13087: @cindex @code{EKEY}, encoding of keyboard events
13088: Keys corresponding to ASCII characters are encoded as ASCII characters.
13089: Other keys are encoded with the constants @code{k-left}, @code{k-right},
13090: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
13091: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
13092: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
13093:
13094:
13095: @item duration of a system clock tick:
13096: @cindex duration of a system clock tick
13097: @cindex clock tick duration
13098: System dependent. With respect to @code{MS}, the time is specified in
13099: microseconds. How well the OS and the hardware implement this, is
13100: another question.
13101:
13102: @item repeatability to be expected from the execution of @code{MS}:
13103: @cindex repeatability to be expected from the execution of @code{MS}
13104: @cindex @code{MS}, repeatability to be expected
13105: System dependent. On Unix, a lot depends on load. If the system is
13106: lightly loaded, and the delay is short enough that Gforth does not get
13107: swapped out, the performance should be acceptable. Under MS-DOS and
13108: other single-tasking systems, it should be good.
13109:
13110: @end table
13111:
13112:
13113: @c ---------------------------------------------------------------------
13114: @node facility-ambcond, , facility-idef, The optional Facility word set
13115: @subsection Ambiguous conditions
13116: @c ---------------------------------------------------------------------
13117: @cindex facility words, ambiguous conditions
13118: @cindex ambiguous conditions, facility words
13119:
13120: @table @i
13121: @item @code{AT-XY} can't be performed on user output device:
13122: @cindex @code{AT-XY} can't be performed on user output device
13123: Largely terminal dependent. No range checks are done on the arguments.
13124: No errors are reported. You may see some garbage appearing, you may see
13125: simply nothing happen.
13126:
13127: @end table
13128:
13129:
13130: @c =====================================================================
13131: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
13132: @section The optional File-Access word set
13133: @c =====================================================================
13134: @cindex system documentation, file words
13135: @cindex file words, system documentation
13136:
13137: @menu
13138: * file-idef:: Implementation Defined Options
13139: * file-ambcond:: Ambiguous Conditions
13140: @end menu
13141:
13142: @c ---------------------------------------------------------------------
13143: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
13144: @subsection Implementation Defined Options
13145: @c ---------------------------------------------------------------------
13146: @cindex implementation-defined options, file words
13147: @cindex file words, implementation-defined options
13148:
13149: @table @i
13150: @item file access methods used:
13151: @cindex file access methods used
13152: @code{R/O}, @code{R/W} and @code{BIN} work as you would
13153: expect. @code{W/O} translates into the C file opening mode @code{w} (or
13154: @code{wb}): The file is cleared, if it exists, and created, if it does
13155: not (with both @code{open-file} and @code{create-file}). Under Unix
13156: @code{create-file} creates a file with 666 permissions modified by your
13157: umask.
13158:
13159: @item file exceptions:
13160: @cindex file exceptions
13161: The file words do not raise exceptions (except, perhaps, memory access
13162: faults when you pass illegal addresses or file-ids).
13163:
13164: @item file line terminator:
13165: @cindex file line terminator
13166: System-dependent. Gforth uses C's newline character as line
13167: terminator. What the actual character code(s) of this are is
13168: system-dependent.
13169:
13170: @item file name format:
13171: @cindex file name format
13172: System dependent. Gforth just uses the file name format of your OS.
13173:
13174: @item information returned by @code{FILE-STATUS}:
13175: @cindex @code{FILE-STATUS}, returned information
13176: @code{FILE-STATUS} returns the most powerful file access mode allowed
13177: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
13178: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
13179: along with the returned mode.
13180:
13181: @item input file state after an exception when including source:
13182: @cindex exception when including source
13183: All files that are left via the exception are closed.
13184:
13185: @item @i{ior} values and meaning:
13186: @cindex @i{ior} values and meaning
13187: @cindex @i{wior} values and meaning
13188: The @i{ior}s returned by the file and memory allocation words are
13189: intended as throw codes. They typically are in the range
13190: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
13191: @i{ior}s is -512@minus{}@i{errno}.
13192:
13193: @item maximum depth of file input nesting:
13194: @cindex maximum depth of file input nesting
13195: @cindex file input nesting, maximum depth
13196: limited by the amount of return stack, locals/TIB stack, and the number
13197: of open files available. This should not give you troubles.
13198:
13199: @item maximum size of input line:
13200: @cindex maximum size of input line
13201: @cindex input line size, maximum
13202: @code{/line}. Currently 255.
13203:
13204: @item methods of mapping block ranges to files:
13205: @cindex mapping block ranges to files
13206: @cindex files containing blocks
13207: @cindex blocks in files
13208: By default, blocks are accessed in the file @file{blocks.fb} in the
13209: current working directory. The file can be switched with @code{USE}.
13210:
13211: @item number of string buffers provided by @code{S"}:
13212: @cindex @code{S"}, number of string buffers
13213: 1
13214:
13215: @item size of string buffer used by @code{S"}:
13216: @cindex @code{S"}, size of string buffer
13217: @code{/line}. currently 255.
13218:
13219: @end table
13220:
13221: @c ---------------------------------------------------------------------
13222: @node file-ambcond, , file-idef, The optional File-Access word set
13223: @subsection Ambiguous conditions
13224: @c ---------------------------------------------------------------------
13225: @cindex file words, ambiguous conditions
13226: @cindex ambiguous conditions, file words
13227:
13228: @table @i
13229: @item attempting to position a file outside its boundaries:
13230: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
13231: @code{REPOSITION-FILE} is performed as usual: Afterwards,
13232: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
13233:
13234: @item attempting to read from file positions not yet written:
13235: @cindex reading from file positions not yet written
13236: End-of-file, i.e., zero characters are read and no error is reported.
13237:
13238: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
13239: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
13240: An appropriate exception may be thrown, but a memory fault or other
13241: problem is more probable.
13242:
13243: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
13244: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
13245: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
13246: The @i{ior} produced by the operation, that discovered the problem, is
13247: thrown.
13248:
13249: @item named file cannot be opened (@code{INCLUDED}):
13250: @cindex @code{INCLUDED}, named file cannot be opened
13251: The @i{ior} produced by @code{open-file} is thrown.
13252:
13253: @item requesting an unmapped block number:
13254: @cindex unmapped block numbers
13255: There are no unmapped legal block numbers. On some operating systems,
13256: writing a block with a large number may overflow the file system and
13257: have an error message as consequence.
13258:
13259: @item using @code{source-id} when @code{blk} is non-zero:
13260: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
13261: @code{source-id} performs its function. Typically it will give the id of
13262: the source which loaded the block. (Better ideas?)
13263:
13264: @end table
13265:
13266:
13267: @c =====================================================================
13268: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
13269: @section The optional Floating-Point word set
13270: @c =====================================================================
13271: @cindex system documentation, floating-point words
13272: @cindex floating-point words, system documentation
13273:
13274: @menu
13275: * floating-idef:: Implementation Defined Options
13276: * floating-ambcond:: Ambiguous Conditions
13277: @end menu
13278:
13279:
13280: @c ---------------------------------------------------------------------
13281: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
13282: @subsection Implementation Defined Options
13283: @c ---------------------------------------------------------------------
13284: @cindex implementation-defined options, floating-point words
13285: @cindex floating-point words, implementation-defined options
13286:
13287: @table @i
13288: @item format and range of floating point numbers:
13289: @cindex format and range of floating point numbers
13290: @cindex floating point numbers, format and range
13291: System-dependent; the @code{double} type of C.
13292:
13293: @item results of @code{REPRESENT} when @i{float} is out of range:
13294: @cindex @code{REPRESENT}, results when @i{float} is out of range
13295: System dependent; @code{REPRESENT} is implemented using the C library
13296: function @code{ecvt()} and inherits its behaviour in this respect.
13297:
13298: @item rounding or truncation of floating-point numbers:
13299: @cindex rounding of floating-point numbers
13300: @cindex truncation of floating-point numbers
13301: @cindex floating-point numbers, rounding or truncation
13302: System dependent; the rounding behaviour is inherited from the hosting C
13303: compiler. IEEE-FP-based (i.e., most) systems by default round to
13304: nearest, and break ties by rounding to even (i.e., such that the last
13305: bit of the mantissa is 0).
13306:
13307: @item size of floating-point stack:
13308: @cindex floating-point stack size
13309: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
13310: the floating-point stack (in floats). You can specify this on startup
13311: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
13312:
13313: @item width of floating-point stack:
13314: @cindex floating-point stack width
13315: @code{1 floats}.
13316:
13317: @end table
13318:
13319:
13320: @c ---------------------------------------------------------------------
13321: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
13322: @subsection Ambiguous conditions
13323: @c ---------------------------------------------------------------------
13324: @cindex floating-point words, ambiguous conditions
13325: @cindex ambiguous conditions, floating-point words
13326:
13327: @table @i
13328: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
13329: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
13330: System-dependent. Typically results in a @code{-23 THROW} like other
13331: alignment violations.
13332:
13333: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
13334: @cindex @code{f@@} used with an address that is not float aligned
13335: @cindex @code{f!} used with an address that is not float aligned
13336: System-dependent. Typically results in a @code{-23 THROW} like other
13337: alignment violations.
13338:
13339: @item floating-point result out of range:
13340: @cindex floating-point result out of range
13341: System-dependent. Can result in a @code{-43 throw} (floating point
13342: overflow), @code{-54 throw} (floating point underflow), @code{-41 throw}
13343: (floating point inexact result), @code{-55 THROW} (Floating-point
13344: unidentified fault), or can produce a special value representing, e.g.,
13345: Infinity.
13346:
13347: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
13348: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
13349: System-dependent. Typically results in an alignment fault like other
13350: alignment violations.
13351:
13352: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
13353: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
13354: The floating-point number is converted into decimal nonetheless.
13355:
13356: @item Both arguments are equal to zero (@code{FATAN2}):
13357: @cindex @code{FATAN2}, both arguments are equal to zero
13358: System-dependent. @code{FATAN2} is implemented using the C library
13359: function @code{atan2()}.
13360:
13361: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
13362: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
13363: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
13364: because of small errors and the tan will be a very large (or very small)
13365: but finite number.
13366:
13367: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
13368: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
13369: The result is rounded to the nearest float.
13370:
13371: @item dividing by zero:
13372: @cindex dividing by zero, floating-point
13373: @cindex floating-point dividing by zero
13374: @cindex floating-point unidentified fault, FP divide-by-zero
13375: Platform-dependent; can produce an Infinity, NaN, @code{-42 throw}
13376: (floating point divide by zero) or @code{-55 throw} (Floating-point
13377: unidentified fault).
13378:
13379: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
13380: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
13381: System dependent. On IEEE-FP based systems the number is converted into
13382: an infinity.
13383:
13384: @item @i{float}<1 (@code{FACOSH}):
13385: @cindex @code{FACOSH}, @i{float}<1
13386: @cindex floating-point unidentified fault, @code{FACOSH}
13387: Platform-dependent; on IEEE-FP systems typically produces a NaN.
13388:
13389: @item @i{float}=<-1 (@code{FLNP1}):
13390: @cindex @code{FLNP1}, @i{float}=<-1
13391: @cindex floating-point unidentified fault, @code{FLNP1}
13392: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13393: negative infinity for @i{float}=-1).
13394:
13395: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
13396: @cindex @code{FLN}, @i{float}=<0
13397: @cindex @code{FLOG}, @i{float}=<0
13398: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
13399: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13400: negative infinity for @i{float}=0).
13401:
13402: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
13403: @cindex @code{FASINH}, @i{float}<0
13404: @cindex @code{FSQRT}, @i{float}<0
13405: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
13406: Platform-dependent; for @code{fsqrt} this typically gives a NaN, for
13407: @code{fasinh} some platforms produce a NaN, others a number (bug in the
13408: C library?).
13409:
13410: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
13411: @cindex @code{FACOS}, |@i{float}|>1
13412: @cindex @code{FASIN}, |@i{float}|>1
13413: @cindex @code{FATANH}, |@i{float}|>1
13414: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
13415: Platform-dependent; IEEE-FP systems typically produce a NaN.
13416:
13417: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
13418: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
13419: @cindex floating-point unidentified fault, @code{F>D}
13420: Platform-dependent; typically, some double number is produced and no
13421: error is reported.
13422:
13423: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
13424: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
13425: @code{Precision} characters of the numeric output area are used. If
13426: @code{precision} is too high, these words will smash the data or code
13427: close to @code{here}.
13428: @end table
13429:
13430: @c =====================================================================
13431: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
13432: @section The optional Locals word set
13433: @c =====================================================================
13434: @cindex system documentation, locals words
13435: @cindex locals words, system documentation
13436:
13437: @menu
13438: * locals-idef:: Implementation Defined Options
13439: * locals-ambcond:: Ambiguous Conditions
13440: @end menu
13441:
13442:
13443: @c ---------------------------------------------------------------------
13444: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
13445: @subsection Implementation Defined Options
13446: @c ---------------------------------------------------------------------
13447: @cindex implementation-defined options, locals words
13448: @cindex locals words, implementation-defined options
13449:
13450: @table @i
13451: @item maximum number of locals in a definition:
13452: @cindex maximum number of locals in a definition
13453: @cindex locals, maximum number in a definition
13454: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
13455: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
13456: characters. The number of locals in a definition is bounded by the size
13457: of locals-buffer, which contains the names of the locals.
13458:
13459: @end table
13460:
13461:
13462: @c ---------------------------------------------------------------------
13463: @node locals-ambcond, , locals-idef, The optional Locals word set
13464: @subsection Ambiguous conditions
13465: @c ---------------------------------------------------------------------
13466: @cindex locals words, ambiguous conditions
13467: @cindex ambiguous conditions, locals words
13468:
13469: @table @i
13470: @item executing a named local in interpretation state:
13471: @cindex local in interpretation state
13472: @cindex Interpreting a compile-only word, for a local
13473: Locals have no interpretation semantics. If you try to perform the
13474: interpretation semantics, you will get a @code{-14 throw} somewhere
13475: (Interpreting a compile-only word). If you perform the compilation
13476: semantics, the locals access will be compiled (irrespective of state).
13477:
13478: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
13479: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
13480: @cindex @code{TO} on non-@code{VALUE}s and non-locals
13481: @cindex Invalid name argument, @code{TO}
13482: @code{-32 throw} (Invalid name argument)
13483:
13484: @end table
13485:
13486:
13487: @c =====================================================================
13488: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
13489: @section The optional Memory-Allocation word set
13490: @c =====================================================================
13491: @cindex system documentation, memory-allocation words
13492: @cindex memory-allocation words, system documentation
13493:
13494: @menu
13495: * memory-idef:: Implementation Defined Options
13496: @end menu
13497:
13498:
13499: @c ---------------------------------------------------------------------
13500: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
13501: @subsection Implementation Defined Options
13502: @c ---------------------------------------------------------------------
13503: @cindex implementation-defined options, memory-allocation words
13504: @cindex memory-allocation words, implementation-defined options
13505:
13506: @table @i
13507: @item values and meaning of @i{ior}:
13508: @cindex @i{ior} values and meaning
13509: The @i{ior}s returned by the file and memory allocation words are
13510: intended as throw codes. They typically are in the range
13511: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
13512: @i{ior}s is -512@minus{}@i{errno}.
13513:
13514: @end table
13515:
13516: @c =====================================================================
13517: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
13518: @section The optional Programming-Tools word set
13519: @c =====================================================================
13520: @cindex system documentation, programming-tools words
13521: @cindex programming-tools words, system documentation
13522:
13523: @menu
13524: * programming-idef:: Implementation Defined Options
13525: * programming-ambcond:: Ambiguous Conditions
13526: @end menu
13527:
13528:
13529: @c ---------------------------------------------------------------------
13530: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
13531: @subsection Implementation Defined Options
13532: @c ---------------------------------------------------------------------
13533: @cindex implementation-defined options, programming-tools words
13534: @cindex programming-tools words, implementation-defined options
13535:
13536: @table @i
13537: @item ending sequence for input following @code{;CODE} and @code{CODE}:
13538: @cindex @code{;CODE} ending sequence
13539: @cindex @code{CODE} ending sequence
13540: @code{END-CODE}
13541:
13542: @item manner of processing input following @code{;CODE} and @code{CODE}:
13543: @cindex @code{;CODE}, processing input
13544: @cindex @code{CODE}, processing input
13545: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
13546: the input is processed by the text interpreter, (starting) in interpret
13547: state.
13548:
13549: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
13550: @cindex @code{ASSEMBLER}, search order capability
13551: The ANS Forth search order word set.
13552:
13553: @item source and format of display by @code{SEE}:
13554: @cindex @code{SEE}, source and format of output
13555: The source for @code{see} is the executable code used by the inner
13556: interpreter. The current @code{see} tries to output Forth source code
13557: (and on some platforms, assembly code for primitives) as well as
13558: possible.
13559:
13560: @end table
13561:
13562: @c ---------------------------------------------------------------------
13563: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
13564: @subsection Ambiguous conditions
13565: @c ---------------------------------------------------------------------
13566: @cindex programming-tools words, ambiguous conditions
13567: @cindex ambiguous conditions, programming-tools words
13568:
13569: @table @i
13570:
13571: @item deleting the compilation word list (@code{FORGET}):
13572: @cindex @code{FORGET}, deleting the compilation word list
13573: Not implemented (yet).
13574:
13575: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
13576: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
13577: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
13578: @cindex control-flow stack underflow
13579: This typically results in an @code{abort"} with a descriptive error
13580: message (may change into a @code{-22 throw} (Control structure mismatch)
13581: in the future). You may also get a memory access error. If you are
13582: unlucky, this ambiguous condition is not caught.
13583:
13584: @item @i{name} can't be found (@code{FORGET}):
13585: @cindex @code{FORGET}, @i{name} can't be found
13586: Not implemented (yet).
13587:
13588: @item @i{name} not defined via @code{CREATE}:
13589: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
13590: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
13591: the execution semantics of the last defined word no matter how it was
13592: defined.
13593:
13594: @item @code{POSTPONE} applied to @code{[IF]}:
13595: @cindex @code{POSTPONE} applied to @code{[IF]}
13596: @cindex @code{[IF]} and @code{POSTPONE}
13597: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
13598: equivalent to @code{[IF]}.
13599:
13600: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
13601: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
13602: Continue in the same state of conditional compilation in the next outer
13603: input source. Currently there is no warning to the user about this.
13604:
13605: @item removing a needed definition (@code{FORGET}):
13606: @cindex @code{FORGET}, removing a needed definition
13607: Not implemented (yet).
13608:
13609: @end table
13610:
13611:
13612: @c =====================================================================
13613: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
13614: @section The optional Search-Order word set
13615: @c =====================================================================
13616: @cindex system documentation, search-order words
13617: @cindex search-order words, system documentation
13618:
13619: @menu
13620: * search-idef:: Implementation Defined Options
13621: * search-ambcond:: Ambiguous Conditions
13622: @end menu
13623:
13624:
13625: @c ---------------------------------------------------------------------
13626: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
13627: @subsection Implementation Defined Options
13628: @c ---------------------------------------------------------------------
13629: @cindex implementation-defined options, search-order words
13630: @cindex search-order words, implementation-defined options
13631:
13632: @table @i
13633: @item maximum number of word lists in search order:
13634: @cindex maximum number of word lists in search order
13635: @cindex search order, maximum depth
13636: @code{s" wordlists" environment? drop .}. Currently 16.
13637:
13638: @item minimum search order:
13639: @cindex minimum search order
13640: @cindex search order, minimum
13641: @code{root root}.
13642:
13643: @end table
13644:
13645: @c ---------------------------------------------------------------------
13646: @node search-ambcond, , search-idef, The optional Search-Order word set
13647: @subsection Ambiguous conditions
13648: @c ---------------------------------------------------------------------
13649: @cindex search-order words, ambiguous conditions
13650: @cindex ambiguous conditions, search-order words
13651:
13652: @table @i
13653: @item changing the compilation word list (during compilation):
13654: @cindex changing the compilation word list (during compilation)
13655: @cindex compilation word list, change before definition ends
13656: The word is entered into the word list that was the compilation word list
13657: at the start of the definition. Any changes to the name field (e.g.,
13658: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
13659: are applied to the latest defined word (as reported by @code{last} or
13660: @code{lastxt}), if possible, irrespective of the compilation word list.
13661:
13662: @item search order empty (@code{previous}):
13663: @cindex @code{previous}, search order empty
13664: @cindex vocstack empty, @code{previous}
13665: @code{abort" Vocstack empty"}.
13666:
13667: @item too many word lists in search order (@code{also}):
13668: @cindex @code{also}, too many word lists in search order
13669: @cindex vocstack full, @code{also}
13670: @code{abort" Vocstack full"}.
13671:
13672: @end table
13673:
13674: @c ***************************************************************
13675: @node Standard vs Extensions, Model, ANS conformance, Top
13676: @chapter Should I use Gforth extensions?
13677: @cindex Gforth extensions
13678:
13679: As you read through the rest of this manual, you will see documentation
13680: for @i{Standard} words, and documentation for some appealing Gforth
13681: @i{extensions}. You might ask yourself the question: @i{``Should I
13682: restrict myself to the standard, or should I use the extensions?''}
13683:
13684: The answer depends on the goals you have for the program you are working
13685: on:
13686:
13687: @itemize @bullet
13688:
13689: @item Is it just for yourself or do you want to share it with others?
13690:
13691: @item
13692: If you want to share it, do the others all use Gforth?
13693:
13694: @item
13695: If it is just for yourself, do you want to restrict yourself to Gforth?
13696:
13697: @end itemize
13698:
13699: If restricting the program to Gforth is ok, then there is no reason not
13700: to use extensions. It is still a good idea to keep to the standard
13701: where it is easy, in case you want to reuse these parts in another
13702: program that you want to be portable.
13703:
13704: If you want to be able to port the program to other Forth systems, there
13705: are the following points to consider:
13706:
13707: @itemize @bullet
13708:
13709: @item
13710: Most Forth systems that are being maintained support the ANS Forth
13711: standard. So if your program complies with the standard, it will be
13712: portable among many systems.
13713:
13714: @item
13715: A number of the Gforth extensions can be implemented in ANS Forth using
13716: public-domain files provided in the @file{compat/} directory. These are
13717: mentioned in the text in passing. There is no reason not to use these
13718: extensions, your program will still be ANS Forth compliant; just include
13719: the appropriate compat files with your program.
13720:
13721: @item
13722: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
13723: analyse your program and determine what non-Standard words it relies
13724: upon. However, it does not check whether you use standard words in a
13725: non-standard way.
13726:
13727: @item
13728: Some techniques are not standardized by ANS Forth, and are hard or
13729: impossible to implement in a standard way, but can be implemented in
13730: most Forth systems easily, and usually in similar ways (e.g., accessing
13731: word headers). Forth has a rich historical precedent for programmers
13732: taking advantage of implementation-dependent features of their tools
13733: (for example, relying on a knowledge of the dictionary
13734: structure). Sometimes these techniques are necessary to extract every
13735: last bit of performance from the hardware, sometimes they are just a
13736: programming shorthand.
13737:
13738: @item
13739: Does using a Gforth extension save more work than the porting this part
13740: to other Forth systems (if any) will cost?
13741:
13742: @item
13743: Is the additional functionality worth the reduction in portability and
13744: the additional porting problems?
13745:
13746: @end itemize
13747:
13748: In order to perform these consideratios, you need to know what's
13749: standard and what's not. This manual generally states if something is
13750: non-standard, but the authoritative source is the
13751: @uref{http://www.taygeta.com/forth/dpans.html,standard document}.
13752: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
13753: into the thought processes of the technical committee.
13754:
13755: Note also that portability between Forth systems is not the only
13756: portability issue; there is also the issue of portability between
13757: different platforms (processor/OS combinations).
13758:
13759: @c ***************************************************************
13760: @node Model, Integrating Gforth, Standard vs Extensions, Top
13761: @chapter Model
13762:
13763: This chapter has yet to be written. It will contain information, on
13764: which internal structures you can rely.
13765:
13766: @c ***************************************************************
13767: @node Integrating Gforth, Emacs and Gforth, Model, Top
13768: @chapter Integrating Gforth into C programs
13769:
13770: This is not yet implemented.
13771:
13772: Several people like to use Forth as scripting language for applications
13773: that are otherwise written in C, C++, or some other language.
13774:
13775: The Forth system ATLAST provides facilities for embedding it into
13776: applications; unfortunately it has several disadvantages: most
13777: importantly, it is not based on ANS Forth, and it is apparently dead
13778: (i.e., not developed further and not supported). The facilities
13779: provided by Gforth in this area are inspired by ATLAST's facilities, so
13780: making the switch should not be hard.
13781:
13782: We also tried to design the interface such that it can easily be
13783: implemented by other Forth systems, so that we may one day arrive at a
13784: standardized interface. Such a standard interface would allow you to
13785: replace the Forth system without having to rewrite C code.
13786:
13787: You embed the Gforth interpreter by linking with the library
13788: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
13789: global symbols in this library that belong to the interface, have the
13790: prefix @code{forth_}. (Global symbols that are used internally have the
13791: prefix @code{gforth_}).
13792:
13793: You can include the declarations of Forth types and the functions and
13794: variables of the interface with @code{#include <forth.h>}.
13795:
13796: Types.
13797:
13798: Variables.
13799:
13800: Data and FP Stack pointer. Area sizes.
13801:
13802: functions.
13803:
13804: forth_init(imagefile)
13805: forth_evaluate(string) exceptions?
13806: forth_goto(address) (or forth_execute(xt)?)
13807: forth_continue() (a corountining mechanism)
13808:
13809: Adding primitives.
13810:
13811: No checking.
13812:
13813: Signals?
13814:
13815: Accessing the Stacks
13816:
13817: @c ******************************************************************
13818: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
13819: @chapter Emacs and Gforth
13820: @cindex Emacs and Gforth
13821:
13822: @cindex @file{gforth.el}
13823: @cindex @file{forth.el}
13824: @cindex Rydqvist, Goran
13825: @cindex Kuehling, David
13826: @cindex comment editing commands
13827: @cindex @code{\}, editing with Emacs
13828: @cindex debug tracer editing commands
13829: @cindex @code{~~}, removal with Emacs
13830: @cindex Forth mode in Emacs
13831:
13832: Gforth comes with @file{gforth.el}, an improved version of
13833: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
13834: improvements are:
13835:
13836: @itemize @bullet
13837: @item
13838: A better handling of indentation.
13839: @item
13840: A custom hilighting engine for Forth-code.
13841: @item
13842: Comment paragraph filling (@kbd{M-q})
13843: @item
13844: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
13845: @item
13846: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
13847: @item
13848: Support of the @code{info-lookup} feature for looking up the
13849: documentation of a word.
13850: @item
13851: Support for reading and writing blocks files.
13852: @end itemize
13853:
13854: To get a basic description of these features, enter Forth mode and
13855: type @kbd{C-h m}.
13856:
13857: @cindex source location of error or debugging output in Emacs
13858: @cindex error output, finding the source location in Emacs
13859: @cindex debugging output, finding the source location in Emacs
13860: In addition, Gforth supports Emacs quite well: The source code locations
13861: given in error messages, debugging output (from @code{~~}) and failed
13862: assertion messages are in the right format for Emacs' compilation mode
13863: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
13864: Manual}) so the source location corresponding to an error or other
13865: message is only a few keystrokes away (@kbd{C-x `} for the next error,
13866: @kbd{C-c C-c} for the error under the cursor).
13867:
13868: @cindex viewing the documentation of a word in Emacs
13869: @cindex context-sensitive help
13870: Moreover, for words documented in this manual, you can look up the
13871: glossary entry quickly by using @kbd{C-h TAB}
13872: (@code{info-lookup-symbol}, @pxref{Documentation, ,Documentation
13873: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
13874: later and does not work for words containing @code{:}.
13875:
13876: @menu
13877: * Installing gforth.el:: Making Emacs aware of Forth.
13878: * Emacs Tags:: Viewing the source of a word in Emacs.
13879: * Hilighting:: Making Forth code look prettier.
13880: * Auto-Indentation:: Customizing auto-indentation.
13881: * Blocks Files:: Reading and writing blocks files.
13882: @end menu
13883:
13884: @c ----------------------------------
13885: @node Installing gforth.el, Emacs Tags, , Emacs and Gforth
13886: @section Installing gforth.el
13887: @cindex @file{.emacs}
13888: @cindex @file{gforth.el}, installation
13889: To make the features from @file{gforth.el} available in Emacs, add
13890: the following lines to your @file{.emacs} file:
13891:
13892: @example
13893: (autoload 'forth-mode "gforth.el")
13894: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode)
13895: auto-mode-alist))
13896: (autoload 'forth-block-mode "gforth.el")
13897: (setq auto-mode-alist (cons '("\\.fb\\'" . forth-block-mode)
13898: auto-mode-alist))
13899: (add-hook 'forth-mode-hook (function (lambda ()
13900: ;; customize variables here:
13901: (setq forth-indent-level 4)
13902: (setq forth-minor-indent-level 2)
13903: (setq forth-hilight-level 3)
13904: ;;; ...
13905: )))
13906: @end example
13907:
13908: @c ----------------------------------
13909: @node Emacs Tags, Hilighting, Installing gforth.el, Emacs and Gforth
13910: @section Emacs Tags
13911: @cindex @file{TAGS} file
13912: @cindex @file{etags.fs}
13913: @cindex viewing the source of a word in Emacs
13914: @cindex @code{require}, placement in files
13915: @cindex @code{include}, placement in files
13916: If you @code{require} @file{etags.fs}, a new @file{TAGS} file will be
13917: produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
13918: contains the definitions of all words defined afterwards. You can then
13919: find the source for a word using @kbd{M-.}. Note that Emacs can use
13920: several tags files at the same time (e.g., one for the Gforth sources
13921: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
13922: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
13923: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
13924: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
13925: with @file{etags.fs}, you should avoid putting definitions both before
13926: and after @code{require} etc., otherwise you will see the same file
13927: visited several times by commands like @code{tags-search}.
13928:
13929: @c ----------------------------------
13930: @node Hilighting, Auto-Indentation, Emacs Tags, Emacs and Gforth
13931: @section Hilighting
13932: @cindex hilighting Forth code in Emacs
13933: @cindex highlighting Forth code in Emacs
13934: @file{gforth.el} comes with a custom source hilighting engine. When
13935: you open a file in @code{forth-mode}, it will be completely parsed,
13936: assigning faces to keywords, comments, strings etc. While you edit
13937: the file, modified regions get parsed and updated on-the-fly.
13938:
13939: Use the variable `forth-hilight-level' to change the level of
13940: decoration from 0 (no hilighting at all) to 3 (the default). Even if
13941: you set the hilighting level to 0, the parser will still work in the
13942: background, collecting information about whether regions of text are
13943: ``compiled'' or ``interpreted''. Those information are required for
13944: auto-indentation to work properly. Set `forth-disable-parser' to
13945: non-nil if your computer is too slow to handle parsing. This will
13946: have an impact on the smartness of the auto-indentation engine,
13947: though.
13948:
13949: Sometimes Forth sources define new features that should be hilighted,
13950: new control structures, defining-words etc. You can use the variable
13951: `forth-custom-words' to make @code{forth-mode} hilight additional
13952: words and constructs. See the docstring of `forth-words' for details
13953: (in Emacs, type @kbd{C-h v forth-words}).
13954:
13955: `forth-custom-words' is meant to be customized in your
13956: @file{.emacs} file. To customize hilighing in a file-specific manner,
13957: set `forth-local-words' in a local-variables section at the end of
13958: your source file (@pxref{Local Variables in Files,, Variables, emacs, Emacs Manual}).
13959:
13960: Example:
13961: @example
13962: 0 [IF]
13963: Local Variables:
13964: forth-local-words:
13965: ((("t:") definition-starter (font-lock-keyword-face . 1)
13966: "[ \t\n]" t name (font-lock-function-name-face . 3))
13967: ((";t") definition-ender (font-lock-keyword-face . 1)))
13968: End:
13969: [THEN]
13970: @end example
13971:
13972: @c ----------------------------------
13973: @node Auto-Indentation, Blocks Files, Hilighting, Emacs and Gforth
13974: @section Auto-Indentation
13975: @cindex auto-indentation of Forth code in Emacs
13976: @cindex indentation of Forth code in Emacs
13977: @code{forth-mode} automatically tries to indent lines in a smart way,
13978: whenever you type @key{TAB} or break a line with @kbd{C-m}.
13979:
13980: Simple customization can be achieved by setting
13981: `forth-indent-level' and `forth-minor-indent-level' in your
13982: @file{.emacs} file. For historical reasons @file{gforth.el} indents
13983: per default by multiples of 4 columns. To use the more traditional
13984: 3-column indentation, add the following lines to your @file{.emacs}:
13985:
13986: @example
13987: (add-hook 'forth-mode-hook (function (lambda ()
13988: ;; customize variables here:
13989: (setq forth-indent-level 3)
13990: (setq forth-minor-indent-level 1)
13991: )))
13992: @end example
13993:
13994: If you want indentation to recognize non-default words, customize it
13995: by setting `forth-custom-indent-words' in your @file{.emacs}. See the
13996: docstring of `forth-indent-words' for details (in Emacs, type @kbd{C-h
13997: v forth-indent-words}).
13998:
13999: To customize indentation in a file-specific manner, set
14000: `forth-local-indent-words' in a local-variables section at the end of
14001: your source file (@pxref{Local Variables in Files, Variables,,emacs,
14002: Emacs Manual}).
14003:
14004: Example:
14005: @example
14006: 0 [IF]
14007: Local Variables:
14008: forth-local-indent-words:
14009: ((("t:") (0 . 2) (0 . 2))
14010: ((";t") (-2 . 0) (0 . -2)))
14011: End:
14012: [THEN]
14013: @end example
14014:
14015: @c ----------------------------------
14016: @node Blocks Files,, Auto-Indentation, Emacs and Gforth
14017: @section Blocks Files
14018: @cindex blocks files, use with Emacs
14019: @code{forth-mode} Autodetects blocks files by checking whether the
14020: length of the first line exceeds 1023 characters. It then tries to
14021: convert the file into normal text format. When you save the file, it
14022: will be written to disk as normal stream-source file.
14023:
14024: If you want to write blocks files, use @code{forth-blocks-mode}. It
14025: inherits all the features from @code{forth-mode}, plus some additions:
14026:
14027: @itemize @bullet
14028: @item
14029: Files are written to disk in blocks file format.
14030: @item
14031: Screen numbers are displayed in the mode line (enumerated beginning
14032: with the value of `forth-block-base')
14033: @item
14034: Warnings are displayed when lines exceed 64 characters.
14035: @item
14036: The beginning of the currently edited block is marked with an
14037: overlay-arrow.
14038: @end itemize
14039:
14040: There are some restrictions you should be aware of. When you open a
14041: blocks file that contains tabulator or newline characters, these
14042: characters will be translated into spaces when the file is written
14043: back to disk. If tabs or newlines are encountered during blocks file
14044: reading, an error is output to the echo area. So have a look at the
14045: `*Messages*' buffer, when Emacs' bell rings during reading.
14046:
14047: Please consult the docstring of @code{forth-blocks-mode} for more
14048: information by typing @kbd{C-h v forth-blocks-mode}).
14049:
14050: @c ******************************************************************
14051: @node Image Files, Engine, Emacs and Gforth, Top
14052: @chapter Image Files
14053: @cindex image file
14054: @cindex @file{.fi} files
14055: @cindex precompiled Forth code
14056: @cindex dictionary in persistent form
14057: @cindex persistent form of dictionary
14058:
14059: An image file is a file containing an image of the Forth dictionary,
14060: i.e., compiled Forth code and data residing in the dictionary. By
14061: convention, we use the extension @code{.fi} for image files.
14062:
14063: @menu
14064: * Image Licensing Issues:: Distribution terms for images.
14065: * Image File Background:: Why have image files?
14066: * Non-Relocatable Image Files:: don't always work.
14067: * Data-Relocatable Image Files:: are better.
14068: * Fully Relocatable Image Files:: better yet.
14069: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
14070: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
14071: * Modifying the Startup Sequence:: and turnkey applications.
14072: @end menu
14073:
14074: @node Image Licensing Issues, Image File Background, Image Files, Image Files
14075: @section Image Licensing Issues
14076: @cindex license for images
14077: @cindex image license
14078:
14079: An image created with @code{gforthmi} (@pxref{gforthmi}) or
14080: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
14081: original image; i.e., according to copyright law it is a derived work of
14082: the original image.
14083:
14084: Since Gforth is distributed under the GNU GPL, the newly created image
14085: falls under the GNU GPL, too. In particular, this means that if you
14086: distribute the image, you have to make all of the sources for the image
14087: available, including those you wrote. For details see @ref{License, ,
14088: GNU General Public License (Section 3)}.
14089:
14090: If you create an image with @code{cross} (@pxref{cross.fs}), the image
14091: contains only code compiled from the sources you gave it; if none of
14092: these sources is under the GPL, the terms discussed above do not apply
14093: to the image. However, if your image needs an engine (a gforth binary)
14094: that is under the GPL, you should make sure that you distribute both in
14095: a way that is at most a @emph{mere aggregation}, if you don't want the
14096: terms of the GPL to apply to the image.
14097:
14098: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
14099: @section Image File Background
14100: @cindex image file background
14101:
14102: Gforth consists not only of primitives (in the engine), but also of
14103: definitions written in Forth. Since the Forth compiler itself belongs to
14104: those definitions, it is not possible to start the system with the
14105: engine and the Forth source alone. Therefore we provide the Forth
14106: code as an image file in nearly executable form. When Gforth starts up,
14107: a C routine loads the image file into memory, optionally relocates the
14108: addresses, then sets up the memory (stacks etc.) according to
14109: information in the image file, and (finally) starts executing Forth
14110: code.
14111:
14112: The image file variants represent different compromises between the
14113: goals of making it easy to generate image files and making them
14114: portable.
14115:
14116: @cindex relocation at run-time
14117: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
14118: run-time. This avoids many of the complications discussed below (image
14119: files are data relocatable without further ado), but costs performance
14120: (one addition per memory access).
14121:
14122: @cindex relocation at load-time
14123: By contrast, the Gforth loader performs relocation at image load time. The
14124: loader also has to replace tokens that represent primitive calls with the
14125: appropriate code-field addresses (or code addresses in the case of
14126: direct threading).
14127:
14128: There are three kinds of image files, with different degrees of
14129: relocatability: non-relocatable, data-relocatable, and fully relocatable
14130: image files.
14131:
14132: @cindex image file loader
14133: @cindex relocating loader
14134: @cindex loader for image files
14135: These image file variants have several restrictions in common; they are
14136: caused by the design of the image file loader:
14137:
14138: @itemize @bullet
14139: @item
14140: There is only one segment; in particular, this means, that an image file
14141: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
14142: them). The contents of the stacks are not represented, either.
14143:
14144: @item
14145: The only kinds of relocation supported are: adding the same offset to
14146: all cells that represent data addresses; and replacing special tokens
14147: with code addresses or with pieces of machine code.
14148:
14149: If any complex computations involving addresses are performed, the
14150: results cannot be represented in the image file. Several applications that
14151: use such computations come to mind:
14152: @itemize @minus
14153: @item
14154: Hashing addresses (or data structures which contain addresses) for table
14155: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
14156: purpose, you will have no problem, because the hash tables are
14157: recomputed automatically when the system is started. If you use your own
14158: hash tables, you will have to do something similar.
14159:
14160: @item
14161: There's a cute implementation of doubly-linked lists that uses
14162: @code{XOR}ed addresses. You could represent such lists as singly-linked
14163: in the image file, and restore the doubly-linked representation on
14164: startup.@footnote{In my opinion, though, you should think thrice before
14165: using a doubly-linked list (whatever implementation).}
14166:
14167: @item
14168: The code addresses of run-time routines like @code{docol:} cannot be
14169: represented in the image file (because their tokens would be replaced by
14170: machine code in direct threaded implementations). As a workaround,
14171: compute these addresses at run-time with @code{>code-address} from the
14172: executions tokens of appropriate words (see the definitions of
14173: @code{docol:} and friends in @file{kernel/getdoers.fs}).
14174:
14175: @item
14176: On many architectures addresses are represented in machine code in some
14177: shifted or mangled form. You cannot put @code{CODE} words that contain
14178: absolute addresses in this form in a relocatable image file. Workarounds
14179: are representing the address in some relative form (e.g., relative to
14180: the CFA, which is present in some register), or loading the address from
14181: a place where it is stored in a non-mangled form.
14182: @end itemize
14183: @end itemize
14184:
14185: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
14186: @section Non-Relocatable Image Files
14187: @cindex non-relocatable image files
14188: @cindex image file, non-relocatable
14189:
14190: These files are simple memory dumps of the dictionary. They are specific
14191: to the executable (i.e., @file{gforth} file) they were created
14192: with. What's worse, they are specific to the place on which the
14193: dictionary resided when the image was created. Now, there is no
14194: guarantee that the dictionary will reside at the same place the next
14195: time you start Gforth, so there's no guarantee that a non-relocatable
14196: image will work the next time (Gforth will complain instead of crashing,
14197: though).
14198:
14199: You can create a non-relocatable image file with
14200:
14201:
14202: doc-savesystem
14203:
14204:
14205: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
14206: @section Data-Relocatable Image Files
14207: @cindex data-relocatable image files
14208: @cindex image file, data-relocatable
14209:
14210: These files contain relocatable data addresses, but fixed code addresses
14211: (instead of tokens). They are specific to the executable (i.e.,
14212: @file{gforth} file) they were created with. For direct threading on some
14213: architectures (e.g., the i386), data-relocatable images do not work. You
14214: get a data-relocatable image, if you use @file{gforthmi} with a
14215: Gforth binary that is not doubly indirect threaded (@pxref{Fully
14216: Relocatable Image Files}).
14217:
14218: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
14219: @section Fully Relocatable Image Files
14220: @cindex fully relocatable image files
14221: @cindex image file, fully relocatable
14222:
14223: @cindex @file{kern*.fi}, relocatability
14224: @cindex @file{gforth.fi}, relocatability
14225: These image files have relocatable data addresses, and tokens for code
14226: addresses. They can be used with different binaries (e.g., with and
14227: without debugging) on the same machine, and even across machines with
14228: the same data formats (byte order, cell size, floating point
14229: format). However, they are usually specific to the version of Gforth
14230: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
14231: are fully relocatable.
14232:
14233: There are two ways to create a fully relocatable image file:
14234:
14235: @menu
14236: * gforthmi:: The normal way
14237: * cross.fs:: The hard way
14238: @end menu
14239:
14240: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
14241: @subsection @file{gforthmi}
14242: @cindex @file{comp-i.fs}
14243: @cindex @file{gforthmi}
14244:
14245: You will usually use @file{gforthmi}. If you want to create an
14246: image @i{file} that contains everything you would load by invoking
14247: Gforth with @code{gforth @i{options}}, you simply say:
14248: @example
14249: gforthmi @i{file} @i{options}
14250: @end example
14251:
14252: E.g., if you want to create an image @file{asm.fi} that has the file
14253: @file{asm.fs} loaded in addition to the usual stuff, you could do it
14254: like this:
14255:
14256: @example
14257: gforthmi asm.fi asm.fs
14258: @end example
14259:
14260: @file{gforthmi} is implemented as a sh script and works like this: It
14261: produces two non-relocatable images for different addresses and then
14262: compares them. Its output reflects this: first you see the output (if
14263: any) of the two Gforth invocations that produce the non-relocatable image
14264: files, then you see the output of the comparing program: It displays the
14265: offset used for data addresses and the offset used for code addresses;
14266: moreover, for each cell that cannot be represented correctly in the
14267: image files, it displays a line like this:
14268:
14269: @example
14270: 78DC BFFFFA50 BFFFFA40
14271: @end example
14272:
14273: This means that at offset $78dc from @code{forthstart}, one input image
14274: contains $bffffa50, and the other contains $bffffa40. Since these cells
14275: cannot be represented correctly in the output image, you should examine
14276: these places in the dictionary and verify that these cells are dead
14277: (i.e., not read before they are written).
14278:
14279: @cindex --application, @code{gforthmi} option
14280: If you insert the option @code{--application} in front of the image file
14281: name, you will get an image that uses the @code{--appl-image} option
14282: instead of the @code{--image-file} option (@pxref{Invoking
14283: Gforth}). When you execute such an image on Unix (by typing the image
14284: name as command), the Gforth engine will pass all options to the image
14285: instead of trying to interpret them as engine options.
14286:
14287: If you type @file{gforthmi} with no arguments, it prints some usage
14288: instructions.
14289:
14290: @cindex @code{savesystem} during @file{gforthmi}
14291: @cindex @code{bye} during @file{gforthmi}
14292: @cindex doubly indirect threaded code
14293: @cindex environment variables
14294: @cindex @code{GFORTHD} -- environment variable
14295: @cindex @code{GFORTH} -- environment variable
14296: @cindex @code{gforth-ditc}
14297: There are a few wrinkles: After processing the passed @i{options}, the
14298: words @code{savesystem} and @code{bye} must be visible. A special doubly
14299: indirect threaded version of the @file{gforth} executable is used for
14300: creating the non-relocatable images; you can pass the exact filename of
14301: this executable through the environment variable @code{GFORTHD}
14302: (default: @file{gforth-ditc}); if you pass a version that is not doubly
14303: indirect threaded, you will not get a fully relocatable image, but a
14304: data-relocatable image (because there is no code address offset). The
14305: normal @file{gforth} executable is used for creating the relocatable
14306: image; you can pass the exact filename of this executable through the
14307: environment variable @code{GFORTH}.
14308:
14309: @node cross.fs, , gforthmi, Fully Relocatable Image Files
14310: @subsection @file{cross.fs}
14311: @cindex @file{cross.fs}
14312: @cindex cross-compiler
14313: @cindex metacompiler
14314: @cindex target compiler
14315:
14316: You can also use @code{cross}, a batch compiler that accepts a Forth-like
14317: programming language (@pxref{Cross Compiler}).
14318:
14319: @code{cross} allows you to create image files for machines with
14320: different data sizes and data formats than the one used for generating
14321: the image file. You can also use it to create an application image that
14322: does not contain a Forth compiler. These features are bought with
14323: restrictions and inconveniences in programming. E.g., addresses have to
14324: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
14325: order to make the code relocatable.
14326:
14327:
14328: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
14329: @section Stack and Dictionary Sizes
14330: @cindex image file, stack and dictionary sizes
14331: @cindex dictionary size default
14332: @cindex stack size default
14333:
14334: If you invoke Gforth with a command line flag for the size
14335: (@pxref{Invoking Gforth}), the size you specify is stored in the
14336: dictionary. If you save the dictionary with @code{savesystem} or create
14337: an image with @file{gforthmi}, this size will become the default
14338: for the resulting image file. E.g., the following will create a
14339: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
14340:
14341: @example
14342: gforthmi gforth.fi -m 1M
14343: @end example
14344:
14345: In other words, if you want to set the default size for the dictionary
14346: and the stacks of an image, just invoke @file{gforthmi} with the
14347: appropriate options when creating the image.
14348:
14349: @cindex stack size, cache-friendly
14350: Note: For cache-friendly behaviour (i.e., good performance), you should
14351: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
14352: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
14353: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
14354:
14355: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
14356: @section Running Image Files
14357: @cindex running image files
14358: @cindex invoking image files
14359: @cindex image file invocation
14360:
14361: @cindex -i, invoke image file
14362: @cindex --image file, invoke image file
14363: You can invoke Gforth with an image file @i{image} instead of the
14364: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
14365: @example
14366: gforth -i @i{image}
14367: @end example
14368:
14369: @cindex executable image file
14370: @cindex image file, executable
14371: If your operating system supports starting scripts with a line of the
14372: form @code{#! ...}, you just have to type the image file name to start
14373: Gforth with this image file (note that the file extension @code{.fi} is
14374: just a convention). I.e., to run Gforth with the image file @i{image},
14375: you can just type @i{image} instead of @code{gforth -i @i{image}}.
14376: This works because every @code{.fi} file starts with a line of this
14377: format:
14378:
14379: @example
14380: #! /usr/local/bin/gforth-0.4.0 -i
14381: @end example
14382:
14383: The file and pathname for the Gforth engine specified on this line is
14384: the specific Gforth executable that it was built against; i.e. the value
14385: of the environment variable @code{GFORTH} at the time that
14386: @file{gforthmi} was executed.
14387:
14388: You can make use of the same shell capability to make a Forth source
14389: file into an executable. For example, if you place this text in a file:
14390:
14391: @example
14392: #! /usr/local/bin/gforth
14393:
14394: ." Hello, world" CR
14395: bye
14396: @end example
14397:
14398: @noindent
14399: and then make the file executable (chmod +x in Unix), you can run it
14400: directly from the command line. The sequence @code{#!} is used in two
14401: ways; firstly, it is recognised as a ``magic sequence'' by the operating
14402: system@footnote{The Unix kernel actually recognises two types of files:
14403: executable files and files of data, where the data is processed by an
14404: interpreter that is specified on the ``interpreter line'' -- the first
14405: line of the file, starting with the sequence #!. There may be a small
14406: limit (e.g., 32) on the number of characters that may be specified on
14407: the interpreter line.} secondly it is treated as a comment character by
14408: Gforth. Because of the second usage, a space is required between
14409: @code{#!} and the path to the executable (moreover, some Unixes
14410: require the sequence @code{#! /}).
14411:
14412: The disadvantage of this latter technique, compared with using
14413: @file{gforthmi}, is that it is slightly slower; the Forth source code is
14414: compiled on-the-fly, each time the program is invoked.
14415:
14416: doc-#!
14417:
14418:
14419: @node Modifying the Startup Sequence, , Running Image Files, Image Files
14420: @section Modifying the Startup Sequence
14421: @cindex startup sequence for image file
14422: @cindex image file initialization sequence
14423: @cindex initialization sequence of image file
14424:
14425: You can add your own initialization to the startup sequence through the
14426: deferred word @code{'cold}. @code{'cold} is invoked just before the
14427: image-specific command line processing (i.e., loading files and
14428: evaluating (@code{-e}) strings) starts.
14429:
14430: A sequence for adding your initialization usually looks like this:
14431:
14432: @example
14433: :noname
14434: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
14435: ... \ your stuff
14436: ; IS 'cold
14437: @end example
14438:
14439: @cindex turnkey image files
14440: @cindex image file, turnkey applications
14441: You can make a turnkey image by letting @code{'cold} execute a word
14442: (your turnkey application) that never returns; instead, it exits Gforth
14443: via @code{bye} or @code{throw}.
14444:
14445: @cindex command-line arguments, access
14446: @cindex arguments on the command line, access
14447: You can access the (image-specific) command-line arguments through the
14448: variables @code{argc} and @code{argv}. @code{arg} provides convenient
14449: access to @code{argv}.
14450:
14451: If @code{'cold} exits normally, Gforth processes the command-line
14452: arguments as files to be loaded and strings to be evaluated. Therefore,
14453: @code{'cold} should remove the arguments it has used in this case.
14454:
14455:
14456:
14457: doc-'cold
14458: doc-argc
14459: doc-argv
14460: doc-arg
14461:
14462:
14463:
14464: @c ******************************************************************
14465: @node Engine, Binding to System Library, Image Files, Top
14466: @chapter Engine
14467: @cindex engine
14468: @cindex virtual machine
14469:
14470: Reading this chapter is not necessary for programming with Gforth. It
14471: may be helpful for finding your way in the Gforth sources.
14472:
14473: The ideas in this section have also been published in Bernd Paysan,
14474: @cite{ANS fig/GNU/??? Forth} (in German), Forth-Tagung '93 and M. Anton
14475: Ertl, @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
14476: Portable Forth Engine}}, EuroForth '93.
14477:
14478: @menu
14479: * Portability::
14480: * Threading::
14481: * Primitives::
14482: * Performance::
14483: @end menu
14484:
14485: @node Portability, Threading, Engine, Engine
14486: @section Portability
14487: @cindex engine portability
14488:
14489: An important goal of the Gforth Project is availability across a wide
14490: range of personal machines. fig-Forth, and, to a lesser extent, F83,
14491: achieved this goal by manually coding the engine in assembly language
14492: for several then-popular processors. This approach is very
14493: labor-intensive and the results are short-lived due to progress in
14494: computer architecture.
14495:
14496: @cindex C, using C for the engine
14497: Others have avoided this problem by coding in C, e.g., Mitch Bradley
14498: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
14499: particularly popular for UNIX-based Forths due to the large variety of
14500: architectures of UNIX machines. Unfortunately an implementation in C
14501: does not mix well with the goals of efficiency and with using
14502: traditional techniques: Indirect or direct threading cannot be expressed
14503: in C, and switch threading, the fastest technique available in C, is
14504: significantly slower. Another problem with C is that it is very
14505: cumbersome to express double integer arithmetic.
14506:
14507: @cindex GNU C for the engine
14508: @cindex long long
14509: Fortunately, there is a portable language that does not have these
14510: limitations: GNU C, the version of C processed by the GNU C compiler
14511: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
14512: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
14513: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
14514: threading possible, its @code{long long} type (@pxref{Long Long, ,
14515: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
14516: double numbers@footnote{Unfortunately, long longs are not implemented
14517: properly on all machines (e.g., on alpha-osf1, long longs are only 64
14518: bits, the same size as longs (and pointers), but they should be twice as
14519: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
14520: C Manual}). So, we had to implement doubles in C after all. Still, on
14521: most machines we can use long longs and achieve better performance than
14522: with the emulation package.}. GNU C is available for free on all
14523: important (and many unimportant) UNIX machines, VMS, 80386s running
14524: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
14525: on all these machines.
14526:
14527: Writing in a portable language has the reputation of producing code that
14528: is slower than assembly. For our Forth engine we repeatedly looked at
14529: the code produced by the compiler and eliminated most compiler-induced
14530: inefficiencies by appropriate changes in the source code.
14531:
14532: @cindex explicit register declarations
14533: @cindex --enable-force-reg, configuration flag
14534: @cindex -DFORCE_REG
14535: However, register allocation cannot be portably influenced by the
14536: programmer, leading to some inefficiencies on register-starved
14537: machines. We use explicit register declarations (@pxref{Explicit Reg
14538: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
14539: improve the speed on some machines. They are turned on by using the
14540: configuration flag @code{--enable-force-reg} (@code{gcc} switch
14541: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
14542: machine, but also on the compiler version: On some machines some
14543: compiler versions produce incorrect code when certain explicit register
14544: declarations are used. So by default @code{-DFORCE_REG} is not used.
14545:
14546: @node Threading, Primitives, Portability, Engine
14547: @section Threading
14548: @cindex inner interpreter implementation
14549: @cindex threaded code implementation
14550:
14551: @cindex labels as values
14552: GNU C's labels as values extension (available since @code{gcc-2.0},
14553: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
14554: makes it possible to take the address of @i{label} by writing
14555: @code{&&@i{label}}. This address can then be used in a statement like
14556: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
14557: @code{goto x}.
14558:
14559: @cindex @code{NEXT}, indirect threaded
14560: @cindex indirect threaded inner interpreter
14561: @cindex inner interpreter, indirect threaded
14562: With this feature an indirect threaded @code{NEXT} looks like:
14563: @example
14564: cfa = *ip++;
14565: ca = *cfa;
14566: goto *ca;
14567: @end example
14568: @cindex instruction pointer
14569: For those unfamiliar with the names: @code{ip} is the Forth instruction
14570: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
14571: execution token and points to the code field of the next word to be
14572: executed; The @code{ca} (code address) fetched from there points to some
14573: executable code, e.g., a primitive or the colon definition handler
14574: @code{docol}.
14575:
14576: @cindex @code{NEXT}, direct threaded
14577: @cindex direct threaded inner interpreter
14578: @cindex inner interpreter, direct threaded
14579: Direct threading is even simpler:
14580: @example
14581: ca = *ip++;
14582: goto *ca;
14583: @end example
14584:
14585: Of course we have packaged the whole thing neatly in macros called
14586: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
14587:
14588: @menu
14589: * Scheduling::
14590: * Direct or Indirect Threaded?::
14591: * DOES>::
14592: @end menu
14593:
14594: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
14595: @subsection Scheduling
14596: @cindex inner interpreter optimization
14597:
14598: There is a little complication: Pipelined and superscalar processors,
14599: i.e., RISC and some modern CISC machines can process independent
14600: instructions while waiting for the results of an instruction. The
14601: compiler usually reorders (schedules) the instructions in a way that
14602: achieves good usage of these delay slots. However, on our first tries
14603: the compiler did not do well on scheduling primitives. E.g., for
14604: @code{+} implemented as
14605: @example
14606: n=sp[0]+sp[1];
14607: sp++;
14608: sp[0]=n;
14609: NEXT;
14610: @end example
14611: the @code{NEXT} comes strictly after the other code, i.e., there is
14612: nearly no scheduling. After a little thought the problem becomes clear:
14613: The compiler cannot know that @code{sp} and @code{ip} point to different
14614: addresses (and the version of @code{gcc} we used would not know it even
14615: if it was possible), so it could not move the load of the cfa above the
14616: store to the TOS. Indeed the pointers could be the same, if code on or
14617: very near the top of stack were executed. In the interest of speed we
14618: chose to forbid this probably unused ``feature'' and helped the compiler
14619: in scheduling: @code{NEXT} is divided into several parts:
14620: @code{NEXT_P0}, @code{NEXT_P1} and @code{NEXT_P2}). @code{+} now looks
14621: like:
14622: @example
14623: NEXT_P0;
14624: n=sp[0]+sp[1];
14625: sp++;
14626: NEXT_P1;
14627: sp[0]=n;
14628: NEXT_P2;
14629: @end example
14630:
14631: There are various schemes that distribute the different operations of
14632: NEXT between these parts in several ways; in general, different schemes
14633: perform best on different processors. We use a scheme for most
14634: architectures that performs well for most processors of this
14635: architecture; in the furture we may switch to benchmarking and chosing
14636: the scheme on installation time.
14637:
14638:
14639: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
14640: @subsection Direct or Indirect Threaded?
14641: @cindex threading, direct or indirect?
14642:
14643: @cindex -DDIRECT_THREADED
14644: Both! After packaging the nasty details in macro definitions we
14645: realized that we could switch between direct and indirect threading by
14646: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
14647: defining a few machine-specific macros for the direct-threading case.
14648: On the Forth level we also offer access words that hide the
14649: differences between the threading methods (@pxref{Threading Words}).
14650:
14651: Indirect threading is implemented completely machine-independently.
14652: Direct threading needs routines for creating jumps to the executable
14653: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
14654: machine-dependent, but they do not amount to many source lines. Therefore,
14655: even porting direct threading to a new machine requires little effort.
14656:
14657: @cindex --enable-indirect-threaded, configuration flag
14658: @cindex --enable-direct-threaded, configuration flag
14659: The default threading method is machine-dependent. You can enforce a
14660: specific threading method when building Gforth with the configuration
14661: flag @code{--enable-direct-threaded} or
14662: @code{--enable-indirect-threaded}. Note that direct threading is not
14663: supported on all machines.
14664:
14665: @node DOES>, , Direct or Indirect Threaded?, Threading
14666: @subsection DOES>
14667: @cindex @code{DOES>} implementation
14668:
14669: @cindex @code{dodoes} routine
14670: @cindex @code{DOES>}-code
14671: One of the most complex parts of a Forth engine is @code{dodoes}, i.e.,
14672: the chunk of code executed by every word defined by a
14673: @code{CREATE}...@code{DOES>} pair. The main problem here is: How to find
14674: the Forth code to be executed, i.e. the code after the
14675: @code{DOES>} (the @code{DOES>}-code)? There are two solutions:
14676:
14677: In fig-Forth the code field points directly to the @code{dodoes} and the
14678: @code{DOES>}-code address is stored in the cell after the code address (i.e. at
14679: @code{@i{CFA} cell+}). It may seem that this solution is illegal in
14680: the Forth-79 and all later standards, because in fig-Forth this address
14681: lies in the body (which is illegal in these standards). However, by
14682: making the code field larger for all words this solution becomes legal
14683: again. We use this approach for the indirect threaded version and for
14684: direct threading on some machines. Leaving a cell unused in most words
14685: is a bit wasteful, but on the machines we are targeting this is hardly a
14686: problem. The other reason for having a code field size of two cells is
14687: to avoid having different image files for direct and indirect threaded
14688: systems (direct threaded systems require two-cell code fields on many
14689: machines).
14690:
14691: @cindex @code{DOES>}-handler
14692: The other approach is that the code field points or jumps to the cell
14693: after @code{DOES>}. In this variant there is a jump to @code{dodoes} at
14694: this address (the @code{DOES>}-handler). @code{dodoes} can then get the
14695: @code{DOES>}-code address by computing the code address, i.e., the address of
14696: the jump to @code{dodoes}, and add the length of that jump field. A variant of
14697: this is to have a call to @code{dodoes} after the @code{DOES>}; then the
14698: return address (which can be found in the return register on RISCs) is
14699: the @code{DOES>}-code address. Since the two cells available in the code field
14700: are used up by the jump to the code address in direct threading on many
14701: architectures, we use this approach for direct threading on these
14702: architectures. We did not want to add another cell to the code field.
14703:
14704: @node Primitives, Performance, Threading, Engine
14705: @section Primitives
14706: @cindex primitives, implementation
14707: @cindex virtual machine instructions, implementation
14708:
14709: @menu
14710: * Automatic Generation::
14711: * TOS Optimization::
14712: * Produced code::
14713: @end menu
14714:
14715: @node Automatic Generation, TOS Optimization, Primitives, Primitives
14716: @subsection Automatic Generation
14717: @cindex primitives, automatic generation
14718:
14719: @cindex @file{prims2x.fs}
14720: Since the primitives are implemented in a portable language, there is no
14721: longer any need to minimize the number of primitives. On the contrary,
14722: having many primitives has an advantage: speed. In order to reduce the
14723: number of errors in primitives and to make programming them easier, we
14724: provide a tool, the primitive generator (@file{prims2x.fs}), that
14725: automatically generates most (and sometimes all) of the C code for a
14726: primitive from the stack effect notation. The source for a primitive
14727: has the following form:
14728:
14729: @cindex primitive source format
14730: @format
14731: @i{Forth-name} ( @i{stack-effect} ) @i{category} [@i{pronounc.}]
14732: [@code{""}@i{glossary entry}@code{""}]
14733: @i{C code}
14734: [@code{:}
14735: @i{Forth code}]
14736: @end format
14737:
14738: The items in brackets are optional. The category and glossary fields
14739: are there for generating the documentation, the Forth code is there
14740: for manual implementations on machines without GNU C. E.g., the source
14741: for the primitive @code{+} is:
14742: @example
14743: + ( n1 n2 -- n ) core plus
14744: n = n1+n2;
14745: @end example
14746:
14747: This looks like a specification, but in fact @code{n = n1+n2} is C
14748: code. Our primitive generation tool extracts a lot of information from
14749: the stack effect notations@footnote{We use a one-stack notation, even
14750: though we have separate data and floating-point stacks; The separate
14751: notation can be generated easily from the unified notation.}: The number
14752: of items popped from and pushed on the stack, their type, and by what
14753: name they are referred to in the C code. It then generates a C code
14754: prelude and postlude for each primitive. The final C code for @code{+}
14755: looks like this:
14756:
14757: @example
14758: I_plus: /* + ( n1 n2 -- n ) */ /* label, stack effect */
14759: /* */ /* documentation */
14760: NAME("+") /* debugging output (with -DDEBUG) */
14761: @{
14762: DEF_CA /* definition of variable ca (indirect threading) */
14763: Cell n1; /* definitions of variables */
14764: Cell n2;
14765: Cell n;
14766: NEXT_P0; /* NEXT part 0 */
14767: n1 = (Cell) sp[1]; /* input */
14768: n2 = (Cell) TOS;
14769: sp += 1; /* stack adjustment */
14770: @{
14771: n = n1+n2; /* C code taken from the source */
14772: @}
14773: NEXT_P1; /* NEXT part 1 */
14774: TOS = (Cell)n; /* output */
14775: NEXT_P2; /* NEXT part 2 */
14776: @}
14777: @end example
14778:
14779: This looks long and inefficient, but the GNU C compiler optimizes quite
14780: well and produces optimal code for @code{+} on, e.g., the R3000 and the
14781: HP RISC machines: Defining the @code{n}s does not produce any code, and
14782: using them as intermediate storage also adds no cost.
14783:
14784: There are also other optimizations that are not illustrated by this
14785: example: assignments between simple variables are usually for free (copy
14786: propagation). If one of the stack items is not used by the primitive
14787: (e.g. in @code{drop}), the compiler eliminates the load from the stack
14788: (dead code elimination). On the other hand, there are some things that
14789: the compiler does not do, therefore they are performed by
14790: @file{prims2x.fs}: The compiler does not optimize code away that stores
14791: a stack item to the place where it just came from (e.g., @code{over}).
14792:
14793: While programming a primitive is usually easy, there are a few cases
14794: where the programmer has to take the actions of the generator into
14795: account, most notably @code{?dup}, but also words that do not (always)
14796: fall through to @code{NEXT}.
14797:
14798: @node TOS Optimization, Produced code, Automatic Generation, Primitives
14799: @subsection TOS Optimization
14800: @cindex TOS optimization for primitives
14801: @cindex primitives, keeping the TOS in a register
14802:
14803: An important optimization for stack machine emulators, e.g., Forth
14804: engines, is keeping one or more of the top stack items in
14805: registers. If a word has the stack effect @i{in1}...@i{inx} @code{--}
14806: @i{out1}...@i{outy}, keeping the top @i{n} items in registers
14807: @itemize @bullet
14808: @item
14809: is better than keeping @i{n-1} items, if @i{x>=n} and @i{y>=n},
14810: due to fewer loads from and stores to the stack.
14811: @item is slower than keeping @i{n-1} items, if @i{x<>y} and @i{x<n} and
14812: @i{y<n}, due to additional moves between registers.
14813: @end itemize
14814:
14815: @cindex -DUSE_TOS
14816: @cindex -DUSE_NO_TOS
14817: In particular, keeping one item in a register is never a disadvantage,
14818: if there are enough registers. Keeping two items in registers is a
14819: disadvantage for frequent words like @code{?branch}, constants,
14820: variables, literals and @code{i}. Therefore our generator only produces
14821: code that keeps zero or one items in registers. The generated C code
14822: covers both cases; the selection between these alternatives is made at
14823: C-compile time using the switch @code{-DUSE_TOS}. @code{TOS} in the C
14824: code for @code{+} is just a simple variable name in the one-item case,
14825: otherwise it is a macro that expands into @code{sp[0]}. Note that the
14826: GNU C compiler tries to keep simple variables like @code{TOS} in
14827: registers, and it usually succeeds, if there are enough registers.
14828:
14829: @cindex -DUSE_FTOS
14830: @cindex -DUSE_NO_FTOS
14831: The primitive generator performs the TOS optimization for the
14832: floating-point stack, too (@code{-DUSE_FTOS}). For floating-point
14833: operations the benefit of this optimization is even larger:
14834: floating-point operations take quite long on most processors, but can be
14835: performed in parallel with other operations as long as their results are
14836: not used. If the FP-TOS is kept in a register, this works. If
14837: it is kept on the stack, i.e., in memory, the store into memory has to
14838: wait for the result of the floating-point operation, lengthening the
14839: execution time of the primitive considerably.
14840:
14841: The TOS optimization makes the automatic generation of primitives a
14842: bit more complicated. Just replacing all occurrences of @code{sp[0]} by
14843: @code{TOS} is not sufficient. There are some special cases to
14844: consider:
14845: @itemize @bullet
14846: @item In the case of @code{dup ( w -- w w )} the generator must not
14847: eliminate the store to the original location of the item on the stack,
14848: if the TOS optimization is turned on.
14849: @item Primitives with stack effects of the form @code{--}
14850: @i{out1}...@i{outy} must store the TOS to the stack at the start.
14851: Likewise, primitives with the stack effect @i{in1}...@i{inx} @code{--}
14852: must load the TOS from the stack at the end. But for the null stack
14853: effect @code{--} no stores or loads should be generated.
14854: @end itemize
14855:
14856: @node Produced code, , TOS Optimization, Primitives
14857: @subsection Produced code
14858: @cindex primitives, assembly code listing
14859:
14860: @cindex @file{engine.s}
14861: To see what assembly code is produced for the primitives on your machine
14862: with your compiler and your flag settings, type @code{make engine.s} and
14863: look at the resulting file @file{engine.s}. Alternatively, you can also
14864: disassemble the code of primitives with @code{see} on some architectures.
14865:
14866: @node Performance, , Primitives, Engine
14867: @section Performance
14868: @cindex performance of some Forth interpreters
14869: @cindex engine performance
14870: @cindex benchmarking Forth systems
14871: @cindex Gforth performance
14872:
14873: On RISCs the Gforth engine is very close to optimal; i.e., it is usually
14874: impossible to write a significantly faster engine.
14875:
14876: On register-starved machines like the 386 architecture processors
14877: improvements are possible, because @code{gcc} does not utilize the
14878: registers as well as a human, even with explicit register declarations;
14879: e.g., Bernd Beuster wrote a Forth system fragment in assembly language
14880: and hand-tuned it for the 486; this system is 1.19 times faster on the
14881: Sieve benchmark on a 486DX2/66 than Gforth compiled with
14882: @code{gcc-2.6.3} with @code{-DFORCE_REG}. The situation has improved
14883: with gcc-2.95 and gforth-0.4.9; now the most important virtual machine
14884: registers fit in real registers (and we can even afford to use the TOS
14885: optimization), resulting in a speedup of 1.14 on the sieve over the
14886: earlier results.
14887:
14888: @cindex Win32Forth performance
14889: @cindex NT Forth performance
14890: @cindex eforth performance
14891: @cindex ThisForth performance
14892: @cindex PFE performance
14893: @cindex TILE performance
14894: The potential advantage of assembly language implementations is not
14895: necessarily realized in complete Forth systems: We compared Gforth-0.4.9
14896: (direct threaded, compiled with @code{gcc-2.95.1} and
14897: @code{-DFORCE_REG}) with Win32Forth 1.2093 (newer versions are
14898: reportedly much faster), LMI's NT Forth (Beta, May 1994) and Eforth
14899: (with and without peephole (aka pinhole) optimization of the threaded
14900: code); all these systems were written in assembly language. We also
14901: compared Gforth with three systems written in C: PFE-0.9.14 (compiled
14902: with @code{gcc-2.6.3} with the default configuration for Linux:
14903: @code{-O2 -fomit-frame-pointer -DUSE_REGS -DUNROLL_NEXT}), ThisForth
14904: Beta (compiled with @code{gcc-2.6.3 -O3 -fomit-frame-pointer}; ThisForth
14905: employs peephole optimization of the threaded code) and TILE (compiled
14906: with @code{make opt}). We benchmarked Gforth, PFE, ThisForth and TILE on
14907: a 486DX2/66 under Linux. Kenneth O'Heskin kindly provided the results
14908: for Win32Forth and NT Forth on a 486DX2/66 with similar memory
14909: performance under Windows NT. Marcel Hendrix ported Eforth to Linux,
14910: then extended it to run the benchmarks, added the peephole optimizer,
14911: ran the benchmarks and reported the results.
14912:
14913: We used four small benchmarks: the ubiquitous Sieve; bubble-sorting and
14914: matrix multiplication come from the Stanford integer benchmarks and have
14915: been translated into Forth by Martin Fraeman; we used the versions
14916: included in the TILE Forth package, but with bigger data set sizes; and
14917: a recursive Fibonacci number computation for benchmarking calling
14918: performance. The following table shows the time taken for the benchmarks
14919: scaled by the time taken by Gforth (in other words, it shows the speedup
14920: factor that Gforth achieved over the other systems).
14921:
14922: @example
14923: relative Win32- NT eforth This-
14924: time Gforth Forth Forth eforth +opt PFE Forth TILE
14925: sieve 1.00 1.60 1.32 1.60 0.98 1.82 3.67 9.91
14926: bubble 1.00 1.55 1.66 1.75 1.04 1.78 4.58
14927: matmul 1.00 1.71 1.57 1.69 0.86 1.83 4.74
14928: fib 1.00 1.76 1.54 1.41 1.00 2.01 3.45 4.96
14929: @end example
14930:
14931: You may be quite surprised by the good performance of Gforth when
14932: compared with systems written in assembly language. One important reason
14933: for the disappointing performance of these other systems is probably
14934: that they are not written optimally for the 486 (e.g., they use the
14935: @code{lods} instruction). In addition, Win32Forth uses a comfortable,
14936: but costly method for relocating the Forth image: like @code{cforth}, it
14937: computes the actual addresses at run time, resulting in two address
14938: computations per @code{NEXT} (@pxref{Image File Background}).
14939:
14940: Only Eforth with the peephole optimizer performs comparable to
14941: Gforth. The speedups achieved with peephole optimization of threaded
14942: code are quite remarkable. Adding a peephole optimizer to Gforth should
14943: cause similar speedups.
14944:
14945: The speedup of Gforth over PFE, ThisForth and TILE can be easily
14946: explained with the self-imposed restriction of the latter systems to
14947: standard C, which makes efficient threading impossible (however, the
14948: measured implementation of PFE uses a GNU C extension: @pxref{Global Reg
14949: Vars, , Defining Global Register Variables, gcc.info, GNU C Manual}).
14950: Moreover, current C compilers have a hard time optimizing other aspects
14951: of the ThisForth and the TILE source.
14952:
14953: The performance of Gforth on 386 architecture processors varies widely
14954: with the version of @code{gcc} used. E.g., @code{gcc-2.5.8} failed to
14955: allocate any of the virtual machine registers into real machine
14956: registers by itself and would not work correctly with explicit register
14957: declarations, giving a 1.5 times slower engine (on a 486DX2/66 running
14958: the Sieve) than the one measured above.
14959:
14960: Note that there have been several releases of Win32Forth since the
14961: release presented here, so the results presented above may have little
14962: predictive value for the performance of Win32Forth today (results for
14963: the current release on an i486DX2/66 are welcome).
14964:
14965: @cindex @file{Benchres}
14966: In
14967: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl&maierhofer95.ps.gz,
14968: Translating Forth to Efficient C}} by M. Anton Ertl and Martin
14969: Maierhofer (presented at EuroForth '95), an indirect threaded version of
14970: Gforth is compared with Win32Forth, NT Forth, PFE, ThisForth, and
14971: several native code systems; that version of Gforth is slower on a 486
14972: than the direct threaded version used here. You can find a newer version
14973: of these measurements at
14974: @uref{http://www.complang.tuwien.ac.at/forth/performance.html}. You can
14975: find numbers for Gforth on various machines in @file{Benchres}.
14976:
14977: @c ******************************************************************
14978: @node Binding to System Library, Cross Compiler, Engine, Top
14979: @chapter Binding to System Library
14980:
14981: @node Cross Compiler, Bugs, Binding to System Library, Top
14982: @chapter Cross Compiler
14983: @cindex @file{cross.fs}
14984: @cindex cross-compiler
14985: @cindex metacompiler
14986: @cindex target compiler
14987:
14988: The cross compiler is used to bootstrap a Forth kernel. Since Gforth is
14989: mostly written in Forth, including crucial parts like the outer
14990: interpreter and compiler, it needs compiled Forth code to get
14991: started. The cross compiler allows to create new images for other
14992: architectures, even running under another Forth system.
14993:
14994: @menu
14995: * Using the Cross Compiler::
14996: * How the Cross Compiler Works::
14997: @end menu
14998:
14999: @node Using the Cross Compiler, How the Cross Compiler Works, Cross Compiler, Cross Compiler
15000: @section Using the Cross Compiler
15001:
15002: The cross compiler uses a language that resembles Forth, but isn't. The
15003: main difference is that you can execute Forth code after definition,
15004: while you usually can't execute the code compiled by cross, because the
15005: code you are compiling is typically for a different computer than the
15006: one you are compiling on.
15007:
15008: @c anton: This chapter is somewhat different from waht I would expect: I
15009: @c would expect an explanation of the cross language and how to create an
15010: @c application image with it. The section explains some aspects of
15011: @c creating a Gforth kernel.
15012:
15013: The Makefile is already set up to allow you to create kernels for new
15014: architectures with a simple make command. The generic kernels using the
15015: GCC compiled virtual machine are created in the normal build process
15016: with @code{make}. To create a embedded Gforth executable for e.g. the
15017: 8086 processor (running on a DOS machine), type
15018:
15019: @example
15020: make kernl-8086.fi
15021: @end example
15022:
15023: This will use the machine description from the @file{arch/8086}
15024: directory to create a new kernel. A machine file may look like that:
15025:
15026: @example
15027: \ Parameter for target systems 06oct92py
15028:
15029: 4 Constant cell \ cell size in bytes
15030: 2 Constant cell<< \ cell shift to bytes
15031: 5 Constant cell>bit \ cell shift to bits
15032: 8 Constant bits/char \ bits per character
15033: 8 Constant bits/byte \ bits per byte [default: 8]
15034: 8 Constant float \ bytes per float
15035: 8 Constant /maxalign \ maximum alignment in bytes
15036: false Constant bigendian \ byte order
15037: ( true=big, false=little )
15038:
15039: include machpc.fs \ feature list
15040: @end example
15041:
15042: This part is obligatory for the cross compiler itself, the feature list
15043: is used by the kernel to conditionally compile some features in and out,
15044: depending on whether the target supports these features.
15045:
15046: There are some optional features, if you define your own primitives,
15047: have an assembler, or need special, nonstandard preparation to make the
15048: boot process work. @code{asm-include} includes an assembler,
15049: @code{prims-include} includes primitives, and @code{>boot} prepares for
15050: booting.
15051:
15052: @example
15053: : asm-include ." Include assembler" cr
15054: s" arch/8086/asm.fs" included ;
15055:
15056: : prims-include ." Include primitives" cr
15057: s" arch/8086/prim.fs" included ;
15058:
15059: : >boot ." Prepare booting" cr
15060: s" ' boot >body into-forth 1+ !" evaluate ;
15061: @end example
15062:
15063: These words are used as sort of macro during the cross compilation in
15064: the file @file{kernel/main.fs}. Instead of using these macros, it would
15065: be possible --- but more complicated --- to write a new kernel project
15066: file, too.
15067:
15068: @file{kernel/main.fs} expects the machine description file name on the
15069: stack; the cross compiler itself (@file{cross.fs}) assumes that either
15070: @code{mach-file} leaves a counted string on the stack, or
15071: @code{machine-file} leaves an address, count pair of the filename on the
15072: stack.
15073:
15074: The feature list is typically controlled using @code{SetValue}, generic
15075: files that are used by several projects can use @code{DefaultValue}
15076: instead. Both functions work like @code{Value}, when the value isn't
15077: defined, but @code{SetValue} works like @code{to} if the value is
15078: defined, and @code{DefaultValue} doesn't set anything, if the value is
15079: defined.
15080:
15081: @example
15082: \ generic mach file for pc gforth 03sep97jaw
15083:
15084: true DefaultValue NIL \ relocating
15085:
15086: >ENVIRON
15087:
15088: true DefaultValue file \ controls the presence of the
15089: \ file access wordset
15090: true DefaultValue OS \ flag to indicate a operating system
15091:
15092: true DefaultValue prims \ true: primitives are c-code
15093:
15094: true DefaultValue floating \ floating point wordset is present
15095:
15096: true DefaultValue glocals \ gforth locals are present
15097: \ will be loaded
15098: true DefaultValue dcomps \ double number comparisons
15099:
15100: true DefaultValue hash \ hashing primitives are loaded/present
15101:
15102: true DefaultValue xconds \ used together with glocals,
15103: \ special conditionals supporting gforths'
15104: \ local variables
15105: true DefaultValue header \ save a header information
15106:
15107: true DefaultValue backtrace \ enables backtrace code
15108:
15109: false DefaultValue ec
15110: false DefaultValue crlf
15111:
15112: cell 2 = [IF] &32 [ELSE] &256 [THEN] KB DefaultValue kernel-size
15113:
15114: &16 KB DefaultValue stack-size
15115: &15 KB &512 + DefaultValue fstack-size
15116: &15 KB DefaultValue rstack-size
15117: &14 KB &512 + DefaultValue lstack-size
15118: @end example
15119:
15120: @node How the Cross Compiler Works, , Using the Cross Compiler, Cross Compiler
15121: @section How the Cross Compiler Works
15122:
15123: @node Bugs, Origin, Cross Compiler, Top
15124: @appendix Bugs
15125: @cindex bug reporting
15126:
15127: Known bugs are described in the file @file{BUGS} in the Gforth distribution.
15128:
15129: If you find a bug, please submit a bug report through
15130: @uref{https://savannah.gnu.org/bugs/?func=addbug&group=gforth}.
15131:
15132: @itemize @bullet
15133: @item
15134: A program (or a sequence of keyboard commands) that reproduces the bug.
15135: @item
15136: A description of what you think constitutes the buggy behaviour.
15137: @item
15138: The Gforth version used (it is announced at the start of an
15139: interactive Gforth session).
15140: @item
15141: The machine and operating system (on Unix
15142: systems @code{uname -a} will report this information).
15143: @item
15144: The installation options (you can find the configure options at the
15145: start of @file{config.status}) and configuration (@code{configure}
15146: output or @file{config.cache}).
15147: @item
15148: A complete list of changes (if any) you (or your installer) have made to the
15149: Gforth sources.
15150: @end itemize
15151:
15152: For a thorough guide on reporting bugs read @ref{Bug Reporting, , How
15153: to Report Bugs, gcc.info, GNU C Manual}.
15154:
15155:
15156: @node Origin, Forth-related information, Bugs, Top
15157: @appendix Authors and Ancestors of Gforth
15158:
15159: @section Authors and Contributors
15160: @cindex authors of Gforth
15161: @cindex contributors to Gforth
15162:
15163: The Gforth project was started in mid-1992 by Bernd Paysan and Anton
15164: Ertl. The third major author was Jens Wilke. Neal Crook contributed a
15165: lot to the manual. Assemblers and disassemblers were contributed by
15166: Andrew McKewan, Christian Pirker, and Bernd Thallner. Lennart Benschop
15167: (who was one of Gforth's first users, in mid-1993) and Stuart Ramsden
15168: inspired us with their continuous feedback. Lennart Benshop contributed
15169: @file{glosgen.fs}, while Stuart Ramsden has been working on automatic
15170: support for calling C libraries. Helpful comments also came from Paul
15171: Kleinrubatscher, Christian Pirker, Dirk Zoller, Marcel Hendrix, John
15172: Wavrik, Barrie Stott, Marc de Groot, Jorge Acerada, Bruce Hoyt, and
15173: Robert Epprecht. Since the release of Gforth-0.2.1 there were also
15174: helpful comments from many others; thank you all, sorry for not listing
15175: you here (but digging through my mailbox to extract your names is on my
15176: to-do list).
15177:
15178: Gforth also owes a lot to the authors of the tools we used (GCC, CVS,
15179: and autoconf, among others), and to the creators of the Internet: Gforth
15180: was developed across the Internet, and its authors did not meet
15181: physically for the first 4 years of development.
15182:
15183: @section Pedigree
15184: @cindex pedigree of Gforth
15185:
15186: Gforth descends from bigFORTH (1993) and fig-Forth. Of course, a
15187: significant part of the design of Gforth was prescribed by ANS Forth.
15188:
15189: Bernd Paysan wrote bigFORTH, a descendent from TurboForth, an unreleased
15190: 32 bit native code version of VolksForth for the Atari ST, written
15191: mostly by Dietrich Weineck.
15192:
15193: VolksForth was written by Klaus Schleisiek, Bernd Pennemann, Georg
15194: Rehfeld and Dietrich Weineck for the C64 (called UltraForth there) in
15195: the mid-80s and ported to the Atari ST in 1986. It descends from F83.
15196:
15197: Henry Laxen and Mike Perry wrote F83 as a model implementation of the
15198: Forth-83 standard. !! Pedigree? When?
15199:
15200: A team led by Bill Ragsdale implemented fig-Forth on many processors in
15201: 1979. Robert Selzer and Bill Ragsdale developed the original
15202: implementation of fig-Forth for the 6502 based on microForth.
15203:
15204: The principal architect of microForth was Dean Sanderson. microForth was
15205: FORTH, Inc.'s first off-the-shelf product. It was developed in 1976 for
15206: the 1802, and subsequently implemented on the 8080, the 6800 and the
15207: Z80.
15208:
15209: All earlier Forth systems were custom-made, usually by Charles Moore,
15210: who discovered (as he puts it) Forth during the late 60s. The first full
15211: Forth existed in 1971.
15212:
15213: A part of the information in this section comes from
15214: @cite{@uref{http://www.forth.com/Content/History/History1.htm,The
15215: Evolution of Forth}} by Elizabeth D. Rather, Donald R. Colburn and
15216: Charles H. Moore, presented at the HOPL-II conference and preprinted in
15217: SIGPLAN Notices 28(3), 1993. You can find more historical and
15218: genealogical information about Forth there.
15219:
15220: @c ------------------------------------------------------------------
15221: @node Forth-related information, Word Index, Origin, Top
15222: @appendix Other Forth-related information
15223: @cindex Forth-related information
15224:
15225: @c anton: I threw most of this stuff out, because it can be found through
15226: @c the FAQ and the FAQ is more likely to be up-to-date.
15227:
15228: @cindex comp.lang.forth
15229: @cindex frequently asked questions
15230: There is an active news group (comp.lang.forth) discussing Forth
15231: (including Gforth) and Forth-related issues. Its
15232: @uref{http://www.complang.tuwien.ac.at/forth/faq/faq-general-2.html,FAQs}
15233: (frequently asked questions and their answers) contains a lot of
15234: information on Forth. You should read it before posting to
15235: comp.lang.forth.
15236:
15237: The ANS Forth standard is most usable in its
15238: @uref{http://www.taygeta.com/forth/dpans.html, HTML form}.
15239:
15240: @c ------------------------------------------------------------------
15241: @node Word Index, Concept Index, Forth-related information, Top
15242: @unnumbered Word Index
15243:
15244: This index is a list of Forth words that have ``glossary'' entries
15245: within this manual. Each word is listed with its stack effect and
15246: wordset.
15247:
15248: @printindex fn
15249:
15250: @c anton: the name index seems superfluous given the word and concept indices.
15251:
15252: @c @node Name Index, Concept Index, Word Index, Top
15253: @c @unnumbered Name Index
15254:
15255: @c This index is a list of Forth words that have ``glossary'' entries
15256: @c within this manual.
15257:
15258: @c @printindex ky
15259:
15260: @node Concept Index, , Word Index, Top
15261: @unnumbered Concept and Word Index
15262:
15263: Not all entries listed in this index are present verbatim in the
15264: text. This index also duplicates, in abbreviated form, all of the words
15265: listed in the Word Index (only the names are listed for the words here).
15266:
15267: @printindex cp
15268:
15269: @contents
15270: @bye
15271:
15272:
15273:
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>