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 --- The Detailed Node Listing ---
173:
174: Gforth Environment
175:
176: * Invoking Gforth:: Getting in
177: * Leaving Gforth:: Getting out
178: * Command-line editing::
179: * Environment variables:: that affect how Gforth starts up
180: * Gforth Files:: What gets installed and where
181: * Startup speed:: When 35ms is not fast enough ...
182:
183: Forth Tutorial
184:
185: * Starting Gforth Tutorial::
186: * Syntax Tutorial::
187: * Crash Course Tutorial::
188: * Stack Tutorial::
189: * Arithmetics Tutorial::
190: * Stack Manipulation Tutorial::
191: * Using files for Forth code Tutorial::
192: * Comments Tutorial::
193: * Colon Definitions Tutorial::
194: * Decompilation Tutorial::
195: * Stack-Effect Comments Tutorial::
196: * Types Tutorial::
197: * Factoring Tutorial::
198: * Designing the stack effect Tutorial::
199: * Local Variables Tutorial::
200: * Conditional execution Tutorial::
201: * Flags and Comparisons Tutorial::
202: * General Loops Tutorial::
203: * Counted loops Tutorial::
204: * Recursion Tutorial::
205: * Leaving definitions or loops Tutorial::
206: * Return Stack Tutorial::
207: * Memory Tutorial::
208: * Characters and Strings Tutorial::
209: * Alignment Tutorial::
210: * Files Tutorial::
211: * Interpretation and Compilation Semantics and Immediacy Tutorial::
212: * Execution Tokens Tutorial::
213: * Exceptions Tutorial::
214: * Defining Words Tutorial::
215: * Arrays and Records Tutorial::
216: * POSTPONE Tutorial::
217: * Literal Tutorial::
218: * Advanced macros Tutorial::
219: * Compilation Tokens Tutorial::
220: * Wordlists and Search Order Tutorial::
221:
222: An Introduction to ANS Forth
223:
224: * Introducing the Text Interpreter::
225: * Stacks and Postfix notation::
226: * Your first definition::
227: * How does that work?::
228: * Forth is written in Forth::
229: * Review - elements of a Forth system::
230: * Where to go next::
231: * Exercises::
232:
233: Forth Words
234:
235: * Notation::
236: * Case insensitivity::
237: * Comments::
238: * Boolean Flags::
239: * Arithmetic::
240: * Stack Manipulation::
241: * Memory::
242: * Control Structures::
243: * Defining Words::
244: * Interpretation and Compilation Semantics::
245: * Tokens for Words::
246: * Compiling words::
247: * The Text Interpreter::
248: * Word Lists::
249: * Environmental Queries::
250: * Files::
251: * Blocks::
252: * Other I/O::
253: * Locals::
254: * Structures::
255: * Object-oriented Forth::
256: * Programming Tools::
257: * Assembler and Code Words::
258: * Threading Words::
259: * Passing Commands to the OS::
260: * Keeping track of Time::
261: * Miscellaneous Words::
262:
263: Arithmetic
264:
265: * Single precision::
266: * Double precision:: Double-cell integer arithmetic
267: * Bitwise operations::
268: * Numeric comparison::
269: * Mixed precision:: Operations with single and double-cell integers
270: * Floating Point::
271:
272: Stack Manipulation
273:
274: * Data stack::
275: * Floating point stack::
276: * Return stack::
277: * Locals stack::
278: * Stack pointer manipulation::
279:
280: Memory
281:
282: * Memory model::
283: * Dictionary allocation::
284: * Heap Allocation::
285: * Memory Access::
286: * Address arithmetic::
287: * Memory Blocks::
288:
289: Control Structures
290:
291: * Selection:: IF ... ELSE ... ENDIF
292: * Simple Loops:: BEGIN ...
293: * Counted Loops:: DO
294: * Arbitrary control structures::
295: * Calls and returns::
296: * Exception Handling::
297:
298: Defining Words
299:
300: * CREATE::
301: * Variables:: Variables and user variables
302: * Constants::
303: * Values:: Initialised variables
304: * Colon Definitions::
305: * Anonymous Definitions:: Definitions without names
306: * Supplying names:: Passing definition names as strings
307: * User-defined Defining Words::
308: * Deferred words:: Allow forward references
309: * Aliases::
310:
311: User-defined Defining Words
312:
313: * CREATE..DOES> applications::
314: * CREATE..DOES> details::
315: * Advanced does> usage example::
316:
317: Interpretation and Compilation Semantics
318:
319: * Combined words::
320:
321: Tokens for Words
322:
323: * Execution token:: represents execution/interpretation semantics
324: * Compilation token:: represents compilation semantics
325: * Name token:: represents named words
326:
327: Compiling words
328:
329: * Literals:: Compiling data values
330: * Macros:: Compiling words
331:
332: The Text Interpreter
333:
334: * Input Sources::
335: * Number Conversion::
336: * Interpret/Compile states::
337: * Interpreter Directives::
338:
339: Word Lists
340:
341: * Vocabularies::
342: * Why use word lists?::
343: * Word list example::
344:
345: Files
346:
347: * Forth source files::
348: * General files::
349: * Search Paths::
350:
351: Search Paths
352:
353: * Source Search Paths::
354: * General Search Paths::
355:
356: Other I/O
357:
358: * Simple numeric output:: Predefined formats
359: * Formatted numeric output:: Formatted (pictured) output
360: * String Formats:: How Forth stores strings in memory
361: * Displaying characters and strings:: Other stuff
362: * Input:: Input
363:
364: Locals
365:
366: * Gforth locals::
367: * ANS Forth locals::
368:
369: Gforth locals
370:
371: * Where are locals visible by name?::
372: * How long do locals live?::
373: * Locals programming style::
374: * Locals implementation::
375:
376: Structures
377:
378: * Why explicit structure support?::
379: * Structure Usage::
380: * Structure Naming Convention::
381: * Structure Implementation::
382: * Structure Glossary::
383:
384: Object-oriented Forth
385:
386: * Why object-oriented programming?::
387: * Object-Oriented Terminology::
388: * Objects::
389: * OOF::
390: * Mini-OOF::
391: * Comparison with other object models::
392:
393: The @file{objects.fs} model
394:
395: * Properties of the Objects model::
396: * Basic Objects Usage::
397: * The Objects base class::
398: * Creating objects::
399: * Object-Oriented Programming Style::
400: * Class Binding::
401: * Method conveniences::
402: * Classes and Scoping::
403: * Dividing classes::
404: * Object Interfaces::
405: * Objects Implementation::
406: * Objects Glossary::
407:
408: The @file{oof.fs} model
409:
410: * Properties of the OOF model::
411: * Basic OOF Usage::
412: * The OOF base class::
413: * Class Declaration::
414: * Class Implementation::
415:
416: The @file{mini-oof.fs} model
417:
418: * Basic Mini-OOF Usage::
419: * Mini-OOF Example::
420: * Mini-OOF Implementation::
421:
422: Programming Tools
423:
424: * Examining::
425: * Forgetting words::
426: * Debugging:: Simple and quick.
427: * Assertions:: Making your programs self-checking.
428: * Singlestep Debugger:: Executing your program word by word.
429:
430: Assembler and Code Words
431:
432: * Code and ;code::
433: * Common Assembler:: Assembler Syntax
434: * Common Disassembler::
435: * 386 Assembler:: Deviations and special cases
436: * Alpha Assembler:: Deviations and special cases
437: * MIPS assembler:: Deviations and special cases
438: * Other assemblers:: How to write them
439:
440: Tools
441:
442: * ANS Report:: Report the words used, sorted by wordset.
443:
444: ANS conformance
445:
446: * The Core Words::
447: * The optional Block word set::
448: * The optional Double Number word set::
449: * The optional Exception word set::
450: * The optional Facility word set::
451: * The optional File-Access word set::
452: * The optional Floating-Point word set::
453: * The optional Locals word set::
454: * The optional Memory-Allocation word set::
455: * The optional Programming-Tools word set::
456: * The optional Search-Order word set::
457:
458: The Core Words
459:
460: * core-idef:: Implementation Defined Options
461: * core-ambcond:: Ambiguous Conditions
462: * core-other:: Other System Documentation
463:
464: The optional Block word set
465:
466: * block-idef:: Implementation Defined Options
467: * block-ambcond:: Ambiguous Conditions
468: * block-other:: Other System Documentation
469:
470: The optional Double Number word set
471:
472: * double-ambcond:: Ambiguous Conditions
473:
474: The optional Exception word set
475:
476: * exception-idef:: Implementation Defined Options
477:
478: The optional Facility word set
479:
480: * facility-idef:: Implementation Defined Options
481: * facility-ambcond:: Ambiguous Conditions
482:
483: The optional File-Access word set
484:
485: * file-idef:: Implementation Defined Options
486: * file-ambcond:: Ambiguous Conditions
487:
488: The optional Floating-Point word set
489:
490: * floating-idef:: Implementation Defined Options
491: * floating-ambcond:: Ambiguous Conditions
492:
493: The optional Locals word set
494:
495: * locals-idef:: Implementation Defined Options
496: * locals-ambcond:: Ambiguous Conditions
497:
498: The optional Memory-Allocation word set
499:
500: * memory-idef:: Implementation Defined Options
501:
502: The optional Programming-Tools word set
503:
504: * programming-idef:: Implementation Defined Options
505: * programming-ambcond:: Ambiguous Conditions
506:
507: The optional Search-Order word set
508:
509: * search-idef:: Implementation Defined Options
510: * search-ambcond:: Ambiguous Conditions
511:
512: Image Files
513:
514: * Image Licensing Issues:: Distribution terms for images.
515: * Image File Background:: Why have image files?
516: * Non-Relocatable Image Files:: don't always work.
517: * Data-Relocatable Image Files:: are better.
518: * Fully Relocatable Image Files:: better yet.
519: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
520: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
521: * Modifying the Startup Sequence:: and turnkey applications.
522:
523: Fully Relocatable Image Files
524:
525: * gforthmi:: The normal way
526: * cross.fs:: The hard way
527:
528: Engine
529:
530: * Portability::
531: * Threading::
532: * Primitives::
533: * Performance::
534:
535: Threading
536:
537: * Scheduling::
538: * Direct or Indirect Threaded?::
539: * DOES>::
540:
541: Primitives
542:
543: * Automatic Generation::
544: * TOS Optimization::
545: * Produced code::
546:
547: Cross Compiler
548:
549: * Using the Cross Compiler::
550: * How the Cross Compiler Works::
551:
552: @end detailmenu
553: @end menu
554:
555: @node License, Goals, Top, Top
556: @unnumbered GNU GENERAL PUBLIC LICENSE
557: @center Version 2, June 1991
558:
559: @display
560: Copyright @copyright{} 1989, 1991 Free Software Foundation, Inc.
561: 59 Temple Place, Suite 330, Boston, MA 02111, USA
562:
563: Everyone is permitted to copy and distribute verbatim copies
564: of this license document, but changing it is not allowed.
565: @end display
566:
567: @unnumberedsec Preamble
568:
569: The licenses for most software are designed to take away your
570: freedom to share and change it. By contrast, the GNU General Public
571: License is intended to guarantee your freedom to share and change free
572: software---to make sure the software is free for all its users. This
573: General Public License applies to most of the Free Software
574: Foundation's software and to any other program whose authors commit to
575: using it. (Some other Free Software Foundation software is covered by
576: the GNU Library General Public License instead.) You can apply it to
577: your programs, too.
578:
579: When we speak of free software, we are referring to freedom, not
580: price. Our General Public Licenses are designed to make sure that you
581: have the freedom to distribute copies of free software (and charge for
582: this service if you wish), that you receive source code or can get it
583: if you want it, that you can change the software or use pieces of it
584: in new free programs; and that you know you can do these things.
585:
586: To protect your rights, we need to make restrictions that forbid
587: anyone to deny you these rights or to ask you to surrender the rights.
588: These restrictions translate to certain responsibilities for you if you
589: distribute copies of the software, or if you modify it.
590:
591: For example, if you distribute copies of such a program, whether
592: gratis or for a fee, you must give the recipients all the rights that
593: you have. You must make sure that they, too, receive or can get the
594: source code. And you must show them these terms so they know their
595: rights.
596:
597: We protect your rights with two steps: (1) copyright the software, and
598: (2) offer you this license which gives you legal permission to copy,
599: distribute and/or modify the software.
600:
601: Also, for each author's protection and ours, we want to make certain
602: that everyone understands that there is no warranty for this free
603: software. If the software is modified by someone else and passed on, we
604: want its recipients to know that what they have is not the original, so
605: that any problems introduced by others will not reflect on the original
606: authors' reputations.
607:
608: Finally, any free program is threatened constantly by software
609: patents. We wish to avoid the danger that redistributors of a free
610: program will individually obtain patent licenses, in effect making the
611: program proprietary. To prevent this, we have made it clear that any
612: patent must be licensed for everyone's free use or not licensed at all.
613:
614: The precise terms and conditions for copying, distribution and
615: modification follow.
616:
617: @iftex
618: @unnumberedsec TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
619: @end iftex
620: @ifnottex
621: @center TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
622: @end ifnottex
623:
624: @enumerate 0
625: @item
626: This License applies to any program or other work which contains
627: a notice placed by the copyright holder saying it may be distributed
628: under the terms of this General Public License. The ``Program'', below,
629: refers to any such program or work, and a ``work based on the Program''
630: means either the Program or any derivative work under copyright law:
631: that is to say, a work containing the Program or a portion of it,
632: either verbatim or with modifications and/or translated into another
633: language. (Hereinafter, translation is included without limitation in
634: the term ``modification''.) Each licensee is addressed as ``you''.
635:
636: Activities other than copying, distribution and modification are not
637: covered by this License; they are outside its scope. The act of
638: running the Program is not restricted, and the output from the Program
639: is covered only if its contents constitute a work based on the
640: Program (independent of having been made by running the Program).
641: Whether that is true depends on what the Program does.
642:
643: @item
644: You may copy and distribute verbatim copies of the Program's
645: source code as you receive it, in any medium, provided that you
646: conspicuously and appropriately publish on each copy an appropriate
647: copyright notice and disclaimer of warranty; keep intact all the
648: notices that refer to this License and to the absence of any warranty;
649: and give any other recipients of the Program a copy of this License
650: along with the Program.
651:
652: You may charge a fee for the physical act of transferring a copy, and
653: you may at your option offer warranty protection in exchange for a fee.
654:
655: @item
656: You may modify your copy or copies of the Program or any portion
657: of it, thus forming a work based on the Program, and copy and
658: distribute such modifications or work under the terms of Section 1
659: above, provided that you also meet all of these conditions:
660:
661: @enumerate a
662: @item
663: You must cause the modified files to carry prominent notices
664: stating that you changed the files and the date of any change.
665:
666: @item
667: You must cause any work that you distribute or publish, that in
668: whole or in part contains or is derived from the Program or any
669: part thereof, to be licensed as a whole at no charge to all third
670: parties under the terms of this License.
671:
672: @item
673: If the modified program normally reads commands interactively
674: when run, you must cause it, when started running for such
675: interactive use in the most ordinary way, to print or display an
676: announcement including an appropriate copyright notice and a
677: notice that there is no warranty (or else, saying that you provide
678: a warranty) and that users may redistribute the program under
679: these conditions, and telling the user how to view a copy of this
680: License. (Exception: if the Program itself is interactive but
681: does not normally print such an announcement, your work based on
682: the Program is not required to print an announcement.)
683: @end enumerate
684:
685: These requirements apply to the modified work as a whole. If
686: identifiable sections of that work are not derived from the Program,
687: and can be reasonably considered independent and separate works in
688: themselves, then this License, and its terms, do not apply to those
689: sections when you distribute them as separate works. But when you
690: distribute the same sections as part of a whole which is a work based
691: on the Program, the distribution of the whole must be on the terms of
692: this License, whose permissions for other licensees extend to the
693: entire whole, and thus to each and every part regardless of who wrote it.
694:
695: Thus, it is not the intent of this section to claim rights or contest
696: your rights to work written entirely by you; rather, the intent is to
697: exercise the right to control the distribution of derivative or
698: collective works based on the Program.
699:
700: In addition, mere aggregation of another work not based on the Program
701: with the Program (or with a work based on the Program) on a volume of
702: a storage or distribution medium does not bring the other work under
703: the scope of this License.
704:
705: @item
706: You may copy and distribute the Program (or a work based on it,
707: under Section 2) in object code or executable form under the terms of
708: Sections 1 and 2 above provided that you also do one of the following:
709:
710: @enumerate a
711: @item
712: Accompany it with the complete corresponding machine-readable
713: source code, which must be distributed under the terms of Sections
714: 1 and 2 above on a medium customarily used for software interchange; or,
715:
716: @item
717: Accompany it with a written offer, valid for at least three
718: years, to give any third party, for a charge no more than your
719: cost of physically performing source distribution, a complete
720: machine-readable copy of the corresponding source code, to be
721: distributed under the terms of Sections 1 and 2 above on a medium
722: customarily used for software interchange; or,
723:
724: @item
725: Accompany it with the information you received as to the offer
726: to distribute corresponding source code. (This alternative is
727: allowed only for noncommercial distribution and only if you
728: received the program in object code or executable form with such
729: an offer, in accord with Subsection b above.)
730: @end enumerate
731:
732: The source code for a work means the preferred form of the work for
733: making modifications to it. For an executable work, complete source
734: code means all the source code for all modules it contains, plus any
735: associated interface definition files, plus the scripts used to
736: control compilation and installation of the executable. However, as a
737: special exception, the source code distributed need not include
738: anything that is normally distributed (in either source or binary
739: form) with the major components (compiler, kernel, and so on) of the
740: operating system on which the executable runs, unless that component
741: itself accompanies the executable.
742:
743: If distribution of executable or object code is made by offering
744: access to copy from a designated place, then offering equivalent
745: access to copy the source code from the same place counts as
746: distribution of the source code, even though third parties are not
747: compelled to copy the source along with the object code.
748:
749: @item
750: You may not copy, modify, sublicense, or distribute the Program
751: except as expressly provided under this License. Any attempt
752: otherwise to copy, modify, sublicense or distribute the Program is
753: void, and will automatically terminate your rights under this License.
754: However, parties who have received copies, or rights, from you under
755: this License will not have their licenses terminated so long as such
756: parties remain in full compliance.
757:
758: @item
759: You are not required to accept this License, since you have not
760: signed it. However, nothing else grants you permission to modify or
761: distribute the Program or its derivative works. These actions are
762: prohibited by law if you do not accept this License. Therefore, by
763: modifying or distributing the Program (or any work based on the
764: Program), you indicate your acceptance of this License to do so, and
765: all its terms and conditions for copying, distributing or modifying
766: the Program or works based on it.
767:
768: @item
769: Each time you redistribute the Program (or any work based on the
770: Program), the recipient automatically receives a license from the
771: original licensor to copy, distribute or modify the Program subject to
772: these terms and conditions. You may not impose any further
773: restrictions on the recipients' exercise of the rights granted herein.
774: You are not responsible for enforcing compliance by third parties to
775: this License.
776:
777: @item
778: If, as a consequence of a court judgment or allegation of patent
779: infringement or for any other reason (not limited to patent issues),
780: conditions are imposed on you (whether by court order, agreement or
781: otherwise) that contradict the conditions of this License, they do not
782: excuse you from the conditions of this License. If you cannot
783: distribute so as to satisfy simultaneously your obligations under this
784: License and any other pertinent obligations, then as a consequence you
785: may not distribute the Program at all. For example, if a patent
786: license would not permit royalty-free redistribution of the Program by
787: all those who receive copies directly or indirectly through you, then
788: the only way you could satisfy both it and this License would be to
789: refrain entirely from distribution of the Program.
790:
791: If any portion of this section is held invalid or unenforceable under
792: any particular circumstance, the balance of the section is intended to
793: apply and the section as a whole is intended to apply in other
794: circumstances.
795:
796: It is not the purpose of this section to induce you to infringe any
797: patents or other property right claims or to contest validity of any
798: such claims; this section has the sole purpose of protecting the
799: integrity of the free software distribution system, which is
800: implemented by public license practices. Many people have made
801: generous contributions to the wide range of software distributed
802: through that system in reliance on consistent application of that
803: system; it is up to the author/donor to decide if he or she is willing
804: to distribute software through any other system and a licensee cannot
805: impose that choice.
806:
807: This section is intended to make thoroughly clear what is believed to
808: be a consequence of the rest of this License.
809:
810: @item
811: If the distribution and/or use of the Program is restricted in
812: certain countries either by patents or by copyrighted interfaces, the
813: original copyright holder who places the Program under this License
814: may add an explicit geographical distribution limitation excluding
815: those countries, so that distribution is permitted only in or among
816: countries not thus excluded. In such case, this License incorporates
817: the limitation as if written in the body of this License.
818:
819: @item
820: The Free Software Foundation may publish revised and/or new versions
821: of the General Public License from time to time. Such new versions will
822: be similar in spirit to the present version, but may differ in detail to
823: address new problems or concerns.
824:
825: Each version is given a distinguishing version number. If the Program
826: specifies a version number of this License which applies to it and ``any
827: later version'', you have the option of following the terms and conditions
828: either of that version or of any later version published by the Free
829: Software Foundation. If the Program does not specify a version number of
830: this License, you may choose any version ever published by the Free Software
831: Foundation.
832:
833: @item
834: If you wish to incorporate parts of the Program into other free
835: programs whose distribution conditions are different, write to the author
836: to ask for permission. For software which is copyrighted by the Free
837: Software Foundation, write to the Free Software Foundation; we sometimes
838: make exceptions for this. Our decision will be guided by the two goals
839: of preserving the free status of all derivatives of our free software and
840: of promoting the sharing and reuse of software generally.
841:
842: @iftex
843: @heading NO WARRANTY
844: @end iftex
845: @ifnottex
846: @center NO WARRANTY
847: @end ifnottex
848:
849: @item
850: BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
851: FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
852: OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
853: PROVIDE THE PROGRAM ``AS IS'' WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
854: OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
855: MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
856: TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
857: PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
858: REPAIR OR CORRECTION.
859:
860: @item
861: IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
862: WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
863: REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
864: INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
865: OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
866: TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
867: YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
868: PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
869: POSSIBILITY OF SUCH DAMAGES.
870: @end enumerate
871:
872: @iftex
873: @heading END OF TERMS AND CONDITIONS
874: @end iftex
875: @ifnottex
876: @center END OF TERMS AND CONDITIONS
877: @end ifnottex
878:
879: @page
880: @unnumberedsec How to Apply These Terms to Your New Programs
881:
882: If you develop a new program, and you want it to be of the greatest
883: possible use to the public, the best way to achieve this is to make it
884: free software which everyone can redistribute and change under these terms.
885:
886: To do so, attach the following notices to the program. It is safest
887: to attach them to the start of each source file to most effectively
888: convey the exclusion of warranty; and each file should have at least
889: the ``copyright'' line and a pointer to where the full notice is found.
890:
891: @smallexample
892: @var{one line to give the program's name and a brief idea of what it does.}
893: Copyright (C) 19@var{yy} @var{name of author}
894:
895: This program is free software; you can redistribute it and/or modify
896: it under the terms of the GNU General Public License as published by
897: the Free Software Foundation; either version 2 of the License, or
898: (at your option) any later version.
899:
900: This program is distributed in the hope that it will be useful,
901: but WITHOUT ANY WARRANTY; without even the implied warranty of
902: MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
903: GNU General Public License for more details.
904:
905: You should have received a copy of the GNU General Public License
906: along with this program; if not, write to the Free Software
907: Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
908: @end smallexample
909:
910: Also add information on how to contact you by electronic and paper mail.
911:
912: If the program is interactive, make it output a short notice like this
913: when it starts in an interactive mode:
914:
915: @smallexample
916: Gnomovision version 69, Copyright (C) 19@var{yy} @var{name of author}
917: Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
918: type `show w'.
919: This is free software, and you are welcome to redistribute it
920: under certain conditions; type `show c' for details.
921: @end smallexample
922:
923: The hypothetical commands @samp{show w} and @samp{show c} should show
924: the appropriate parts of the General Public License. Of course, the
925: commands you use may be called something other than @samp{show w} and
926: @samp{show c}; they could even be mouse-clicks or menu items---whatever
927: suits your program.
928:
929: You should also get your employer (if you work as a programmer) or your
930: school, if any, to sign a ``copyright disclaimer'' for the program, if
931: necessary. Here is a sample; alter the names:
932:
933: @smallexample
934: Yoyodyne, Inc., hereby disclaims all copyright interest in the program
935: `Gnomovision' (which makes passes at compilers) written by James Hacker.
936:
937: @var{signature of Ty Coon}, 1 April 1989
938: Ty Coon, President of Vice
939: @end smallexample
940:
941: This General Public License does not permit incorporating your program into
942: proprietary programs. If your program is a subroutine library, you may
943: consider it more useful to permit linking proprietary applications with the
944: library. If this is what you want to do, use the GNU Library General
945: Public License instead of this License.
946:
947: @iftex
948: @unnumbered Preface
949: @cindex Preface
950: This manual documents Gforth. Some introductory material is provided for
951: readers who are unfamiliar with Forth or who are migrating to Gforth
952: from other Forth compilers. However, this manual is primarily a
953: reference manual.
954: @end iftex
955:
956: @comment TODO much more blurb here.
957:
958: @c ******************************************************************
959: @node Goals, Gforth Environment, License, Top
960: @comment node-name, next, previous, up
961: @chapter Goals of Gforth
962: @cindex goals of the Gforth project
963: The goal of the Gforth Project is to develop a standard model for
964: ANS Forth. This can be split into several subgoals:
965:
966: @itemize @bullet
967: @item
968: Gforth should conform to the ANS Forth Standard.
969: @item
970: It should be a model, i.e. it should define all the
971: implementation-dependent things.
972: @item
973: It should become standard, i.e. widely accepted and used. This goal
974: is the most difficult one.
975: @end itemize
976:
977: To achieve these goals Gforth should be
978: @itemize @bullet
979: @item
980: Similar to previous models (fig-Forth, F83)
981: @item
982: Powerful. It should provide for all the things that are considered
983: necessary today and even some that are not yet considered necessary.
984: @item
985: Efficient. It should not get the reputation of being exceptionally
986: slow.
987: @item
988: Free.
989: @item
990: Available on many machines/easy to port.
991: @end itemize
992:
993: Have we achieved these goals? Gforth conforms to the ANS Forth
994: standard. It may be considered a model, but we have not yet documented
995: which parts of the model are stable and which parts we are likely to
996: change. It certainly has not yet become a de facto standard, but it
997: appears to be quite popular. It has some similarities to and some
998: differences from previous models. It has some powerful features, but not
999: yet everything that we envisioned. We certainly have achieved our
1000: execution speed goals (@pxref{Performance})@footnote{However, in 1998
1001: the bar was raised when the major commercial Forth vendors switched to
1002: native code compilers.}. It is free and available on many machines.
1003:
1004: @c ******************************************************************
1005: @node Gforth Environment, Tutorial, Goals, Top
1006: @chapter Gforth Environment
1007: @cindex Gforth environment
1008:
1009: Note: ultimately, the Gforth man page will be auto-generated from the
1010: material in this chapter.
1011:
1012: @menu
1013: * Invoking Gforth:: Getting in
1014: * Leaving Gforth:: Getting out
1015: * Command-line editing::
1016: * Environment variables:: that affect how Gforth starts up
1017: * Gforth Files:: What gets installed and where
1018: * Startup speed:: When 35ms is not fast enough ...
1019: @end menu
1020:
1021: For related information about the creation of images see @ref{Image Files}.
1022:
1023: @comment ----------------------------------------------
1024: @node Invoking Gforth, Leaving Gforth, Gforth Environment, Gforth Environment
1025: @section Invoking Gforth
1026: @cindex invoking Gforth
1027: @cindex running Gforth
1028: @cindex command-line options
1029: @cindex options on the command line
1030: @cindex flags on the command line
1031:
1032: Gforth is made up of two parts; an executable ``engine'' (named
1033: @file{gforth} or @file{gforth-fast}) and an image file. To start it, you
1034: will usually just say @code{gforth} -- this automatically loads the
1035: default image file @file{gforth.fi}. In many other cases the default
1036: Gforth image will be invoked like this:
1037: @example
1038: gforth [file | -e forth-code] ...
1039: @end example
1040: @noindent
1041: This interprets the contents of the files and the Forth code in the order they
1042: are given.
1043:
1044: In addition to the @file{gforth} engine, there is also an engine called
1045: @file{gforth-fast}, which is faster, but gives less informative error
1046: messages (@pxref{Error messages}). You should use it for debugged,
1047: performance-critical programs.
1048:
1049: In general, the command line looks like this:
1050:
1051: @example
1052: gforth[-fast] [engine options] [image options]
1053: @end example
1054:
1055: The engine options must come before the rest of the command
1056: line. They are:
1057:
1058: @table @code
1059: @cindex -i, command-line option
1060: @cindex --image-file, command-line option
1061: @item --image-file @i{file}
1062: @itemx -i @i{file}
1063: Loads the Forth image @i{file} instead of the default
1064: @file{gforth.fi} (@pxref{Image Files}).
1065:
1066: @cindex --appl-image, command-line option
1067: @item --appl-image @i{file}
1068: Loads the image @i{file} and leaves all further command-line arguments
1069: to the image (instead of processing them as engine options). This is
1070: useful for building executable application images on Unix, built with
1071: @code{gforthmi --application ...}.
1072:
1073: @cindex --path, command-line option
1074: @cindex -p, command-line option
1075: @item --path @i{path}
1076: @itemx -p @i{path}
1077: Uses @i{path} for searching the image file and Forth source code files
1078: instead of the default in the environment variable @code{GFORTHPATH} or
1079: the path specified at installation time (e.g.,
1080: @file{/usr/local/share/gforth/0.2.0:.}). A path is given as a list of
1081: directories, separated by @samp{:} (on Unix) or @samp{;} (on other OSs).
1082:
1083: @cindex --dictionary-size, command-line option
1084: @cindex -m, command-line option
1085: @cindex @i{size} parameters for command-line options
1086: @cindex size of the dictionary and the stacks
1087: @item --dictionary-size @i{size}
1088: @itemx -m @i{size}
1089: Allocate @i{size} space for the Forth dictionary space instead of
1090: using the default specified in the image (typically 256K). The
1091: @i{size} specification for this and subsequent options consists of
1092: an integer and a unit (e.g.,
1093: @code{4M}). The unit can be one of @code{b} (bytes), @code{e} (element
1094: size, in this case Cells), @code{k} (kilobytes), @code{M} (Megabytes),
1095: @code{G} (Gigabytes), and @code{T} (Terabytes). If no unit is specified,
1096: @code{e} is used.
1097:
1098: @cindex --data-stack-size, command-line option
1099: @cindex -d, command-line option
1100: @item --data-stack-size @i{size}
1101: @itemx -d @i{size}
1102: Allocate @i{size} space for the data stack instead of using the
1103: default specified in the image (typically 16K).
1104:
1105: @cindex --return-stack-size, command-line option
1106: @cindex -r, command-line option
1107: @item --return-stack-size @i{size}
1108: @itemx -r @i{size}
1109: Allocate @i{size} space for the return stack instead of using the
1110: default specified in the image (typically 15K).
1111:
1112: @cindex --fp-stack-size, command-line option
1113: @cindex -f, command-line option
1114: @item --fp-stack-size @i{size}
1115: @itemx -f @i{size}
1116: Allocate @i{size} space for the floating point stack instead of
1117: using the default specified in the image (typically 15.5K). In this case
1118: the unit specifier @code{e} refers to floating point numbers.
1119:
1120: @cindex --locals-stack-size, command-line option
1121: @cindex -l, command-line option
1122: @item --locals-stack-size @i{size}
1123: @itemx -l @i{size}
1124: Allocate @i{size} space for the locals stack instead of using the
1125: default specified in the image (typically 14.5K).
1126:
1127: @cindex -h, command-line option
1128: @cindex --help, command-line option
1129: @item --help
1130: @itemx -h
1131: Print a message about the command-line options
1132:
1133: @cindex -v, command-line option
1134: @cindex --version, command-line option
1135: @item --version
1136: @itemx -v
1137: Print version and exit
1138:
1139: @cindex --debug, command-line option
1140: @item --debug
1141: Print some information useful for debugging on startup.
1142:
1143: @cindex --offset-image, command-line option
1144: @item --offset-image
1145: Start the dictionary at a slightly different position than would be used
1146: otherwise (useful for creating data-relocatable images,
1147: @pxref{Data-Relocatable Image Files}).
1148:
1149: @cindex --no-offset-im, command-line option
1150: @item --no-offset-im
1151: Start the dictionary at the normal position.
1152:
1153: @cindex --clear-dictionary, command-line option
1154: @item --clear-dictionary
1155: Initialize all bytes in the dictionary to 0 before loading the image
1156: (@pxref{Data-Relocatable Image Files}).
1157:
1158: @cindex --die-on-signal, command-line-option
1159: @item --die-on-signal
1160: Normally Gforth handles most signals (e.g., the user interrupt SIGINT,
1161: or the segmentation violation SIGSEGV) by translating it into a Forth
1162: @code{THROW}. With this option, Gforth exits if it receives such a
1163: signal. This option is useful when the engine and/or the image might be
1164: severely broken (such that it causes another signal before recovering
1165: from the first); this option avoids endless loops in such cases.
1166: @end table
1167:
1168: @cindex loading files at startup
1169: @cindex executing code on startup
1170: @cindex batch processing with Gforth
1171: As explained above, the image-specific command-line arguments for the
1172: default image @file{gforth.fi} consist of a sequence of filenames and
1173: @code{-e @var{forth-code}} options that are interpreted in the sequence
1174: in which they are given. The @code{-e @var{forth-code}} or
1175: @code{--evaluate @var{forth-code}} option evaluates the Forth
1176: code. This option takes only one argument; if you want to evaluate more
1177: Forth words, you have to quote them or use @code{-e} several times. To exit
1178: after processing the command line (instead of entering interactive mode)
1179: append @code{-e bye} to the command line.
1180:
1181: @cindex versions, invoking other versions of Gforth
1182: If you have several versions of Gforth installed, @code{gforth} will
1183: invoke the version that was installed last. @code{gforth-@i{version}}
1184: invokes a specific version. If your environment contains the variable
1185: @code{GFORTHPATH}, you may want to override it by using the
1186: @code{--path} option.
1187:
1188: Not yet implemented:
1189: On startup the system first executes the system initialization file
1190: (unless the option @code{--no-init-file} is given; note that the system
1191: resulting from using this option may not be ANS Forth conformant). Then
1192: the user initialization file @file{.gforth.fs} is executed, unless the
1193: option @code{--no-rc} is given; this file is searched for in @file{.},
1194: then in @file{~}, then in the normal path (see above).
1195:
1196:
1197:
1198: @comment ----------------------------------------------
1199: @node Leaving Gforth, Command-line editing, Invoking Gforth, Gforth Environment
1200: @section Leaving Gforth
1201: @cindex Gforth - leaving
1202: @cindex leaving Gforth
1203:
1204: You can leave Gforth by typing @code{bye} or @kbd{Ctrl-d} (at the start
1205: of a line) or (if you invoked Gforth with the @code{--die-on-signal}
1206: option) @kbd{Ctrl-c}. When you leave Gforth, all of your definitions and
1207: data are discarded. For ways of saving the state of the system before
1208: leaving Gforth see @ref{Image Files}.
1209:
1210: doc-bye
1211:
1212:
1213: @comment ----------------------------------------------
1214: @node Command-line editing, Environment variables, Leaving Gforth, Gforth Environment
1215: @section Command-line editing
1216: @cindex command-line editing
1217:
1218: Gforth maintains a history file that records every line that you type to
1219: the text interpreter. This file is preserved between sessions, and is
1220: used to provide a command-line recall facility; if you type @kbd{Ctrl-P}
1221: repeatedly you can recall successively older commands from this (or
1222: previous) session(s). The full list of command-line editing facilities is:
1223:
1224: @itemize @bullet
1225: @item
1226: @kbd{Ctrl-p} (``previous'') (or up-arrow) to recall successively older
1227: commands from the history buffer.
1228: @item
1229: @kbd{Ctrl-n} (``next'') (or down-arrow) to recall successively newer commands
1230: from the history buffer.
1231: @item
1232: @kbd{Ctrl-f} (or right-arrow) to move the cursor right, non-destructively.
1233: @item
1234: @kbd{Ctrl-b} (or left-arrow) to move the cursor left, non-destructively.
1235: @item
1236: @kbd{Ctrl-h} (backspace) to delete the character to the left of the cursor,
1237: closing up the line.
1238: @item
1239: @kbd{Ctrl-k} to delete (``kill'') from the cursor to the end of the line.
1240: @item
1241: @kbd{Ctrl-a} to move the cursor to the start of the line.
1242: @item
1243: @kbd{Ctrl-e} to move the cursor to the end of the line.
1244: @item
1245: @key{RET} (@kbd{Ctrl-m}) or @key{LFD} (@kbd{Ctrl-j}) to submit the current
1246: line.
1247: @item
1248: @key{TAB} to step through all possible full-word completions of the word
1249: currently being typed.
1250: @item
1251: @kbd{Ctrl-d} on an empty line line to terminate Gforth (gracefully,
1252: using @code{bye}).
1253: @item
1254: @kbd{Ctrl-x} (or @code{Ctrl-d} on a non-empty line) to delete the
1255: character under the cursor.
1256: @end itemize
1257:
1258: When editing, displayable characters are inserted to the left of the
1259: cursor position; the line is always in ``insert'' (as opposed to
1260: ``overstrike'') mode.
1261:
1262: @cindex history file
1263: @cindex @file{.gforth-history}
1264: On Unix systems, the history file is @file{~/.gforth-history} by
1265: default@footnote{i.e. it is stored in the user's home directory.}. You
1266: can find out the name and location of your history file using:
1267:
1268: @example
1269: history-file type \ Unix-class systems
1270:
1271: history-file type \ Other systems
1272: history-dir type
1273: @end example
1274:
1275: If you enter long definitions by hand, you can use a text editor to
1276: paste them out of the history file into a Forth source file for reuse at
1277: a later time.
1278:
1279: Gforth never trims the size of the history file, so you should do this
1280: periodically, if necessary.
1281:
1282: @comment this is all defined in history.fs
1283: @comment NAC TODO the ctrl-D behaviour can either do a bye or a beep.. how is that option
1284: @comment chosen?
1285:
1286:
1287: @comment ----------------------------------------------
1288: @node Environment variables, Gforth Files, Command-line editing, Gforth Environment
1289: @section Environment variables
1290: @cindex environment variables
1291:
1292: Gforth uses these environment variables:
1293:
1294: @itemize @bullet
1295: @item
1296: @cindex @code{GFORTHHIST} -- environment variable
1297: @code{GFORTHHIST} -- (Unix systems only) specifies the directory in which to
1298: open/create the history file, @file{.gforth-history}. Default:
1299: @code{$HOME}.
1300:
1301: @item
1302: @cindex @code{GFORTHPATH} -- environment variable
1303: @code{GFORTHPATH} -- specifies the path used when searching for the gforth image file and
1304: for Forth source-code files.
1305:
1306: @item
1307: @cindex @code{GFORTH} -- environment variable
1308: @code{GFORTH} -- used by @file{gforthmi}, @xref{gforthmi}.
1309:
1310: @item
1311: @cindex @code{GFORTHD} -- environment variable
1312: @code{GFORTHD} -- used by @file{gforthmi}, @xref{gforthmi}.
1313:
1314: @item
1315: @cindex @code{TMP}, @code{TEMP} - environment variable
1316: @code{TMP}, @code{TEMP} - (non-Unix systems only) used as a potential
1317: location for the history file.
1318: @end itemize
1319:
1320: @comment also POSIXELY_CORRECT LINES COLUMNS HOME but no interest in
1321: @comment mentioning these.
1322:
1323: All the Gforth environment variables default to sensible values if they
1324: are not set.
1325:
1326:
1327: @comment ----------------------------------------------
1328: @node Gforth Files, Startup speed, Environment variables, Gforth Environment
1329: @section Gforth files
1330: @cindex Gforth files
1331:
1332: When you install Gforth on a Unix system, it installs files in these
1333: locations by default:
1334:
1335: @itemize @bullet
1336: @item
1337: @file{/usr/local/bin/gforth}
1338: @item
1339: @file{/usr/local/bin/gforthmi}
1340: @item
1341: @file{/usr/local/man/man1/gforth.1} - man page.
1342: @item
1343: @file{/usr/local/info} - the Info version of this manual.
1344: @item
1345: @file{/usr/local/lib/gforth/<version>/...} - Gforth @file{.fi} files.
1346: @item
1347: @file{/usr/local/share/gforth/<version>/TAGS} - Emacs TAGS file.
1348: @item
1349: @file{/usr/local/share/gforth/<version>/...} - Gforth source files.
1350: @item
1351: @file{.../emacs/site-lisp/gforth.el} - Emacs gforth mode.
1352: @end itemize
1353:
1354: You can select different places for installation by using
1355: @code{configure} options (listed with @code{configure --help}).
1356:
1357: @comment ----------------------------------------------
1358: @node Startup speed, , Gforth Files, Gforth Environment
1359: @section Startup speed
1360: @cindex Startup speed
1361: @cindex speed, startup
1362:
1363: If Gforth is used for CGI scripts or in shell scripts, its startup
1364: speed may become a problem. On a 300MHz 21064a under Linux-2.2.13 with
1365: glibc-2.0.7, @code{gforth -e bye} takes about 24.6ms user and 11.3ms
1366: system time.
1367:
1368: If startup speed is a problem, you may consider the following ways to
1369: improve it; or you may consider ways to reduce the number of startups
1370: (for example, by using Fast-CGI).
1371:
1372: The first step to improve startup speed is to statically link Gforth, by
1373: building it with @code{XLDFLAGS=-static}. This requires more memory for
1374: the code and will therefore slow down the first invocation, but
1375: subsequent invocations avoid the dynamic linking overhead. Another
1376: disadvantage is that Gforth won't profit from library upgrades. As a
1377: result, @code{gforth-static -e bye} takes about 17.1ms user and
1378: 8.2ms system time.
1379:
1380: The next step to improve startup speed is to use a non-relocatable image
1381: (@pxref{Non-Relocatable Image Files}). You can create this image with
1382: @code{gforth -e "savesystem gforthnr.fi bye"} and later use it with
1383: @code{gforth -i gforthnr.fi ...}. This avoids the relocation overhead
1384: and a part of the copy-on-write overhead. The disadvantage is that the
1385: non-relocatable image does not work if the OS gives Gforth a different
1386: address for the dictionary, for whatever reason; so you better provide a
1387: fallback on a relocatable image. @code{gforth-static -i gforthnr.fi -e
1388: bye} takes about 15.3ms user and 7.5ms system time.
1389:
1390: The final step is to disable dictionary hashing in Gforth. Gforth
1391: builds the hash table on startup, which takes much of the startup
1392: overhead. You can do this by commenting out the @code{include hash.fs}
1393: in @file{startup.fs} and everything that requires @file{hash.fs} (at the
1394: moment @file{table.fs} and @file{ekey.fs}) and then doing @code{make}.
1395: The disadvantages are that functionality like @code{table} and
1396: @code{ekey} is missing and that text interpretation (e.g., compiling)
1397: now takes much longer. So, you should only use this method if there is
1398: no significant text interpretation to perform (the script should be
1399: compiled into the image, amongst other things). @code{gforth-static -i
1400: gforthnrnh.fi -e bye} takes about 2.1ms user and 6.1ms system time.
1401:
1402: @c ******************************************************************
1403: @node Tutorial, Introduction, Gforth Environment, Top
1404: @chapter Forth Tutorial
1405: @cindex Tutorial
1406: @cindex Forth Tutorial
1407:
1408: @c Topics from nac's Introduction that could be mentioned:
1409: @c press <ret> after each line
1410: @c Prompt
1411: @c numbers vs. words in dictionary on text interpretation
1412: @c what happens on redefinition
1413: @c parsing words (in particular, defining words)
1414:
1415: The difference of this chapter from the Introduction
1416: (@pxref{Introduction}) is that this tutorial is more fast-paced, should
1417: be used while sitting in front of a computer, and covers much more
1418: material, but does not explain how the Forth system works.
1419:
1420: This tutorial can be used with any ANS-compliant Forth; any
1421: Gforth-specific features are marked as such and you can skip them if you
1422: work with another Forth. This tutorial does not explain all features of
1423: Forth, just enough to get you started and give you some ideas about the
1424: facilities available in Forth. Read the rest of the manual and the
1425: standard when you are through this.
1426:
1427: The intended way to use this tutorial is that you work through it while
1428: sitting in front of the console, take a look at the examples and predict
1429: what they will do, then try them out; if the outcome is not as expected,
1430: find out why (e.g., by trying out variations of the example), so you
1431: understand what's going on. There are also some assignments that you
1432: should solve.
1433:
1434: This tutorial assumes that you have programmed before and know what,
1435: e.g., a loop is.
1436:
1437: @c !! explain compat library
1438:
1439: @menu
1440: * Starting Gforth Tutorial::
1441: * Syntax Tutorial::
1442: * Crash Course Tutorial::
1443: * Stack Tutorial::
1444: * Arithmetics Tutorial::
1445: * Stack Manipulation Tutorial::
1446: * Using files for Forth code Tutorial::
1447: * Comments Tutorial::
1448: * Colon Definitions Tutorial::
1449: * Decompilation Tutorial::
1450: * Stack-Effect Comments Tutorial::
1451: * Types Tutorial::
1452: * Factoring Tutorial::
1453: * Designing the stack effect Tutorial::
1454: * Local Variables Tutorial::
1455: * Conditional execution Tutorial::
1456: * Flags and Comparisons Tutorial::
1457: * General Loops Tutorial::
1458: * Counted loops Tutorial::
1459: * Recursion Tutorial::
1460: * Leaving definitions or loops Tutorial::
1461: * Return Stack Tutorial::
1462: * Memory Tutorial::
1463: * Characters and Strings Tutorial::
1464: * Alignment Tutorial::
1465: * Files Tutorial::
1466: * Interpretation and Compilation Semantics and Immediacy Tutorial::
1467: * Execution Tokens Tutorial::
1468: * Exceptions Tutorial::
1469: * Defining Words Tutorial::
1470: * Arrays and Records Tutorial::
1471: * POSTPONE Tutorial::
1472: * Literal Tutorial::
1473: * Advanced macros Tutorial::
1474: * Compilation Tokens Tutorial::
1475: * Wordlists and Search Order Tutorial::
1476: @end menu
1477:
1478: @node Starting Gforth Tutorial, Syntax Tutorial, Tutorial, Tutorial
1479: @section Starting Gforth
1480: @cindex starting Gforth tutorial
1481: You can start Gforth by typing its name:
1482:
1483: @example
1484: gforth
1485: @end example
1486:
1487: That puts you into interactive mode; you can leave Gforth by typing
1488: @code{bye}. While in Gforth, you can edit the command line and access
1489: the command line history with cursor keys, similar to bash.
1490:
1491:
1492: @node Syntax Tutorial, Crash Course Tutorial, Starting Gforth Tutorial, Tutorial
1493: @section Syntax
1494: @cindex syntax tutorial
1495:
1496: A @dfn{word} is a sequence of arbitrary characters (expcept white
1497: space). Words are separated by white space. E.g., each of the
1498: following lines contains exactly one word:
1499:
1500: @example
1501: word
1502: !@@#$%^&*()
1503: 1234567890
1504: 5!a
1505: @end example
1506:
1507: A frequent beginner's error is to leave away necessary white space,
1508: resulting in an error like @samp{Undefined word}; so if you see such an
1509: error, check if you have put spaces wherever necessary.
1510:
1511: @example
1512: ." hello, world" \ correct
1513: ."hello, world" \ gives an "Undefined word" error
1514: @end example
1515:
1516: Gforth and most other Forth systems ignore differences in case (they are
1517: case-insensitive), i.e., @samp{word} is the same as @samp{Word}. If
1518: your system is case-sensitive, you may have to type all the examples
1519: given here in upper case.
1520:
1521:
1522: @node Crash Course Tutorial, Stack Tutorial, Syntax Tutorial, Tutorial
1523: @section Crash Course
1524:
1525: Type
1526:
1527: @example
1528: 0 0 !
1529: here execute
1530: ' catch >body 20 erase abort
1531: ' (quit) >body 20 erase
1532: @end example
1533:
1534: The last two examples are guaranteed to destroy parts of Gforth (and
1535: most other systems), so you better leave Gforth afterwards (if it has
1536: not finished by itself). On some systems you may have to kill gforth
1537: from outside (e.g., in Unix with @code{kill}).
1538:
1539: Now that you know how to produce crashes (and that there's not much to
1540: them), let's learn how to produce meaningful programs.
1541:
1542:
1543: @node Stack Tutorial, Arithmetics Tutorial, Crash Course Tutorial, Tutorial
1544: @section Stack
1545: @cindex stack tutorial
1546:
1547: The most obvious feature of Forth is the stack. When you type in a
1548: number, it is pushed on the stack. You can display the content of the
1549: stack with @code{.s}.
1550:
1551: @example
1552: 1 2 .s
1553: 3 .s
1554: @end example
1555:
1556: @code{.s} displays the top-of-stack to the right, i.e., the numbers
1557: appear in @code{.s} output as they appeared in the input.
1558:
1559: You can print the top of stack element with @code{.}.
1560:
1561: @example
1562: 1 2 3 . . .
1563: @end example
1564:
1565: In general, words consume their stack arguments (@code{.s} is an
1566: exception).
1567:
1568: @assignment
1569: What does the stack contain after @code{5 6 7 .}?
1570: @endassignment
1571:
1572:
1573: @node Arithmetics Tutorial, Stack Manipulation Tutorial, Stack Tutorial, Tutorial
1574: @section Arithmetics
1575: @cindex arithmetics tutorial
1576:
1577: The words @code{+}, @code{-}, @code{*}, @code{/}, and @code{mod} always
1578: operate on the top two stack items:
1579:
1580: @example
1581: 2 2 .s
1582: + .s
1583: .
1584: 2 1 - .
1585: 7 3 mod .
1586: @end example
1587:
1588: The operands of @code{-}, @code{/}, and @code{mod} are in the same order
1589: as in the corresponding infix expression (this is generally the case in
1590: Forth).
1591:
1592: Parentheses are superfluous (and not available), because the order of
1593: the words unambiguously determines the order of evaluation and the
1594: operands:
1595:
1596: @example
1597: 3 4 + 5 * .
1598: 3 4 5 * + .
1599: @end example
1600:
1601: @assignment
1602: What are the infix expressions corresponding to the Forth code above?
1603: Write @code{6-7*8+9} in Forth notation@footnote{This notation is also
1604: known as Postfix or RPN (Reverse Polish Notation).}.
1605: @endassignment
1606:
1607: To change the sign, use @code{negate}:
1608:
1609: @example
1610: 2 negate .
1611: @end example
1612:
1613: @assignment
1614: Convert -(-3)*4-5 to Forth.
1615: @endassignment
1616:
1617: @code{/mod} performs both @code{/} and @code{mod}.
1618:
1619: @example
1620: 7 3 /mod . .
1621: @end example
1622:
1623: Reference: @ref{Arithmetic}.
1624:
1625:
1626: @node Stack Manipulation Tutorial, Using files for Forth code Tutorial, Arithmetics Tutorial, Tutorial
1627: @section Stack Manipulation
1628: @cindex stack manipulation tutorial
1629:
1630: Stack manipulation words rearrange the data on the stack.
1631:
1632: @example
1633: 1 .s drop .s
1634: 1 .s dup .s drop drop .s
1635: 1 2 .s over .s drop drop drop
1636: 1 2 .s swap .s drop drop
1637: 1 2 3 .s rot .s drop drop drop
1638: @end example
1639:
1640: These are the most important stack manipulation words. There are also
1641: variants that manipulate twice as many stack items:
1642:
1643: @example
1644: 1 2 3 4 .s 2swap .s 2drop 2drop
1645: @end example
1646:
1647: Two more stack manipulation words are:
1648:
1649: @example
1650: 1 2 .s nip .s drop
1651: 1 2 .s tuck .s 2drop drop
1652: @end example
1653:
1654: @assignment
1655: Replace @code{nip} and @code{tuck} with combinations of other stack
1656: manipulation words.
1657:
1658: @example
1659: Given: How do you get:
1660: 1 2 3 3 2 1
1661: 1 2 3 1 2 3 2
1662: 1 2 3 1 2 3 3
1663: 1 2 3 1 3 3
1664: 1 2 3 2 1 3
1665: 1 2 3 4 4 3 2 1
1666: 1 2 3 1 2 3 1 2 3
1667: 1 2 3 4 1 2 3 4 1 2
1668: 1 2 3
1669: 1 2 3 1 2 3 4
1670: 1 2 3 1 3
1671: @end example
1672: @endassignment
1673:
1674: @example
1675: 5 dup * .
1676: @end example
1677:
1678: @assignment
1679: Write 17^3 and 17^4 in Forth, without writing @code{17} more than once.
1680: Write a piece of Forth code that expects two numbers on the stack
1681: (@var{a} and @var{b}, with @var{b} on top) and computes
1682: @code{(a-b)(a+1)}.
1683: @endassignment
1684:
1685: Reference: @ref{Stack Manipulation}.
1686:
1687:
1688: @node Using files for Forth code Tutorial, Comments Tutorial, Stack Manipulation Tutorial, Tutorial
1689: @section Using files for Forth code
1690: @cindex loading Forth code, tutorial
1691: @cindex files containing Forth code, tutorial
1692:
1693: While working at the Forth command line is convenient for one-line
1694: examples and short one-off code, you probably want to store your source
1695: code in files for convenient editing and persistence. You can use your
1696: favourite editor (Gforth includes Emacs support, @pxref{Emacs and
1697: Gforth}) to create @var{file} and use
1698:
1699: @example
1700: s" @var{file}" included
1701: @end example
1702:
1703: to load it into your Forth system. The file name extension I use for
1704: Forth files is @samp{.fs}.
1705:
1706: You can easily start Gforth with some files loaded like this:
1707:
1708: @example
1709: gforth @var{file1} @var{file2}
1710: @end example
1711:
1712: If an error occurs during loading these files, Gforth terminates,
1713: whereas an error during @code{INCLUDED} within Gforth usually gives you
1714: a Gforth command line. Starting the Forth system every time gives you a
1715: clean start every time, without interference from the results of earlier
1716: tries.
1717:
1718: I often put all the tests in a file, then load the code and run the
1719: tests with
1720:
1721: @example
1722: gforth @var{code} @var{tests} -e bye
1723: @end example
1724:
1725: (often by performing this command with @kbd{C-x C-e} in Emacs). The
1726: @code{-e bye} ensures that Gforth terminates afterwards so that I can
1727: restart this command without ado.
1728:
1729: The advantage of this approach is that the tests can be repeated easily
1730: every time the program ist changed, making it easy to catch bugs
1731: introduced by the change.
1732:
1733: Reference: @ref{Forth source files}.
1734:
1735:
1736: @node Comments Tutorial, Colon Definitions Tutorial, Using files for Forth code Tutorial, Tutorial
1737: @section Comments
1738: @cindex comments tutorial
1739:
1740: @example
1741: \ That's a comment; it ends at the end of the line
1742: ( Another comment; it ends here: ) .s
1743: @end example
1744:
1745: @code{\} and @code{(} are ordinary Forth words and therefore have to be
1746: separated with white space from the following text.
1747:
1748: @example
1749: \This gives an "Undefined word" error
1750: @end example
1751:
1752: The first @code{)} ends a comment started with @code{(}, so you cannot
1753: nest @code{(}-comments; and you cannot comment out text containing a
1754: @code{)} with @code{( ... )}@footnote{therefore it's a good idea to
1755: avoid @code{)} in word names.}.
1756:
1757: I use @code{\}-comments for descriptive text and for commenting out code
1758: of one or more line; I use @code{(}-comments for describing the stack
1759: effect, the stack contents, or for commenting out sub-line pieces of
1760: code.
1761:
1762: The Emacs mode @file{gforth.el} (@pxref{Emacs and Gforth}) supports
1763: these uses by commenting out a region with @kbd{C-x \}, uncommenting a
1764: region with @kbd{C-u C-x \}, and filling a @code{\}-commented region
1765: with @kbd{M-q}.
1766:
1767: Reference: @ref{Comments}.
1768:
1769:
1770: @node Colon Definitions Tutorial, Decompilation Tutorial, Comments Tutorial, Tutorial
1771: @section Colon Definitions
1772: @cindex colon definitions, tutorial
1773: @cindex definitions, tutorial
1774: @cindex procedures, tutorial
1775: @cindex functions, tutorial
1776:
1777: are similar to procedures and functions in other programming languages.
1778:
1779: @example
1780: : squared ( n -- n^2 )
1781: dup * ;
1782: 5 squared .
1783: 7 squared .
1784: @end example
1785:
1786: @code{:} starts the colon definition; its name is @code{squared}. The
1787: following comment describes its stack effect. The words @code{dup *}
1788: are not executed, but compiled into the definition. @code{;} ends the
1789: colon definition.
1790:
1791: The newly-defined word can be used like any other word, including using
1792: it in other definitions:
1793:
1794: @example
1795: : cubed ( n -- n^3 )
1796: dup squared * ;
1797: -5 cubed .
1798: : fourth-power ( n -- n^4 )
1799: squared squared ;
1800: 3 fourth-power .
1801: @end example
1802:
1803: @assignment
1804: Write colon definitions for @code{nip}, @code{tuck}, @code{negate}, and
1805: @code{/mod} in terms of other Forth words, and check if they work (hint:
1806: test your tests on the originals first). Don't let the
1807: @samp{redefined}-Messages spook you, they are just warnings.
1808: @endassignment
1809:
1810: Reference: @ref{Colon Definitions}.
1811:
1812:
1813: @node Decompilation Tutorial, Stack-Effect Comments Tutorial, Colon Definitions Tutorial, Tutorial
1814: @section Decompilation
1815: @cindex decompilation tutorial
1816: @cindex see tutorial
1817:
1818: You can decompile colon definitions with @code{see}:
1819:
1820: @example
1821: see squared
1822: see cubed
1823: @end example
1824:
1825: In Gforth @code{see} shows you a reconstruction of the source code from
1826: the executable code. Informations that were present in the source, but
1827: not in the executable code, are lost (e.g., comments).
1828:
1829: You can also decompile the predefined words:
1830:
1831: @example
1832: see .
1833: see +
1834: @end example
1835:
1836:
1837: @node Stack-Effect Comments Tutorial, Types Tutorial, Decompilation Tutorial, Tutorial
1838: @section Stack-Effect Comments
1839: @cindex stack-effect comments, tutorial
1840: @cindex --, tutorial
1841: By convention the comment after the name of a definition describes the
1842: stack effect: The part in from of the @samp{--} describes the state of
1843: the stack before the execution of the definition, i.e., the parameters
1844: that are passed into the colon definition; the part behind the @samp{--}
1845: is the state of the stack after the execution of the definition, i.e.,
1846: the results of the definition. The stack comment only shows the top
1847: stack items that the definition accesses and/or changes.
1848:
1849: You should put a correct stack effect on every definition, even if it is
1850: just @code{( -- )}. You should also add some descriptive comment to
1851: more complicated words (I usually do this in the lines following
1852: @code{:}). If you don't do this, your code becomes unreadable (because
1853: you have to work through every definition before you can undertsand
1854: any).
1855:
1856: @assignment
1857: The stack effect of @code{swap} can be written like this: @code{x1 x2 --
1858: x2 x1}. Describe the stack effect of @code{-}, @code{drop}, @code{dup},
1859: @code{over}, @code{rot}, @code{nip}, and @code{tuck}. Hint: When you
1860: are done, you can compare your stack effects to those in this manual
1861: (@pxref{Word Index}).
1862: @endassignment
1863:
1864: Sometimes programmers put comments at various places in colon
1865: definitions that describe the contents of the stack at that place (stack
1866: comments); i.e., they are like the first part of a stack-effect
1867: comment. E.g.,
1868:
1869: @example
1870: : cubed ( n -- n^3 )
1871: dup squared ( n n^2 ) * ;
1872: @end example
1873:
1874: In this case the stack comment is pretty superfluous, because the word
1875: is simple enough. If you think it would be a good idea to add such a
1876: comment to increase readability, you should also consider factoring the
1877: word into several simpler words (@pxref{Factoring Tutorial,,
1878: Factoring}), which typically eliminates the need for the stack comment;
1879: however, if you decide not to refactor it, then having such a comment is
1880: better than not having it.
1881:
1882: The names of the stack items in stack-effect and stack comments in the
1883: standard, in this manual, and in many programs specify the type through
1884: a type prefix, similar to Fortran and Hungarian notation. The most
1885: frequent prefixes are:
1886:
1887: @table @code
1888: @item n
1889: signed integer
1890: @item u
1891: unsigned integer
1892: @item c
1893: character
1894: @item f
1895: Boolean flags, i.e. @code{false} or @code{true}.
1896: @item a-addr,a-
1897: Cell-aligned address
1898: @item c-addr,c-
1899: Char-aligned address (note that a Char may have two bytes in Windows NT)
1900: @item xt
1901: Execution token, same size as Cell
1902: @item w,x
1903: Cell, can contain an integer or an address. It usually takes 32, 64 or
1904: 16 bits (depending on your platform and Forth system). A cell is more
1905: commonly known as machine word, but the term @emph{word} already means
1906: something different in Forth.
1907: @item d
1908: signed double-cell integer
1909: @item ud
1910: unsigned double-cell integer
1911: @item r
1912: Float (on the FP stack)
1913: @end table
1914:
1915: You can find a more complete list in @ref{Notation}.
1916:
1917: @assignment
1918: Write stack-effect comments for all definitions you have written up to
1919: now.
1920: @endassignment
1921:
1922:
1923: @node Types Tutorial, Factoring Tutorial, Stack-Effect Comments Tutorial, Tutorial
1924: @section Types
1925: @cindex types tutorial
1926:
1927: In Forth the names of the operations are not overloaded; so similar
1928: operations on different types need different names; e.g., @code{+} adds
1929: integers, and you have to use @code{f+} to add floating-point numbers.
1930: The following prefixes are often used for related operations on
1931: different types:
1932:
1933: @table @code
1934: @item (none)
1935: signed integer
1936: @item u
1937: unsigned integer
1938: @item c
1939: character
1940: @item d
1941: signed double-cell integer
1942: @item ud, du
1943: unsigned double-cell integer
1944: @item 2
1945: two cells (not-necessarily double-cell numbers)
1946: @item m, um
1947: mixed single-cell and double-cell operations
1948: @item f
1949: floating-point (note that in stack comments @samp{f} represents flags,
1950: and @samp{r} represents FP numbers).
1951: @end table
1952:
1953: If there are no differences between the signed and the unsigned variant
1954: (e.g., for @code{+}), there is only the prefix-less variant.
1955:
1956: Forth does not perform type checking, neither at compile time, nor at
1957: run time. If you use the wrong oeration, the data are interpreted
1958: incorrectly:
1959:
1960: @example
1961: -1 u.
1962: @end example
1963:
1964: If you have only experience with type-checked languages until now, and
1965: have heard how important type-checking is, don't panic! In my
1966: experience (and that of other Forthers), type errors in Forth code are
1967: usually easy to find (once you get used to it), the increased vigilance
1968: of the programmer tends to catch some harder errors in addition to most
1969: type errors, and you never have to work around the type system, so in
1970: most situations the lack of type-checking seems to be a win (projects to
1971: add type checking to Forth have not caught on).
1972:
1973:
1974: @node Factoring Tutorial, Designing the stack effect Tutorial, Types Tutorial, Tutorial
1975: @section Factoring
1976: @cindex factoring tutorial
1977:
1978: If you try to write longer definitions, you will soon find it hard to
1979: keep track of the stack contents. Therefore, good Forth programmers
1980: tend to write only short definitions (e.g., three lines). The art of
1981: finding meaningful short definitions is known as factoring (as in
1982: factoring polynomials).
1983:
1984: Well-factored programs offer additional advantages: smaller, more
1985: general words, are easier to test and debug and can be reused more and
1986: better than larger, specialized words.
1987:
1988: So, if you run into difficulties with stack management, when writing
1989: code, try to define meaningful factors for the word, and define the word
1990: in terms of those. Even if a factor contains only two words, it is
1991: often helpful.
1992:
1993: Good factoring is not easy, and it takes some practice to get the knack
1994: for it; but even experienced Forth programmers often don't find the
1995: right solution right away, but only when rewriting the program. So, if
1996: you don't come up with a good solution immediately, keep trying, don't
1997: despair.
1998:
1999: @c example !!
2000:
2001:
2002: @node Designing the stack effect Tutorial, Local Variables Tutorial, Factoring Tutorial, Tutorial
2003: @section Designing the stack effect
2004: @cindex Stack effect design, tutorial
2005: @cindex design of stack effects, tutorial
2006:
2007: In other languages you can use an arbitrary order of parameters for a
2008: function; and since there is only one result, you don't have to deal with
2009: the order of results, either.
2010:
2011: In Forth (and other stack-based languages, e.g., Postscript) the
2012: parameter and result order of a definition is important and should be
2013: designed well. The general guideline is to design the stack effect such
2014: that the word is simple to use in most cases, even if that complicates
2015: the implementation of the word. Some concrete rules are:
2016:
2017: @itemize @bullet
2018:
2019: @item
2020: Words consume all of their parameters (e.g., @code{.}).
2021:
2022: @item
2023: If there is a convention on the order of parameters (e.g., from
2024: mathematics or another programming language), stick with it (e.g.,
2025: @code{-}).
2026:
2027: @item
2028: If one parameter usually requires only a short computation (e.g., it is
2029: a constant), pass it on the top of the stack. Conversely, parameters
2030: that usually require a long sequence of code to compute should be passed
2031: as the bottom (i.e., first) parameter. This makes the code easier to
2032: read, because reader does not need to keep track of the bottom item
2033: through a long sequence of code (or, alternatively, through stack
2034: manipulations). E.g., @code{!} (store, @pxref{Memory}) expects the
2035: address on top of the stack because it is usually simpler to compute
2036: than the stored value (often the address is just a variable).
2037:
2038: @item
2039: Similarly, results that are usually consumed quickly should be returned
2040: on the top of stack, whereas a result that is often used in long
2041: computations should be passed as bottom result. E.g., the file words
2042: like @code{open-file} return the error code on the top of stack, because
2043: it is usually consumed quickly by @code{throw}; moreover, the error code
2044: has to be checked before doing anything with the other results.
2045:
2046: @end itemize
2047:
2048: These rules are just general guidelines, don't lose sight of the overall
2049: goal to make the words easy to use. E.g., if the convention rule
2050: conflicts with the computation-length rule, you might decide in favour
2051: of the convention if the word will be used rarely, and in favour of the
2052: computation-length rule if the word will be used frequently (because
2053: with frequent use the cost of breaking the computation-length rule would
2054: be quite high, and frequent use makes it easier to remember an
2055: unconventional order).
2056:
2057: @c example !! structure package
2058:
2059:
2060: @node Local Variables Tutorial, Conditional execution Tutorial, Designing the stack effect Tutorial, Tutorial
2061: @section Local Variables
2062: @cindex local variables, tutorial
2063:
2064: You can define local variables (@emph{locals}) in a colon definition:
2065:
2066: @example
2067: : swap @{ a b -- b a @}
2068: b a ;
2069: 1 2 swap .s 2drop
2070: @end example
2071:
2072: (If your Forth system does not support this syntax, include
2073: @file{compat/anslocals.fs} first).
2074:
2075: In this example @code{@{ a b -- b a @}} is the locals definition; it
2076: takes two cells from the stack, puts the top of stack in @code{b} and
2077: the next stack element in @code{a}. @code{--} starts a comment ending
2078: with @code{@}}. After the locals definition, using the name of the
2079: local will push its value on the stack. You can leave the comment
2080: part (@code{-- b a}) away:
2081:
2082: @example
2083: : swap ( x1 x2 -- x2 x1 )
2084: @{ a b @} b a ;
2085: @end example
2086:
2087: In Gforth you can have several locals definitions, anywhere in a colon
2088: definition; in contrast, in a standard program you can have only one
2089: locals definition per colon definition, and that locals definition must
2090: be outside any controll structure.
2091:
2092: With locals you can write slightly longer definitions without running
2093: into stack trouble. However, I recommend trying to write colon
2094: definitions without locals for exercise purposes to help you gain the
2095: essential factoring skills.
2096:
2097: @assignment
2098: Rewrite your definitions until now with locals
2099: @endassignment
2100:
2101: Reference: @ref{Locals}.
2102:
2103:
2104: @node Conditional execution Tutorial, Flags and Comparisons Tutorial, Local Variables Tutorial, Tutorial
2105: @section Conditional execution
2106: @cindex conditionals, tutorial
2107: @cindex if, tutorial
2108:
2109: In Forth you can use control structures only inside colon definitions.
2110: An @code{if}-structure looks like this:
2111:
2112: @example
2113: : abs ( n1 -- +n2 )
2114: dup 0 < if
2115: negate
2116: endif ;
2117: 5 abs .
2118: -5 abs .
2119: @end example
2120:
2121: @code{if} takes a flag from the stack. If the flag is non-zero (true),
2122: the following code is performed, otherwise execution continues after the
2123: @code{endif} (or @code{else}). @code{<} compares the top two stack
2124: elements and prioduces a flag:
2125:
2126: @example
2127: 1 2 < .
2128: 2 1 < .
2129: 1 1 < .
2130: @end example
2131:
2132: Actually the standard name for @code{endif} is @code{then}. This
2133: tutorial presents the examples using @code{endif}, because this is often
2134: less confusing for people familiar with other programming languages
2135: where @code{then} has a different meaning. If your system does not have
2136: @code{endif}, define it with
2137:
2138: @example
2139: : endif postpone then ; immediate
2140: @end example
2141:
2142: You can optionally use an @code{else}-part:
2143:
2144: @example
2145: : min ( n1 n2 -- n )
2146: 2dup < if
2147: drop
2148: else
2149: nip
2150: endif ;
2151: 2 3 min .
2152: 3 2 min .
2153: @end example
2154:
2155: @assignment
2156: Write @code{min} without @code{else}-part (hint: what's the definition
2157: of @code{nip}?).
2158: @endassignment
2159:
2160: Reference: @ref{Selection}.
2161:
2162:
2163: @node Flags and Comparisons Tutorial, General Loops Tutorial, Conditional execution Tutorial, Tutorial
2164: @section Flags and Comparisons
2165: @cindex flags tutorial
2166: @cindex comparison tutorial
2167:
2168: In a false-flag all bits are clear (0 when interpreted as integer). In
2169: a canonical true-flag all bits are set (-1 as a twos-complement signed
2170: integer); in many contexts (e.g., @code{if}) any non-zero value is
2171: treated as true flag.
2172:
2173: @example
2174: false .
2175: true .
2176: true hex u. decimal
2177: @end example
2178:
2179: Comparison words produce canonical flags:
2180:
2181: @example
2182: 1 1 = .
2183: 1 0= .
2184: 0 1 < .
2185: 0 0 < .
2186: -1 1 u< . \ type error, u< interprets -1 as large unsigned number
2187: -1 1 < .
2188: @end example
2189:
2190: Gforth supports all combinations of the prefixes @code{0 u d d0 du f f0}
2191: (or none) and the comparisons @code{= <> < > <= >=}. Only a part of
2192: these combinations are standard (for details see the standard,
2193: @ref{Numeric comparison}, @ref{Floating Point} or @ref{Word Index}).
2194:
2195: You can use @code{and or xor invert} can be used as operations on
2196: canonical flags. Actually they are bitwise operations:
2197:
2198: @example
2199: 1 2 and .
2200: 1 2 or .
2201: 1 3 xor .
2202: 1 invert .
2203: @end example
2204:
2205: You can convert a zero/non-zero flag into a canonical flag with
2206: @code{0<>} (and complement it on the way with @code{0=}).
2207:
2208: @example
2209: 1 0= .
2210: 1 0<> .
2211: @end example
2212:
2213: You can use the all-bits-set feature of canonical flags and the bitwise
2214: operation of the Boolean operations to avoid @code{if}s:
2215:
2216: @example
2217: : foo ( n1 -- n2 )
2218: 0= if
2219: 14
2220: else
2221: 0
2222: endif ;
2223: 0 foo .
2224: 1 foo .
2225:
2226: : foo ( n1 -- n2 )
2227: 0= 14 and ;
2228: 0 foo .
2229: 1 foo .
2230: @end example
2231:
2232: @assignment
2233: Write @code{min} without @code{if}.
2234: @endassignment
2235:
2236: For reference, see @ref{Boolean Flags}, @ref{Numeric comparison}, and
2237: @ref{Bitwise operations}.
2238:
2239:
2240: @node General Loops Tutorial, Counted loops Tutorial, Flags and Comparisons Tutorial, Tutorial
2241: @section General Loops
2242: @cindex loops, indefinite, tutorial
2243:
2244: The endless loop is the most simple one:
2245:
2246: @example
2247: : endless ( -- )
2248: 0 begin
2249: dup . 1+
2250: again ;
2251: endless
2252: @end example
2253:
2254: Terminate this loop by pressing @kbd{Ctrl-C} (in Gforth). @code{begin}
2255: does nothing at run-time, @code{again} jumps back to @code{begin}.
2256:
2257: A loop with one exit at any place looks like this:
2258:
2259: @example
2260: : log2 ( +n1 -- n2 )
2261: \ logarithmus dualis of n1>0, rounded down to the next integer
2262: assert( dup 0> )
2263: 2/ 0 begin
2264: over 0> while
2265: 1+ swap 2/ swap
2266: repeat
2267: nip ;
2268: 7 log2 .
2269: 8 log2 .
2270: @end example
2271:
2272: At run-time @code{while} consumes a flag; if it is 0, execution
2273: continues behind the @code{repeat}; if the flag is non-zero, execution
2274: continues behind the @code{while}. @code{Repeat} jumps back to
2275: @code{begin}, just like @code{again}.
2276:
2277: In Forth there are many combinations/abbreviations, like @code{1+}.
2278: However, @code{2/} is not one of them; it shifts it's argument right by
2279: one bit (arithmetic shift right):
2280:
2281: @example
2282: -5 2 / .
2283: -5 2/ .
2284: @end example
2285:
2286: @code{assert(} is no standard word, but you can get it on systems other
2287: then Gforth by including @file{compat/assert.fs}. You can see what it
2288: does by trying
2289:
2290: @example
2291: 0 log2 .
2292: @end example
2293:
2294: Here's a loop with an exit at the end:
2295:
2296: @example
2297: : log2 ( +n1 -- n2 )
2298: \ logarithmus dualis of n1>0, rounded down to the next integer
2299: assert( dup 0 > )
2300: -1 begin
2301: 1+ swap 2/ swap
2302: over 0 <=
2303: until
2304: nip ;
2305: @end example
2306:
2307: @code{Until} consumes a flag; if it is non-zero, execution continues at
2308: the @code{begin}, otherwise after the @code{until}.
2309:
2310: @assignment
2311: Write a definition for computing the greatest common divisor.
2312: @endassignment
2313:
2314: Reference: @ref{Simple Loops}.
2315:
2316:
2317: @node Counted loops Tutorial, Recursion Tutorial, General Loops Tutorial, Tutorial
2318: @section Counted loops
2319: @cindex loops, counted, tutorial
2320:
2321: @example
2322: : ^ ( n1 u -- n )
2323: \ n = the uth power of u1
2324: 1 swap 0 u+do
2325: over *
2326: loop
2327: nip ;
2328: 3 2 ^ .
2329: 4 3 ^ .
2330: @end example
2331:
2332: @code{U+do} (from @file{compat/loops.fs}, if your Forth system doesn't
2333: have it) takes two numbers of the stack @code{( u3 u4 -- )}, and then
2334: performs the code between @code{u+do} and @code{loop} for @code{u3-u4}
2335: times (or not at all, if @code{u3-u4<0}).
2336:
2337: You can see the stack effect design rules at work in the stack effect of
2338: the loop start words: Since the start value of the loop is more
2339: frequently constant than the end value, the start value is passed on
2340: the top-of-stack.
2341:
2342: You can access the counter of a counted loop with @code{i}:
2343:
2344: @example
2345: : fac ( u -- u! )
2346: 1 swap 1+ 1 u+do
2347: i *
2348: loop ;
2349: 5 fac .
2350: 7 fac .
2351: @end example
2352:
2353: There is also @code{+do}, which expects signed numbers (important for
2354: deciding whether to enter the loop).
2355:
2356: @assignment
2357: Write a definition for computing the nth Fibonacci number.
2358: @endassignment
2359:
2360: You can also use increments other than 1:
2361:
2362: @example
2363: : up2 ( n1 n2 -- )
2364: +do
2365: i .
2366: 2 +loop ;
2367: 10 0 up2
2368:
2369: : down2 ( n1 n2 -- )
2370: -do
2371: i .
2372: 2 -loop ;
2373: 0 10 down2
2374: @end example
2375:
2376: Reference: @ref{Counted Loops}.
2377:
2378:
2379: @node Recursion Tutorial, Leaving definitions or loops Tutorial, Counted loops Tutorial, Tutorial
2380: @section Recursion
2381: @cindex recursion tutorial
2382:
2383: Usually the name of a definition is not visible in the definition; but
2384: earlier definitions are usually visible:
2385:
2386: @example
2387: 1 0 / . \ "Floating-point unidentified fault" in Gforth on most platforms
2388: : / ( n1 n2 -- n )
2389: dup 0= if
2390: -10 throw \ report division by zero
2391: endif
2392: / \ old version
2393: ;
2394: 1 0 /
2395: @end example
2396:
2397: For recursive definitions you can use @code{recursive} (non-standard) or
2398: @code{recurse}:
2399:
2400: @example
2401: : fac1 ( n -- n! ) recursive
2402: dup 0> if
2403: dup 1- fac1 *
2404: else
2405: drop 1
2406: endif ;
2407: 7 fac1 .
2408:
2409: : fac2 ( n -- n! )
2410: dup 0> if
2411: dup 1- recurse *
2412: else
2413: drop 1
2414: endif ;
2415: 8 fac2 .
2416: @end example
2417:
2418: @assignment
2419: Write a recursive definition for computing the nth Fibonacci number.
2420: @endassignment
2421:
2422: Reference (including indirect recursion): @xref{Calls and returns}.
2423:
2424:
2425: @node Leaving definitions or loops Tutorial, Return Stack Tutorial, Recursion Tutorial, Tutorial
2426: @section Leaving definitions or loops
2427: @cindex leaving definitions, tutorial
2428: @cindex leaving loops, tutorial
2429:
2430: @code{EXIT} exits the current definition right away. For every counted
2431: loop that is left in this way, an @code{UNLOOP} has to be performed
2432: before the @code{EXIT}:
2433:
2434: @c !! real examples
2435: @example
2436: : ...
2437: ... u+do
2438: ... if
2439: ... unloop exit
2440: endif
2441: ...
2442: loop
2443: ... ;
2444: @end example
2445:
2446: @code{LEAVE} leaves the innermost counted loop right away:
2447:
2448: @example
2449: : ...
2450: ... u+do
2451: ... if
2452: ... leave
2453: endif
2454: ...
2455: loop
2456: ... ;
2457: @end example
2458:
2459: @c !! example
2460:
2461: Reference: @ref{Calls and returns}, @ref{Counted Loops}.
2462:
2463:
2464: @node Return Stack Tutorial, Memory Tutorial, Leaving definitions or loops Tutorial, Tutorial
2465: @section Return Stack
2466: @cindex return stack tutorial
2467:
2468: In addition to the data stack Forth also has a second stack, the return
2469: stack; most Forth systems store the return addresses of procedure calls
2470: there (thus its name). Programmers can also use this stack:
2471:
2472: @example
2473: : foo ( n1 n2 -- )
2474: .s
2475: >r .s
2476: r@@ .
2477: >r .s
2478: r@@ .
2479: r> .
2480: r@@ .
2481: r> . ;
2482: 1 2 foo
2483: @end example
2484:
2485: @code{>r} takes an element from the data stack and pushes it onto the
2486: return stack; conversely, @code{r>} moves an elementm from the return to
2487: the data stack; @code{r@@} pushes a copy of the top of the return stack
2488: on the return stack.
2489:
2490: Forth programmers usually use the return stack for storing data
2491: temporarily, if using the data stack alone would be too complex, and
2492: factoring and locals are not an option:
2493:
2494: @example
2495: : 2swap ( x1 x2 x3 x4 -- x3 x4 x1 x2 )
2496: rot >r rot r> ;
2497: @end example
2498:
2499: The return address of the definition and the loop control parameters of
2500: counted loops usually reside on the return stack, so you have to take
2501: all items, that you have pushed on the return stack in a colon
2502: definition or counted loop, from the return stack before the definition
2503: or loop ends. You cannot access items that you pushed on the return
2504: stack outside some definition or loop within the definition of loop.
2505:
2506: If you miscount the return stack items, this usually ends in a crash:
2507:
2508: @example
2509: : crash ( n -- )
2510: >r ;
2511: 5 crash
2512: @end example
2513:
2514: You cannot mix using locals and using the return stack (according to the
2515: standard; Gforth has no problem). However, they solve the same
2516: problems, so this shouldn't be an issue.
2517:
2518: @assignment
2519: Can you rewrite any of the definitions you wrote until now in a better
2520: way using the return stack?
2521: @endassignment
2522:
2523: Reference: @ref{Return stack}.
2524:
2525:
2526: @node Memory Tutorial, Characters and Strings Tutorial, Return Stack Tutorial, Tutorial
2527: @section Memory
2528: @cindex memory access/allocation tutorial
2529:
2530: You can create a global variable @code{v} with
2531:
2532: @example
2533: variable v ( -- addr )
2534: @end example
2535:
2536: @code{v} pushes the address of a cell in memory on the stack. This cell
2537: was reserved by @code{variable}. You can use @code{!} (store) to store
2538: values into this cell and @code{@@} (fetch) to load the value from the
2539: stack into memory:
2540:
2541: @example
2542: v .
2543: 5 v ! .s
2544: v @@ .
2545: @end example
2546:
2547: You can see a raw dump of memory with @code{dump}:
2548:
2549: @example
2550: v 1 cells .s dump
2551: @end example
2552:
2553: @code{Cells ( n1 -- n2 )} gives you the number of bytes (or, more
2554: generally, address units (aus)) that @code{n1 cells} occupy. You can
2555: also reserve more memory:
2556:
2557: @example
2558: create v2 20 cells allot
2559: v2 20 cells dump
2560: @end example
2561:
2562: creates a word @code{v2} and reserves 20 uninitialized cells; the
2563: address pushed by @code{v2} points to the start of these 20 cells. You
2564: can use address arithmetic to access these cells:
2565:
2566: @example
2567: 3 v2 5 cells + !
2568: v2 20 cells dump
2569: @end example
2570:
2571: You can reserve and initialize memory with @code{,}:
2572:
2573: @example
2574: create v3
2575: 5 , 4 , 3 , 2 , 1 ,
2576: v3 @@ .
2577: v3 cell+ @@ .
2578: v3 2 cells + @@ .
2579: v3 5 cells dump
2580: @end example
2581:
2582: @assignment
2583: Write a definition @code{vsum ( addr u -- n )} that computes the sum of
2584: @code{u} cells, with the first of these cells at @code{addr}, the next
2585: one at @code{addr cell+} etc.
2586: @endassignment
2587:
2588: You can also reserve memory without creating a new word:
2589:
2590: @example
2591: here 10 cells allot .
2592: here .
2593: @end example
2594:
2595: @code{Here} pushes the start address of the memory area. You should
2596: store it somewhere, or you will have a hard time finding the memory area
2597: again.
2598:
2599: @code{Allot} manages dictionary memory. The dictionary memory contains
2600: the system's data structures for words etc. on Gforth and most other
2601: Forth systems. It is managed like a stack: You can free the memory that
2602: you have just @code{allot}ed with
2603:
2604: @example
2605: -10 cells allot
2606: here .
2607: @end example
2608:
2609: Note that you cannot do this if you have created a new word in the
2610: meantime (because then your @code{allot}ed memory is no longer on the
2611: top of the dictionary ``stack'').
2612:
2613: Alternatively, you can use @code{allocate} and @code{free} which allow
2614: freeing memory in any order:
2615:
2616: @example
2617: 10 cells allocate throw .s
2618: 20 cells allocate throw .s
2619: swap
2620: free throw
2621: free throw
2622: @end example
2623:
2624: The @code{throw}s deal with errors (e.g., out of memory).
2625:
2626: And there is also a
2627: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
2628: garbage collector}, which eliminates the need to @code{free} memory
2629: explicitly.
2630:
2631: Reference: @ref{Memory}.
2632:
2633:
2634: @node Characters and Strings Tutorial, Alignment Tutorial, Memory Tutorial, Tutorial
2635: @section Characters and Strings
2636: @cindex strings tutorial
2637: @cindex characters tutorial
2638:
2639: On the stack characters take up a cell, like numbers. In memory they
2640: have their own size (one 8-bit byte on most systems), and therefore
2641: require their own words for memory access:
2642:
2643: @example
2644: create v4
2645: 104 c, 97 c, 108 c, 108 c, 111 c,
2646: v4 4 chars + c@@ .
2647: v4 5 chars dump
2648: @end example
2649:
2650: The preferred representation of strings on the stack is @code{addr
2651: u-count}, where @code{addr} is the address of the first character and
2652: @code{u-count} is the number of characters in the string.
2653:
2654: @example
2655: v4 5 type
2656: @end example
2657:
2658: You get a string constant with
2659:
2660: @example
2661: s" hello, world" .s
2662: type
2663: @end example
2664:
2665: Make sure you have a space between @code{s"} and the string; @code{s"}
2666: is a normal Forth word and must be delimited with white space (try what
2667: happens when you remove the space).
2668:
2669: However, this interpretive use of @code{s"} is quite restricted: the
2670: string exists only until the next call of @code{s"} (some Forth systems
2671: keep more than one of these strings, but usually they still have a
2672: limited lifetime).
2673:
2674: @example
2675: s" hello," s" world" .s
2676: type
2677: type
2678: @end example
2679:
2680: You can also use @code{s"} in a definition, and the resulting
2681: strings then live forever (well, for as long as the definition):
2682:
2683: @example
2684: : foo s" hello," s" world" ;
2685: foo .s
2686: type
2687: type
2688: @end example
2689:
2690: @assignment
2691: @code{Emit ( c -- )} types @code{c} as character (not a number).
2692: Implement @code{type ( addr u -- )}.
2693: @endassignment
2694:
2695: Reference: @ref{Memory Blocks}.
2696:
2697:
2698: @node Alignment Tutorial, Files Tutorial, Characters and Strings Tutorial, Tutorial
2699: @section Alignment
2700: @cindex alignment tutorial
2701: @cindex memory alignment tutorial
2702:
2703: On many processors cells have to be aligned in memory, if you want to
2704: access them with @code{@@} and @code{!} (and even if the processor does
2705: not require alignment, access to aligned cells is faster).
2706:
2707: @code{Create} aligns @code{here} (i.e., the place where the next
2708: allocation will occur, and that the @code{create}d word points to).
2709: Likewise, the memory produced by @code{allocate} starts at an aligned
2710: address. Adding a number of @code{cells} to an aligned address produces
2711: another aligned address.
2712:
2713: However, address arithmetic involving @code{char+} and @code{chars} can
2714: create an address that is not cell-aligned. @code{Aligned ( addr --
2715: a-addr )} produces the next aligned address:
2716:
2717: @example
2718: v3 char+ aligned .s @@ .
2719: v3 char+ .s @@ .
2720: @end example
2721:
2722: Similarly, @code{align} advances @code{here} to the next aligned
2723: address:
2724:
2725: @example
2726: create v5 97 c,
2727: here .
2728: align here .
2729: 1000 ,
2730: @end example
2731:
2732: Note that you should use aligned addresses even if your processor does
2733: not require them, if you want your program to be portable.
2734:
2735: Reference: @ref{Address arithmetic}.
2736:
2737:
2738: @node Files Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Alignment Tutorial, Tutorial
2739: @section Files
2740: @cindex files tutorial
2741:
2742: This section gives a short introduction into how to use files inside
2743: Forth. It's broken up into five easy steps:
2744:
2745: @enumerate 1
2746: @item Opened an ASCII text file for input
2747: @item Opened a file for output
2748: @item Read input file until string matched (or some other condition matched)
2749: @item Wrote some lines from input ( modified or not) to output
2750: @item Closed the files.
2751: @end enumerate
2752:
2753: @subsection Open file for input
2754:
2755: @example
2756: s" foo.in" r/o open-file throw Value fd-in
2757: @end example
2758:
2759: @subsection Create file for output
2760:
2761: @example
2762: s" foo.out" w/o create-file throw Value fd-out
2763: @end example
2764:
2765: The available file modes are r/o for read-only access, r/w for
2766: read-write access, and w/o for write-only access. You could open both
2767: files with r/w, too, if you like. All file words return error codes; for
2768: most applications, it's best to pass there error codes with @code{throw}
2769: to the outer error handler.
2770:
2771: If you want words for opening and assigning, define them as follows:
2772:
2773: @example
2774: 0 Value fd-in
2775: 0 Value fd-out
2776: : open-input ( addr u -- ) r/o open-file throw to fd-in ;
2777: : open-output ( addr u -- ) w/o create-file throw to fd-out ;
2778: @end example
2779:
2780: Usage example:
2781:
2782: @example
2783: s" foo.in" open-input
2784: s" foo.out" open-output
2785: @end example
2786:
2787: @subsection Scan file for a particular line
2788:
2789: @example
2790: 256 Constant max-line
2791: Create line-buffer max-line 2 + allot
2792:
2793: : scan-file ( addr u -- )
2794: begin
2795: line-buffer max-line fd-in read-line throw
2796: while
2797: >r 2dup line-buffer r> compare 0=
2798: until
2799: else
2800: drop
2801: then
2802: 2drop ;
2803: @end example
2804:
2805: @code{read-line ( addr u1 fd -- u2 flag ior )} reads up to u1 bytes into
2806: the buffer at addr, and returns the number of bytes read, a flag that's
2807: true when the end of file is reached, and an error code.
2808:
2809: @code{compare ( addr1 u1 addr2 u2 -- n )} compares two strings and
2810: returns zero if both strings are equal. It returns a positive number if
2811: the first string is lexically greater, a negative if the second string
2812: is lexically greater.
2813:
2814: We haven't seen this loop here; it has two exits. Since the @code{while}
2815: exits with the number of bytes read on the stack, we have to clean up
2816: that separately; that's after the @code{else}.
2817:
2818: Usage example:
2819:
2820: @example
2821: s" The text I search is here" scan-file
2822: @end example
2823:
2824: @subsection Copy input to output
2825:
2826: @example
2827: : copy-file ( -- )
2828: begin
2829: line-buffer max-line fd-in read-line throw
2830: while
2831: line-buffer swap fd-out write-file throw
2832: repeat ;
2833: @end example
2834:
2835: @subsection Close files
2836:
2837: @example
2838: fd-in close-file throw
2839: fd-out close-file throw
2840: @end example
2841:
2842: Likewise, you can put that into definitions, too:
2843:
2844: @example
2845: : close-input ( -- ) fd-in close-file throw ;
2846: : close-output ( -- ) fd-out close-file throw ;
2847: @end example
2848:
2849: @assignment
2850: How could you modify @code{copy-file} so that it copies until a second line is
2851: matched? Can you write a program that extracts a section of a text file,
2852: given the line that starts and the line that terminates that section?
2853: @endassignment
2854:
2855: @node Interpretation and Compilation Semantics and Immediacy Tutorial, Execution Tokens Tutorial, Files Tutorial, Tutorial
2856: @section Interpretation and Compilation Semantics and Immediacy
2857: @cindex semantics tutorial
2858: @cindex interpretation semantics tutorial
2859: @cindex compilation semantics tutorial
2860: @cindex immediate, tutorial
2861:
2862: When a word is compiled, it behaves differently from being interpreted.
2863: E.g., consider @code{+}:
2864:
2865: @example
2866: 1 2 + .
2867: : foo + ;
2868: @end example
2869:
2870: These two behaviours are known as compilation and interpretation
2871: semantics. For normal words (e.g., @code{+}), the compilation semantics
2872: is to append the interpretation semantics to the currently defined word
2873: (@code{foo} in the example above). I.e., when @code{foo} is executed
2874: later, the interpretation semantics of @code{+} (i.e., adding two
2875: numbers) will be performed.
2876:
2877: However, there are words with non-default compilation semantics, e.g.,
2878: the control-flow words like @code{if}. You can use @code{immediate} to
2879: change the compilation semantics of the last defined word to be equal to
2880: the interpretation semantics:
2881:
2882: @example
2883: : [FOO] ( -- )
2884: 5 . ; immediate
2885:
2886: [FOO]
2887: : bar ( -- )
2888: [FOO] ;
2889: bar
2890: see bar
2891: @end example
2892:
2893: Two conventions to mark words with non-default compilation semnatics are
2894: names with brackets (more frequently used) and to write them all in
2895: upper case (less frequently used).
2896:
2897: In Gforth (and many other systems) you can also remove the
2898: interpretation semantics with @code{compile-only} (the compilation
2899: semantics is derived from the original interpretation semantics):
2900:
2901: @example
2902: : flip ( -- )
2903: 6 . ; compile-only \ but not immediate
2904: flip
2905:
2906: : flop ( -- )
2907: flip ;
2908: flop
2909: @end example
2910:
2911: In this example the interpretation semantics of @code{flop} is equal to
2912: the original interpretation semantics of @code{flip}.
2913:
2914: The text interpreter has two states: in interpret state, it performs the
2915: interpretation semantics of words it encounters; in compile state, it
2916: performs the compilation semantics of these words.
2917:
2918: Among other things, @code{:} switches into compile state, and @code{;}
2919: switches back to interpret state. They contain the factors @code{]}
2920: (switch to compile state) and @code{[} (switch to interpret state), that
2921: do nothing but switch the state.
2922:
2923: @example
2924: : xxx ( -- )
2925: [ 5 . ]
2926: ;
2927:
2928: xxx
2929: see xxx
2930: @end example
2931:
2932: These brackets are also the source of the naming convention mentioned
2933: above.
2934:
2935: Reference: @ref{Interpretation and Compilation Semantics}.
2936:
2937:
2938: @node Execution Tokens Tutorial, Exceptions Tutorial, Interpretation and Compilation Semantics and Immediacy Tutorial, Tutorial
2939: @section Execution Tokens
2940: @cindex execution tokens tutorial
2941: @cindex XT tutorial
2942:
2943: @code{' word} gives you the execution token (XT) of a word. The XT is a
2944: cell representing the interpretation semantics of a word. You can
2945: execute this semantics with @code{execute}:
2946:
2947: @example
2948: ' + .s
2949: 1 2 rot execute .
2950: @end example
2951:
2952: The XT is similar to a function pointer in C. However, parameter
2953: passing through the stack makes it a little more flexible:
2954:
2955: @example
2956: : map-array ( ... addr u xt -- ... )
2957: \ executes xt ( ... x -- ... ) for every element of the array starting
2958: \ at addr and containing u elements
2959: @{ xt @}
2960: cells over + swap ?do
2961: i @@ xt execute
2962: 1 cells +loop ;
2963:
2964: create a 3 , 4 , 2 , -1 , 4 ,
2965: a 5 ' . map-array .s
2966: 0 a 5 ' + map-array .
2967: s" max-n" environment? drop .s
2968: a 5 ' min map-array .
2969: @end example
2970:
2971: You can use map-array with the XTs of words that consume one element
2972: more than they produce. In theory you can also use it with other XTs,
2973: but the stack effect then depends on the size of the array, which is
2974: hard to understand.
2975:
2976: Since XTs are cell-sized, you can store them in memory and manipulate
2977: them on the stack like other cells. You can also compile the XT into a
2978: word with @code{compile,}:
2979:
2980: @example
2981: : foo1 ( n1 n2 -- n )
2982: [ ' + compile, ] ;
2983: see foo
2984: @end example
2985:
2986: This is non-standard, because @code{compile,} has no compilation
2987: semantics in the standard, but it works in good Forth systems. For the
2988: broken ones, use
2989:
2990: @example
2991: : [compile,] compile, ; immediate
2992:
2993: : foo1 ( n1 n2 -- n )
2994: [ ' + ] [compile,] ;
2995: see foo
2996: @end example
2997:
2998: @code{'} is a word with default compilation semantics; it parses the
2999: next word when its interpretation semantics are executed, not during
3000: compilation:
3001:
3002: @example
3003: : foo ( -- xt )
3004: ' ;
3005: see foo
3006: : bar ( ... "word" -- ... )
3007: ' execute ;
3008: see bar
3009: 1 2 bar + .
3010: @end example
3011:
3012: You often want to parse a word during compilation and compile its XT so
3013: it will be pushed on the stack at run-time. @code{[']} does this:
3014:
3015: @example
3016: : xt-+ ( -- xt )
3017: ['] + ;
3018: see xt-+
3019: 1 2 xt-+ execute .
3020: @end example
3021:
3022: Many programmers tend to see @code{'} and the word it parses as one
3023: unit, and expect it to behave like @code{[']} when compiled, and are
3024: confused by the actual behaviour. If you are, just remember that the
3025: Forth system just takes @code{'} as one unit and has no idea that it is
3026: a parsing word (attempts to convenience programmers in this issue have
3027: usually resulted in even worse pitfalls, see
3028: @uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,
3029: @code{State}-smartness---Why it is evil and How to Exorcise it}).
3030:
3031: Note that the state of the interpreter does not come into play when
3032: creating and executing XTs. I.e., even when you execute @code{'} in
3033: compile state, it still gives you the interpretation semantics. And
3034: whatever that state is, @code{execute} performs the semantics
3035: represented by the XT (i.e., for XTs produced with @code{'} the
3036: interpretation semantics).
3037:
3038: Reference: @ref{Tokens for Words}.
3039:
3040:
3041: @node Exceptions Tutorial, Defining Words Tutorial, Execution Tokens Tutorial, Tutorial
3042: @section Exceptions
3043: @cindex exceptions tutorial
3044:
3045: @code{throw ( n -- )} causes an exception unless n is zero.
3046:
3047: @example
3048: 100 throw .s
3049: 0 throw .s
3050: @end example
3051:
3052: @code{catch ( ... xt -- ... n )} behaves similar to @code{execute}, but
3053: it catches exceptions and pushes the number of the exception on the
3054: stack (or 0, if the xt executed without exception). If there was an
3055: exception, the stacks have the same depth as when entering @code{catch}:
3056:
3057: @example
3058: .s
3059: 3 0 ' / catch .s
3060: 3 2 ' / catch .s
3061: @end example
3062:
3063: @assignment
3064: Try the same with @code{execute} instead of @code{catch}.
3065: @endassignment
3066:
3067: @code{Throw} always jumps to the dynamically next enclosing
3068: @code{catch}, even if it has to leave several call levels to achieve
3069: this:
3070:
3071: @example
3072: : foo 100 throw ;
3073: : foo1 foo ." after foo" ;
3074: : bar ['] foo1 catch ;
3075: bar .
3076: @end example
3077:
3078: It is often important to restore a value upon leaving a definition, even
3079: if the definition is left through an exception. You can ensure this
3080: like this:
3081:
3082: @example
3083: : ...
3084: save-x
3085: ['] word-changing-x catch ( ... n )
3086: restore-x
3087: ( ... n ) throw ;
3088: @end example
3089:
3090: Gforth provides an alternative syntax in addition to @code{catch}:
3091: @code{try ... recover ... endtry}. If the code between @code{try} and
3092: @code{recover} has an exception, the stack depths are restored, the
3093: exception number is pushed on the stack, and the code between
3094: @code{recover} and @code{endtry} is performed. E.g., the definition for
3095: @code{catch} is
3096:
3097: @example
3098: : catch ( x1 .. xn xt -- y1 .. ym 0 / z1 .. zn error ) \ exception
3099: try
3100: execute 0
3101: recover
3102: nip
3103: endtry ;
3104: @end example
3105:
3106: The equivalent to the restoration code above is
3107:
3108: @example
3109: : ...
3110: save-x
3111: try
3112: word-changing-x
3113: end-try
3114: restore-x
3115: throw ;
3116: @end example
3117:
3118: As you can see, the @code{recover} part is optional.
3119:
3120: Reference: @ref{Exception Handling}.
3121:
3122:
3123: @node Defining Words Tutorial, Arrays and Records Tutorial, Exceptions Tutorial, Tutorial
3124: @section Defining Words
3125: @cindex defining words tutorial
3126: @cindex does> tutorial
3127: @cindex create...does> tutorial
3128:
3129: @c before semantics?
3130:
3131: @code{:}, @code{create}, and @code{variable} are definition words: They
3132: define other words. @code{Constant} is another definition word:
3133:
3134: @example
3135: 5 constant foo
3136: foo .
3137: @end example
3138:
3139: You can also use the prefixes @code{2} (double-cell) and @code{f}
3140: (floating point) with @code{variable} and @code{constant}.
3141:
3142: You can also define your own defining words. E.g.:
3143:
3144: @example
3145: : variable ( "name" -- )
3146: create 0 , ;
3147: @end example
3148:
3149: You can also define defining words that create words that do something
3150: other than just producing their address:
3151:
3152: @example
3153: : constant ( n "name" -- )
3154: create ,
3155: does> ( -- n )
3156: ( addr ) @@ ;
3157:
3158: 5 constant foo
3159: foo .
3160: @end example
3161:
3162: The definition of @code{constant} above ends at the @code{does>}; i.e.,
3163: @code{does>} replaces @code{;}, but it also does something else: It
3164: changes the last defined word such that it pushes the address of the
3165: body of the word and then performs the code after the @code{does>}
3166: whenever it is called.
3167:
3168: In the example above, @code{constant} uses @code{,} to store 5 into the
3169: body of @code{foo}. When @code{foo} executes, it pushes the address of
3170: the body onto the stack, then (in the code after the @code{does>})
3171: fetches the 5 from there.
3172:
3173: The stack comment near the @code{does>} reflects the stack effect of the
3174: defined word, not the stack effect of the code after the @code{does>}
3175: (the difference is that the code expects the address of the body that
3176: the stack comment does not show).
3177:
3178: You can use these definition words to do factoring in cases that involve
3179: (other) definition words. E.g., a field offset is always added to an
3180: address. Instead of defining
3181:
3182: @example
3183: 2 cells constant offset-field1
3184: @end example
3185:
3186: and using this like
3187:
3188: @example
3189: ( addr ) offset-field1 +
3190: @end example
3191:
3192: you can define a definition word
3193:
3194: @example
3195: : simple-field ( n "name" -- )
3196: create ,
3197: does> ( n1 -- n1+n )
3198: ( addr ) @@ + ;
3199: @end example
3200:
3201: Definition and use of field offsets now look like this:
3202:
3203: @example
3204: 2 cells simple-field field1
3205: create mystruct 4 cells allot
3206: mystruct .s field1 .s drop
3207: @end example
3208:
3209: If you want to do something with the word without performing the code
3210: after the @code{does>}, you can access the body of a @code{create}d word
3211: with @code{>body ( xt -- addr )}:
3212:
3213: @example
3214: : value ( n "name" -- )
3215: create ,
3216: does> ( -- n1 )
3217: @@ ;
3218: : to ( n "name" -- )
3219: ' >body ! ;
3220:
3221: 5 value foo
3222: foo .
3223: 7 to foo
3224: foo .
3225: @end example
3226:
3227: @assignment
3228: Define @code{defer ( "name" -- )}, which creates a word that stores an
3229: XT (at the start the XT of @code{abort}), and upon execution
3230: @code{execute}s the XT. Define @code{is ( xt "name" -- )} that stores
3231: @code{xt} into @code{name}, a word defined with @code{defer}. Indirect
3232: recursion is one application of @code{defer}.
3233: @endassignment
3234:
3235: Reference: @ref{User-defined Defining Words}.
3236:
3237:
3238: @node Arrays and Records Tutorial, POSTPONE Tutorial, Defining Words Tutorial, Tutorial
3239: @section Arrays and Records
3240: @cindex arrays tutorial
3241: @cindex records tutorial
3242: @cindex structs tutorial
3243:
3244: Forth has no standard words for defining data structures such as arrays
3245: and records (structs in C terminology), but you can build them yourself
3246: based on address arithmetic. You can also define words for defining
3247: arrays and records (@pxref{Defining Words Tutorial,, Defining Words}).
3248:
3249: One of the first projects a Forth newcomer sets out upon when learning
3250: about defining words is an array defining word (possibly for
3251: n-dimensional arrays). Go ahead and do it, I did it, too; you will
3252: learn something from it. However, don't be disappointed when you later
3253: learn that you have little use for these words (inappropriate use would
3254: be even worse). I have not yet found a set of useful array words yet;
3255: the needs are just too diverse, and named, global arrays (the result of
3256: naive use of defining words) are often not flexible enough (e.g.,
3257: consider how to pass them as parameters). Another such project is a set
3258: of words to help dealing with strings.
3259:
3260: On the other hand, there is a useful set of record words, and it has
3261: been defined in @file{compat/struct.fs}; these words are predefined in
3262: Gforth. They are explained in depth elsewhere in this manual (see
3263: @pxref{Structures}). The @code{simple-field} example above is
3264: simplified variant of fields in this package.
3265:
3266:
3267: @node POSTPONE Tutorial, Literal Tutorial, Arrays and Records Tutorial, Tutorial
3268: @section @code{POSTPONE}
3269: @cindex postpone tutorial
3270:
3271: You can compile the compilation semantics (instead of compiling the
3272: interpretation semantics) of a word with @code{POSTPONE}:
3273:
3274: @example
3275: : MY-+ ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3276: POSTPONE + ; immediate
3277: : foo ( n1 n2 -- n )
3278: MY-+ ;
3279: 1 2 foo .
3280: see foo
3281: @end example
3282:
3283: During the definition of @code{foo} the text interpreter performs the
3284: compilation semantics of @code{MY-+}, which performs the compilation
3285: semantics of @code{+}, i.e., it compiles @code{+} into @code{foo}.
3286:
3287: This example also displays separate stack comments for the compilation
3288: semantics and for the stack effect of the compiled code. For words with
3289: default compilation semantics these stack effects are usually not
3290: displayed; the stack effect of the compilation semantics is always
3291: @code{( -- )} for these words, the stack effect for the compiled code is
3292: the stack effect of the interpretation semantics.
3293:
3294: Note that the state of the interpreter does not come into play when
3295: performing the compilation semantics in this way. You can also perform
3296: it interpretively, e.g.:
3297:
3298: @example
3299: : foo2 ( n1 n2 -- n )
3300: [ MY-+ ] ;
3301: 1 2 foo .
3302: see foo
3303: @end example
3304:
3305: However, there are some broken Forth systems where this does not always
3306: work, and therefore this practice was been declared non-standard in
3307: 1999.
3308: @c !! repair.fs
3309:
3310: Here is another example for using @code{POSTPONE}:
3311:
3312: @example
3313: : MY-- ( Compilation: -- ; Run-time of compiled code: n1 n2 -- n )
3314: POSTPONE negate POSTPONE + ; immediate compile-only
3315: : bar ( n1 n2 -- n )
3316: MY-- ;
3317: 2 1 bar .
3318: see bar
3319: @end example
3320:
3321: You can define @code{ENDIF} in this way:
3322:
3323: @example
3324: : ENDIF ( Compilation: orig -- )
3325: POSTPONE then ; immediate
3326: @end example
3327:
3328: @assignment
3329: Write @code{MY-2DUP} that has compilation semantics equivalent to
3330: @code{2dup}, but compiles @code{over over}.
3331: @endassignment
3332:
3333: @c !! @xref{Macros} for reference
3334:
3335:
3336: @node Literal Tutorial, Advanced macros Tutorial, POSTPONE Tutorial, Tutorial
3337: @section @code{Literal}
3338: @cindex literal tutorial
3339:
3340: You cannot @code{POSTPONE} numbers:
3341:
3342: @example
3343: : [FOO] POSTPONE 500 ; immediate
3344: @end example
3345:
3346: Instead, you can use @code{LITERAL (compilation: n --; run-time: -- n )}:
3347:
3348: @example
3349: : [FOO] ( compilation: --; run-time: -- n )
3350: 500 POSTPONE literal ; immediate
3351:
3352: : flip [FOO] ;
3353: flip .
3354: see flip
3355: @end example
3356:
3357: @code{LITERAL} consumes a number at compile-time (when it's compilation
3358: semantics are executed) and pushes it at run-time (when the code it
3359: compiled is executed). A frequent use of @code{LITERAL} is to compile a
3360: number computed at compile time into the current word:
3361:
3362: @example
3363: : bar ( -- n )
3364: [ 2 2 + ] literal ;
3365: see bar
3366: @end example
3367:
3368: @assignment
3369: Write @code{]L} which allows writing the example above as @code{: bar (
3370: -- n ) [ 2 2 + ]L ;}
3371: @endassignment
3372:
3373: @c !! @xref{Macros} for reference
3374:
3375:
3376: @node Advanced macros Tutorial, Compilation Tokens Tutorial, Literal Tutorial, Tutorial
3377: @section Advanced macros
3378: @cindex macros, advanced tutorial
3379: @cindex run-time code generation, tutorial
3380:
3381: Reconsider @code{map-array} from @ref{Execution Tokens Tutorial,,
3382: Execution Tokens}. It frequently performs @code{execute}, a relatively
3383: expensive operation in some Forth implementations. You can use
3384: @code{compile,} and @code{POSTPONE} to eliminate these @code{execute}s
3385: and produce a word that contains the word to be performed directly:
3386:
3387: @c use ]] ... [[
3388: @example
3389: : compile-map-array ( compilation: xt -- ; run-time: ... addr u -- ... )
3390: \ at run-time, execute xt ( ... x -- ... ) for each element of the
3391: \ array beginning at addr and containing u elements
3392: @{ xt @}
3393: POSTPONE cells POSTPONE over POSTPONE + POSTPONE swap POSTPONE ?do
3394: POSTPONE i POSTPONE @@ xt compile,
3395: 1 cells POSTPONE literal POSTPONE +loop ;
3396:
3397: : sum-array ( addr u -- n )
3398: 0 rot rot [ ' + compile-map-array ] ;
3399: see sum-array
3400: a 5 sum-array .
3401: @end example
3402:
3403: You can use the full power of Forth for generating the code; here's an
3404: example where the code is generated in a loop:
3405:
3406: @example
3407: : compile-vmul-step ( compilation: n --; run-time: n1 addr1 -- n2 addr2 )
3408: \ n2=n1+(addr1)*n, addr2=addr1+cell
3409: POSTPONE tuck POSTPONE @@
3410: POSTPONE literal POSTPONE * POSTPONE +
3411: POSTPONE swap POSTPONE cell+ ;
3412:
3413: : compile-vmul ( compilation: addr1 u -- ; run-time: addr2 -- n )
3414: \ n=v1*v2 (inner product), where the v_i are represented as addr_i u
3415: 0 postpone literal postpone swap
3416: [ ' compile-vmul-step compile-map-array ]
3417: postpone drop ;
3418: see compile-vmul
3419:
3420: : a-vmul ( addr -- n )
3421: \ n=a*v, where v is a vector that's as long as a and starts at addr
3422: [ a 5 compile-vmul ] ;
3423: see a-vmul
3424: a a-vmul .
3425: @end example
3426:
3427: This example uses @code{compile-map-array} to show off, but you could
3428: also use @code{map-array} instead (try it now!).
3429:
3430: You can use this technique for efficient multiplication of large
3431: matrices. In matrix multiplication, you multiply every line of one
3432: matrix with every column of the other matrix. You can generate the code
3433: for one line once, and use it for every column. The only downside of
3434: this technique is that it is cumbersome to recover the memory consumed
3435: by the generated code when you are done (and in more complicated cases
3436: it is not possible portably).
3437:
3438: @c !! @xref{Macros} for reference
3439:
3440:
3441: @node Compilation Tokens Tutorial, Wordlists and Search Order Tutorial, Advanced macros Tutorial, Tutorial
3442: @section Compilation Tokens
3443: @cindex compilation tokens, tutorial
3444: @cindex CT, tutorial
3445:
3446: This section is Gforth-specific. You can skip it.
3447:
3448: @code{' word compile,} compiles the interpretation semantics. For words
3449: with default compilation semantics this is the same as performing the
3450: compilation semantics. To represent the compilation semantics of other
3451: words (e.g., words like @code{if} that have no interpretation
3452: semantics), Gforth has the concept of a compilation token (CT,
3453: consisting of two cells), and words @code{comp'} and @code{[comp']}.
3454: You can perform the compilation semantics represented by a CT with
3455: @code{execute}:
3456:
3457: @example
3458: : foo2 ( n1 n2 -- n )
3459: [ comp' + execute ] ;
3460: see foo
3461: @end example
3462:
3463: You can compile the compilation semantics represented by a CT with
3464: @code{postpone,}:
3465:
3466: @example
3467: : foo3 ( -- )
3468: [ comp' + postpone, ] ;
3469: see foo3
3470: @end example
3471:
3472: @code{[ comp' word postpone, ]} is equivalent to @code{POSTPONE word}.
3473: @code{comp'} is particularly useful for words that have no
3474: interpretation semantics:
3475:
3476: @example
3477: ' if
3478: comp' if .s 2drop
3479: @end example
3480:
3481: Reference: @ref{Tokens for Words}.
3482:
3483:
3484: @node Wordlists and Search Order Tutorial, , Compilation Tokens Tutorial, Tutorial
3485: @section Wordlists and Search Order
3486: @cindex wordlists tutorial
3487: @cindex search order, tutorial
3488:
3489: The dictionary is not just a memory area that allows you to allocate
3490: memory with @code{allot}, it also contains the Forth words, arranged in
3491: several wordlists. When searching for a word in a wordlist,
3492: conceptually you start searching at the youngest and proceed towards
3493: older words (in reality most systems nowadays use hash-tables); i.e., if
3494: you define a word with the same name as an older word, the new word
3495: shadows the older word.
3496:
3497: Which wordlists are searched in which order is determined by the search
3498: order. You can display the search order with @code{order}. It displays
3499: first the search order, starting with the wordlist searched first, then
3500: it displays the wordlist that will contain newly defined words.
3501:
3502: You can create a new, empty wordlist with @code{wordlist ( -- wid )}:
3503:
3504: @example
3505: wordlist constant mywords
3506: @end example
3507:
3508: @code{Set-current ( wid -- )} sets the wordlist that will contain newly
3509: defined words (the @emph{current} wordlist):
3510:
3511: @example
3512: mywords set-current
3513: order
3514: @end example
3515:
3516: Gforth does not display a name for the wordlist in @code{mywords}
3517: because this wordlist was created anonymously with @code{wordlist}.
3518:
3519: You can get the current wordlist with @code{get-current ( -- wid)}. If
3520: you want to put something into a specific wordlist without overall
3521: effect on the current wordlist, this typically looks like this:
3522:
3523: @example
3524: get-current mywords set-current ( wid )
3525: create someword
3526: ( wid ) set-current
3527: @end example
3528:
3529: You can write the search order with @code{set-order ( wid1 .. widn n --
3530: )} and read it with @code{get-order ( -- wid1 .. widn n )}. The first
3531: searched wordlist is topmost.
3532:
3533: @example
3534: get-order mywords swap 1+ set-order
3535: order
3536: @end example
3537:
3538: Yes, the order of wordlists in the output of @code{order} is reversed
3539: from stack comments and the output of @code{.s} and thus unintuitive.
3540:
3541: @assignment
3542: Define @code{>order ( wid -- )} with adds @code{wid} as first searched
3543: wordlist to the search order. Define @code{previous ( -- )}, which
3544: removes the first searched wordlist from the search order. Experiment
3545: with boundary conditions (you will see some crashes or situations that
3546: are hard or impossible to leave).
3547: @endassignment
3548:
3549: The search order is a powerful foundation for providing features similar
3550: to Modula-2 modules and C++ namespaces. However, trying to modularize
3551: programs in this way has disadvantages for debugging and reuse/factoring
3552: that overcome the advantages in my experience (I don't do huge projects,
3553: though). These disadvantages are not so clear in other
3554: languages/programming environments, because these languages are not so
3555: strong in debugging and reuse.
3556:
3557: @c !! example
3558:
3559: Reference: @ref{Word Lists}.
3560:
3561: @c ******************************************************************
3562: @node Introduction, Words, Tutorial, Top
3563: @comment node-name, next, previous, up
3564: @chapter An Introduction to ANS Forth
3565: @cindex Forth - an introduction
3566:
3567: The difference of this chapter from the Tutorial (@pxref{Tutorial}) is
3568: that it is slower-paced in its examples, but uses them to dive deep into
3569: explaining Forth internals (not covered by the Tutorial). Apart from
3570: that, this chapter covers far less material. It is suitable for reading
3571: without using a computer.
3572:
3573: The primary purpose of this manual is to document Gforth. However, since
3574: Forth is not a widely-known language and there is a lack of up-to-date
3575: teaching material, it seems worthwhile to provide some introductory
3576: material. For other sources of Forth-related
3577: information, see @ref{Forth-related information}.
3578:
3579: The examples in this section should work on any ANS Forth; the
3580: output shown was produced using Gforth. Each example attempts to
3581: reproduce the exact output that Gforth produces. If you try out the
3582: examples (and you should), what you should type is shown @kbd{like this}
3583: and Gforth's response is shown @code{like this}. The single exception is
3584: that, where the example shows @key{RET} it means that you should
3585: press the ``carriage return'' key. Unfortunately, some output formats for
3586: this manual cannot show the difference between @kbd{this} and
3587: @code{this} which will make trying out the examples harder (but not
3588: impossible).
3589:
3590: Forth is an unusual language. It provides an interactive development
3591: environment which includes both an interpreter and compiler. Forth
3592: programming style encourages you to break a problem down into many
3593: @cindex factoring
3594: small fragments (@dfn{factoring}), and then to develop and test each
3595: fragment interactively. Forth advocates assert that breaking the
3596: edit-compile-test cycle used by conventional programming languages can
3597: lead to great productivity improvements.
3598:
3599: @menu
3600: * Introducing the Text Interpreter::
3601: * Stacks and Postfix notation::
3602: * Your first definition::
3603: * How does that work?::
3604: * Forth is written in Forth::
3605: * Review - elements of a Forth system::
3606: * Where to go next::
3607: * Exercises::
3608: @end menu
3609:
3610: @comment ----------------------------------------------
3611: @node Introducing the Text Interpreter, Stacks and Postfix notation, Introduction, Introduction
3612: @section Introducing the Text Interpreter
3613: @cindex text interpreter
3614: @cindex outer interpreter
3615:
3616: @c IMO this is too detailed and the pace is too slow for
3617: @c an introduction. If you know German, take a look at
3618: @c http://www.complang.tuwien.ac.at/anton/lvas/skriptum-stack.html
3619: @c to see how I do it - anton
3620:
3621: @c nac-> Where I have accepted your comments 100% and modified the text
3622: @c accordingly, I have deleted your comments. Elsewhere I have added a
3623: @c response like this to attempt to rationalise what I have done. Of
3624: @c course, this is a very clumsy mechanism for something that would be
3625: @c done far more efficiently over a beer. Please delete any dialogue
3626: @c you consider closed.
3627:
3628: When you invoke the Forth image, you will see a startup banner printed
3629: and nothing else (if you have Gforth installed on your system, try
3630: invoking it now, by typing @kbd{gforth@key{RET}}). Forth is now running
3631: its command line interpreter, which is called the @dfn{Text Interpreter}
3632: (also known as the @dfn{Outer Interpreter}). (You will learn a lot
3633: about the text interpreter as you read through this chapter, for more
3634: detail @pxref{The Text Interpreter}).
3635:
3636: Although it's not obvious, Forth is actually waiting for your
3637: input. Type a number and press the @key{RET} key:
3638:
3639: @example
3640: @kbd{45@key{RET}} ok
3641: @end example
3642:
3643: Rather than give you a prompt to invite you to input something, the text
3644: interpreter prints a status message @i{after} it has processed a line
3645: of input. The status message in this case (``@code{ ok}'' followed by
3646: carriage-return) indicates that the text interpreter was able to process
3647: all of your input successfully. Now type something illegal:
3648:
3649: @example
3650: @kbd{qwer341@key{RET}}
3651: :1: Undefined word
3652: qwer341
3653: ^^^^^^^
3654: $400D2BA8 Bounce
3655: $400DBDA8 no.extensions
3656: @end example
3657:
3658: The exact text, other than the ``Undefined word'' may differ slightly on
3659: your system, but the effect is the same; when the text interpreter
3660: detects an error, it discards any remaining text on a line, resets
3661: certain internal state and prints an error message. For a detailed description of error messages see @ref{Error
3662: messages}.
3663:
3664: The text interpreter waits for you to press carriage-return, and then
3665: processes your input line. Starting at the beginning of the line, it
3666: breaks the line into groups of characters separated by spaces. For each
3667: group of characters in turn, it makes two attempts to do something:
3668:
3669: @itemize @bullet
3670: @item
3671: @cindex name dictionary
3672: It tries to treat it as a command. It does this by searching a @dfn{name
3673: dictionary}. If the group of characters matches an entry in the name
3674: dictionary, the name dictionary provides the text interpreter with
3675: information that allows the text interpreter perform some actions. In
3676: Forth jargon, we say that the group
3677: @cindex word
3678: @cindex definition
3679: @cindex execution token
3680: @cindex xt
3681: of characters names a @dfn{word}, that the dictionary search returns an
3682: @dfn{execution token (xt)} corresponding to the @dfn{definition} of the
3683: word, and that the text interpreter executes the xt. Often, the terms
3684: @dfn{word} and @dfn{definition} are used interchangeably.
3685: @item
3686: If the text interpreter fails to find a match in the name dictionary, it
3687: tries to treat the group of characters as a number in the current number
3688: base (when you start up Forth, the current number base is base 10). If
3689: the group of characters legitimately represents a number, the text
3690: interpreter pushes the number onto a stack (we'll learn more about that
3691: in the next section).
3692: @end itemize
3693:
3694: If the text interpreter is unable to do either of these things with any
3695: group of characters, it discards the group of characters and the rest of
3696: the line, then prints an error message. If the text interpreter reaches
3697: the end of the line without error, it prints the status message ``@code{ ok}''
3698: followed by carriage-return.
3699:
3700: This is the simplest command we can give to the text interpreter:
3701:
3702: @example
3703: @key{RET} ok
3704: @end example
3705:
3706: The text interpreter did everything we asked it to do (nothing) without
3707: an error, so it said that everything is ``@code{ ok}''. Try a slightly longer
3708: command:
3709:
3710: @example
3711: @kbd{12 dup fred dup@key{RET}}
3712: :1: Undefined word
3713: 12 dup fred dup
3714: ^^^^
3715: $400D2BA8 Bounce
3716: $400DBDA8 no.extensions
3717: @end example
3718:
3719: When you press the carriage-return key, the text interpreter starts to
3720: work its way along the line:
3721:
3722: @itemize @bullet
3723: @item
3724: When it gets to the space after the @code{2}, it takes the group of
3725: characters @code{12} and looks them up in the name
3726: dictionary@footnote{We can't tell if it found them or not, but assume
3727: for now that it did not}. There is no match for this group of characters
3728: in the name dictionary, so it tries to treat them as a number. It is
3729: able to do this successfully, so it puts the number, 12, ``on the stack''
3730: (whatever that means).
3731: @item
3732: The text interpreter resumes scanning the line and gets the next group
3733: of characters, @code{dup}. It looks it up in the name dictionary and
3734: (you'll have to take my word for this) finds it, and executes the word
3735: @code{dup} (whatever that means).
3736: @item
3737: Once again, the text interpreter resumes scanning the line and gets the
3738: group of characters @code{fred}. It looks them up in the name
3739: dictionary, but can't find them. It tries to treat them as a number, but
3740: they don't represent any legal number.
3741: @end itemize
3742:
3743: At this point, the text interpreter gives up and prints an error
3744: message. The error message shows exactly how far the text interpreter
3745: got in processing the line. In particular, it shows that the text
3746: interpreter made no attempt to do anything with the final character
3747: group, @code{dup}, even though we have good reason to believe that the
3748: text interpreter would have no problem looking that word up and
3749: executing it a second time.
3750:
3751:
3752: @comment ----------------------------------------------
3753: @node Stacks and Postfix notation, Your first definition, Introducing the Text Interpreter, Introduction
3754: @section Stacks, postfix notation and parameter passing
3755: @cindex text interpreter
3756: @cindex outer interpreter
3757:
3758: In procedural programming languages (like C and Pascal), the
3759: building-block of programs is the @dfn{function} or @dfn{procedure}. These
3760: functions or procedures are called with @dfn{explicit parameters}. For
3761: example, in C we might write:
3762:
3763: @example
3764: total = total + new_volume(length,height,depth);
3765: @end example
3766:
3767: @noindent
3768: where new_volume is a function-call to another piece of code, and total,
3769: length, height and depth are all variables. length, height and depth are
3770: parameters to the function-call.
3771:
3772: In Forth, the equivalent of the function or procedure is the
3773: @dfn{definition} and parameters are implicitly passed between
3774: definitions using a shared stack that is visible to the
3775: programmer. Although Forth does support variables, the existence of the
3776: stack means that they are used far less often than in most other
3777: programming languages. When the text interpreter encounters a number, it
3778: will place (@dfn{push}) it on the stack. There are several stacks (the
3779: actual number is implementation-dependent ...) and the particular stack
3780: used for any operation is implied unambiguously by the operation being
3781: performed. The stack used for all integer operations is called the @dfn{data
3782: stack} and, since this is the stack used most commonly, references to
3783: ``the data stack'' are often abbreviated to ``the stack''.
3784:
3785: The stacks have a last-in, first-out (LIFO) organisation. If you type:
3786:
3787: @example
3788: @kbd{1 2 3@key{RET}} ok
3789: @end example
3790:
3791: Then this instructs the text interpreter to placed three numbers on the
3792: (data) stack. An analogy for the behaviour of the stack is to take a
3793: pack of playing cards and deal out the ace (1), 2 and 3 into a pile on
3794: the table. The 3 was the last card onto the pile (``last-in'') and if
3795: you take a card off the pile then, unless you're prepared to fiddle a
3796: bit, the card that you take off will be the 3 (``first-out''). The
3797: number that will be first-out of the stack is called the @dfn{top of
3798: stack}, which
3799: @cindex TOS definition
3800: is often abbreviated to @dfn{TOS}.
3801:
3802: To understand how parameters are passed in Forth, consider the
3803: behaviour of the definition @code{+} (pronounced ``plus''). You will not
3804: be surprised to learn that this definition performs addition. More
3805: precisely, it adds two number together and produces a result. Where does
3806: it get the two numbers from? It takes the top two numbers off the
3807: stack. Where does it place the result? On the stack. You can act-out the
3808: behaviour of @code{+} with your playing cards like this:
3809:
3810: @itemize @bullet
3811: @item
3812: Pick up two cards from the stack on the table
3813: @item
3814: Stare at them intently and ask yourself ``what @i{is} the sum of these two
3815: numbers''
3816: @item
3817: Decide that the answer is 5
3818: @item
3819: Shuffle the two cards back into the pack and find a 5
3820: @item
3821: Put a 5 on the remaining ace that's on the table.
3822: @end itemize
3823:
3824: If you don't have a pack of cards handy but you do have Forth running,
3825: you can use the definition @code{.s} to show the current state of the stack,
3826: without affecting the stack. Type:
3827:
3828: @example
3829: @kbd{clearstack 1 2 3@key{RET}} ok
3830: @kbd{.s@key{RET}} <3> 1 2 3 ok
3831: @end example
3832:
3833: The text interpreter looks up the word @code{clearstack} and executes
3834: it; it tidies up the stack and removes any entries that may have been
3835: left on it by earlier examples. The text interpreter pushes each of the
3836: three numbers in turn onto the stack. Finally, the text interpreter
3837: looks up the word @code{.s} and executes it. The effect of executing
3838: @code{.s} is to print the ``<3>'' (the total number of items on the stack)
3839: followed by a list of all the items on the stack; the item on the far
3840: right-hand side is the TOS.
3841:
3842: You can now type:
3843:
3844: @example
3845: @kbd{+ .s@key{RET}} <2> 1 5 ok
3846: @end example
3847:
3848: @noindent
3849: which is correct; there are now 2 items on the stack and the result of
3850: the addition is 5.
3851:
3852: If you're playing with cards, try doing a second addition: pick up the
3853: two cards, work out that their sum is 6, shuffle them into the pack,
3854: look for a 6 and place that on the table. You now have just one item on
3855: the stack. What happens if you try to do a third addition? Pick up the
3856: first card, pick up the second card -- ah! There is no second card. This
3857: is called a @dfn{stack underflow} and consitutes an error. If you try to
3858: do the same thing with Forth it will report an error (probably a Stack
3859: Underflow or an Invalid Memory Address error).
3860:
3861: The opposite situation to a stack underflow is a @dfn{stack overflow},
3862: which simply accepts that there is a finite amount of storage space
3863: reserved for the stack. To stretch the playing card analogy, if you had
3864: enough packs of cards and you piled the cards up on the table, you would
3865: eventually be unable to add another card; you'd hit the ceiling. Gforth
3866: allows you to set the maximum size of the stacks. In general, the only
3867: time that you will get a stack overflow is because a definition has a
3868: bug in it and is generating data on the stack uncontrollably.
3869:
3870: There's one final use for the playing card analogy. If you model your
3871: stack using a pack of playing cards, the maximum number of items on
3872: your stack will be 52 (I assume you didn't use the Joker). The maximum
3873: @i{value} of any item on the stack is 13 (the King). In fact, the only
3874: possible numbers are positive integer numbers 1 through 13; you can't
3875: have (for example) 0 or 27 or 3.52 or -2. If you change the way you
3876: think about some of the cards, you can accommodate different
3877: numbers. For example, you could think of the Jack as representing 0,
3878: the Queen as representing -1 and the King as representing -2. Your
3879: @i{range} remains unchanged (you can still only represent a total of 13
3880: numbers) but the numbers that you can represent are -2 through 10.
3881:
3882: In that analogy, the limit was the amount of information that a single
3883: stack entry could hold, and Forth has a similar limit. In Forth, the
3884: size of a stack entry is called a @dfn{cell}. The actual size of a cell is
3885: implementation dependent and affects the maximum value that a stack
3886: entry can hold. A Standard Forth provides a cell size of at least
3887: 16-bits, and most desktop systems use a cell size of 32-bits.
3888:
3889: Forth does not do any type checking for you, so you are free to
3890: manipulate and combine stack items in any way you wish. A convenient way
3891: of treating stack items is as 2's complement signed integers, and that
3892: is what Standard words like @code{+} do. Therefore you can type:
3893:
3894: @example
3895: @kbd{-5 12 + .s@key{RET}} <1> 7 ok
3896: @end example
3897:
3898: If you use numbers and definitions like @code{+} in order to turn Forth
3899: into a great big pocket calculator, you will realise that it's rather
3900: different from a normal calculator. Rather than typing 2 + 3 = you had
3901: to type 2 3 + (ignore the fact that you had to use @code{.s} to see the
3902: result). The terminology used to describe this difference is to say that
3903: your calculator uses @dfn{Infix Notation} (parameters and operators are
3904: mixed) whilst Forth uses @dfn{Postfix Notation} (parameters and
3905: operators are separate), also called @dfn{Reverse Polish Notation}.
3906:
3907: Whilst postfix notation might look confusing to begin with, it has
3908: several important advantages:
3909:
3910: @itemize @bullet
3911: @item
3912: it is unambiguous
3913: @item
3914: it is more concise
3915: @item
3916: it fits naturally with a stack-based system
3917: @end itemize
3918:
3919: To examine these claims in more detail, consider these sums:
3920:
3921: @example
3922: 6 + 5 * 4 =
3923: 4 * 5 + 6 =
3924: @end example
3925:
3926: If you're just learning maths or your maths is very rusty, you will
3927: probably come up with the answer 44 for the first and 26 for the
3928: second. If you are a bit of a whizz at maths you will remember the
3929: @i{convention} that multiplication takes precendence over addition, and
3930: you'd come up with the answer 26 both times. To explain the answer 26
3931: to someone who got the answer 44, you'd probably rewrite the first sum
3932: like this:
3933:
3934: @example
3935: 6 + (5 * 4) =
3936: @end example
3937:
3938: If what you really wanted was to perform the addition before the
3939: multiplication, you would have to use parentheses to force it.
3940:
3941: If you did the first two sums on a pocket calculator you would probably
3942: get the right answers, unless you were very cautious and entered them using
3943: these keystroke sequences:
3944:
3945: 6 + 5 = * 4 =
3946: 4 * 5 = + 6 =
3947:
3948: Postfix notation is unambiguous because the order that the operators
3949: are applied is always explicit; that also means that parentheses are
3950: never required. The operators are @i{active} (the act of quoting the
3951: operator makes the operation occur) which removes the need for ``=''.
3952:
3953: The sum 6 + 5 * 4 can be written (in postfix notation) in two
3954: equivalent ways:
3955:
3956: @example
3957: 6 5 4 * + or:
3958: 5 4 * 6 +
3959: @end example
3960:
3961: An important thing that you should notice about this notation is that
3962: the @i{order} of the numbers does not change; if you want to subtract
3963: 2 from 10 you type @code{10 2 -}.
3964:
3965: The reason that Forth uses postfix notation is very simple to explain: it
3966: makes the implementation extremely simple, and it follows naturally from
3967: using the stack as a mechanism for passing parameters. Another way of
3968: thinking about this is to realise that all Forth definitions are
3969: @i{active}; they execute as they are encountered by the text
3970: interpreter. The result of this is that the syntax of Forth is trivially
3971: simple.
3972:
3973:
3974:
3975: @comment ----------------------------------------------
3976: @node Your first definition, How does that work?, Stacks and Postfix notation, Introduction
3977: @section Your first Forth definition
3978: @cindex first definition
3979:
3980: Until now, the examples we've seen have been trivial; we've just been
3981: using Forth as a bigger-than-pocket calculator. Also, each calculation
3982: we've shown has been a ``one-off'' -- to repeat it we'd need to type it in
3983: again@footnote{That's not quite true. If you press the up-arrow key on
3984: your keyboard you should be able to scroll back to any earlier command,
3985: edit it and re-enter it.} In this section we'll see how to add new
3986: words to Forth's vocabulary.
3987:
3988: The easiest way to create a new word is to use a @dfn{colon
3989: definition}. We'll define a few and try them out before worrying too
3990: much about how they work. Try typing in these examples; be careful to
3991: copy the spaces accurately:
3992:
3993: @example
3994: : add-two 2 + . ;
3995: : greet ." Hello and welcome" ;
3996: : demo 5 add-two ;
3997: @end example
3998:
3999: @noindent
4000: Now try them out:
4001:
4002: @example
4003: @kbd{greet@key{RET}} Hello and welcome ok
4004: @kbd{greet greet@key{RET}} Hello and welcomeHello and welcome ok
4005: @kbd{4 add-two@key{RET}} 6 ok
4006: @kbd{demo@key{RET}} 7 ok
4007: @kbd{9 greet demo add-two@key{RET}} Hello and welcome7 11 ok
4008: @end example
4009:
4010: The first new thing that we've introduced here is the pair of words
4011: @code{:} and @code{;}. These are used to start and terminate a new
4012: definition, respectively. The first word after the @code{:} is the name
4013: for the new definition.
4014:
4015: As you can see from the examples, a definition is built up of words that
4016: have already been defined; Forth makes no distinction between
4017: definitions that existed when you started the system up, and those that
4018: you define yourself.
4019:
4020: The examples also introduce the words @code{.} (dot), @code{."}
4021: (dot-quote) and @code{dup} (dewp). Dot takes the value from the top of
4022: the stack and displays it. It's like @code{.s} except that it only
4023: displays the top item of the stack and it is destructive; after it has
4024: executed, the number is no longer on the stack. There is always one
4025: space printed after the number, and no spaces before it. Dot-quote
4026: defines a string (a sequence of characters) that will be printed when
4027: the word is executed. The string can contain any printable characters
4028: except @code{"}. A @code{"} has a special function; it is not a Forth
4029: word but it acts as a delimiter (the way that delimiters work is
4030: described in the next section). Finally, @code{dup} duplicates the value
4031: at the top of the stack. Try typing @code{5 dup .s} to see what it does.
4032:
4033: We already know that the text interpreter searches through the
4034: dictionary to locate names. If you've followed the examples earlier, you
4035: will already have a definition called @code{add-two}. Lets try modifying
4036: it by typing in a new definition:
4037:
4038: @example
4039: @kbd{: add-two dup . ." + 2 =" 2 + . ;@key{RET}} redefined add-two ok
4040: @end example
4041:
4042: Forth recognised that we were defining a word that already exists, and
4043: printed a message to warn us of that fact. Let's try out the new
4044: definition:
4045:
4046: @example
4047: @kbd{9 add-two@key{RET}} 9 + 2 =11 ok
4048: @end example
4049:
4050: @noindent
4051: All that we've actually done here, though, is to create a new
4052: definition, with a particular name. The fact that there was already a
4053: definition with the same name did not make any difference to the way
4054: that the new definition was created (except that Forth printed a warning
4055: message). The old definition of add-two still exists (try @code{demo}
4056: again to see that this is true). Any new definition will use the new
4057: definition of @code{add-two}, but old definitions continue to use the
4058: version that already existed at the time that they were @code{compiled}.
4059:
4060: Before you go on to the next section, try defining and redefining some
4061: words of your own.
4062:
4063: @comment ----------------------------------------------
4064: @node How does that work?, Forth is written in Forth, Your first definition, Introduction
4065: @section How does that work?
4066: @cindex parsing words
4067:
4068: @c That's pretty deep (IMO way too deep) for an introduction. - anton
4069:
4070: @c Is it a good idea to talk about the interpretation semantics of a
4071: @c number? We don't have an xt to go along with it. - anton
4072:
4073: @c Now that I have eliminated execution semantics, I wonder if it would not
4074: @c be better to keep them (or add run-time semantics), to make it easier to
4075: @c explain what compilation semantics usually does. - anton
4076:
4077: @c nac-> I removed the term ``default compilation sematics'' from the
4078: @c introductory chapter. Removing ``execution semantics'' was making
4079: @c everything simpler to explain, then I think the use of this term made
4080: @c everything more complex again. I replaced it with ``default
4081: @c semantics'' (which is used elsewhere in the manual) by which I mean
4082: @c ``a definition that has neither the immediate nor the compile-only
4083: @c flag set''.
4084:
4085: @c anton: I have eliminated default semantics (except in one place where it
4086: @c means "default interpretation and compilation semantics"), because it
4087: @c makes no sense in the presence of combined words. I reverted to
4088: @c "execution semantics" where necessary.
4089:
4090: @c nac-> I reworded big chunks of the ``how does that work''
4091: @c section (and, unusually for me, I think I even made it shorter!). See
4092: @c what you think -- I know I have not addressed your primary concern
4093: @c that it is too heavy-going for an introduction. From what I understood
4094: @c of your course notes it looks as though they might be a good framework.
4095: @c Things that I've tried to capture here are some things that came as a
4096: @c great revelation here when I first understood them. Also, I like the
4097: @c fact that a very simple code example shows up almost all of the issues
4098: @c that you need to understand to see how Forth works. That's unique and
4099: @c worthwhile to emphasise.
4100:
4101: @c anton: I think it's a good idea to present the details, especially those
4102: @c that you found to be a revelation, and probably the tutorial tries to be
4103: @c too superficial and does not get some of the things across that make
4104: @c Forth special. I do believe that most of the time these things should
4105: @c be discussed at the end of a section or in separate sections instead of
4106: @c in the middle of a section (e.g., the stuff you added in "User-defined
4107: @c defining words" leads in a completely different direction from the rest
4108: @c of the section).
4109:
4110: Now we're going to take another look at the definition of @code{add-two}
4111: from the previous section. From our knowledge of the way that the text
4112: interpreter works, we would have expected this result when we tried to
4113: define @code{add-two}:
4114:
4115: @example
4116: @kbd{: add-two 2 + . ;@key{RET}}
4117: ^^^^^^^
4118: Error: Undefined word
4119: @end example
4120:
4121: The reason that this didn't happen is bound up in the way that @code{:}
4122: works. The word @code{:} does two special things. The first special
4123: thing that it does prevents the text interpreter from ever seeing the
4124: characters @code{add-two}. The text interpreter uses a variable called
4125: @cindex modifying >IN
4126: @code{>IN} (pronounced ``to-in'') to keep track of where it is in the
4127: input line. When it encounters the word @code{:} it behaves in exactly
4128: the same way as it does for any other word; it looks it up in the name
4129: dictionary, finds its xt and executes it. When @code{:} executes, it
4130: looks at the input buffer, finds the word @code{add-two} and advances the
4131: value of @code{>IN} to point past it. It then does some other stuff
4132: associated with creating the new definition (including creating an entry
4133: for @code{add-two} in the name dictionary). When the execution of @code{:}
4134: completes, control returns to the text interpreter, which is oblivious
4135: to the fact that it has been tricked into ignoring part of the input
4136: line.
4137:
4138: @cindex parsing words
4139: Words like @code{:} -- words that advance the value of @code{>IN} and so
4140: prevent the text interpreter from acting on the whole of the input line
4141: -- are called @dfn{parsing words}.
4142:
4143: @cindex @code{state} - effect on the text interpreter
4144: @cindex text interpreter - effect of state
4145: The second special thing that @code{:} does is change the value of a
4146: variable called @code{state}, which affects the way that the text
4147: interpreter behaves. When Gforth starts up, @code{state} has the value
4148: 0, and the text interpreter is said to be @dfn{interpreting}. During a
4149: colon definition (started with @code{:}), @code{state} is set to -1 and
4150: the text interpreter is said to be @dfn{compiling}.
4151:
4152: In this example, the text interpreter is compiling when it processes the
4153: string ``@code{2 + . ;}''. It still breaks the string down into
4154: character sequences in the same way. However, instead of pushing the
4155: number @code{2} onto the stack, it lays down (@dfn{compiles}) some magic
4156: into the definition of @code{add-two} that will make the number @code{2} get
4157: pushed onto the stack when @code{add-two} is @dfn{executed}. Similarly,
4158: the behaviours of @code{+} and @code{.} are also compiled into the
4159: definition.
4160:
4161: One category of words don't get compiled. These so-called @dfn{immediate
4162: words} get executed (performed @i{now}) regardless of whether the text
4163: interpreter is interpreting or compiling. The word @code{;} is an
4164: immediate word. Rather than being compiled into the definition, it
4165: executes. Its effect is to terminate the current definition, which
4166: includes changing the value of @code{state} back to 0.
4167:
4168: When you execute @code{add-two}, it has a @dfn{run-time effect} that is
4169: exactly the same as if you had typed @code{2 + . @key{RET}} outside of a
4170: definition.
4171:
4172: In Forth, every word or number can be described in terms of two
4173: properties:
4174:
4175: @itemize @bullet
4176: @item
4177: @cindex interpretation semantics
4178: Its @dfn{interpretation semantics} describe how it will behave when the
4179: text interpreter encounters it in @dfn{interpret} state. The
4180: interpretation semantics of a word are represented by an @dfn{execution
4181: token}.
4182: @item
4183: @cindex compilation semantics
4184: Its @dfn{compilation semantics} describe how it will behave when the
4185: text interpreter encounters it in @dfn{compile} state. The compilation
4186: semantics of a word are represented in an implementation-dependent way;
4187: Gforth uses a @dfn{compilation token}.
4188: @end itemize
4189:
4190: @noindent
4191: Numbers are always treated in a fixed way:
4192:
4193: @itemize @bullet
4194: @item
4195: When the number is @dfn{interpreted}, its behaviour is to push the
4196: number onto the stack.
4197: @item
4198: When the number is @dfn{compiled}, a piece of code is appended to the
4199: current definition that pushes the number when it runs. (In other words,
4200: the compilation semantics of a number are to postpone its interpretation
4201: semantics until the run-time of the definition that it is being compiled
4202: into.)
4203: @end itemize
4204:
4205: Words don't behave in such a regular way, but most have @i{default
4206: semantics} which means that they behave like this:
4207:
4208: @itemize @bullet
4209: @item
4210: The @dfn{interpretation semantics} of the word are to do something useful.
4211: @item
4212: The @dfn{compilation semantics} of the word are to append its
4213: @dfn{interpretation semantics} to the current definition (so that its
4214: run-time behaviour is to do something useful).
4215: @end itemize
4216:
4217: @cindex immediate words
4218: The actual behaviour of any particular word can be controlled by using
4219: the words @code{immediate} and @code{compile-only} when the word is
4220: defined. These words set flags in the name dictionary entry of the most
4221: recently defined word, and these flags are retrieved by the text
4222: interpreter when it finds the word in the name dictionary.
4223:
4224: A word that is marked as @dfn{immediate} has compilation semantics that
4225: are identical to its interpretation semantics. In other words, it
4226: behaves like this:
4227:
4228: @itemize @bullet
4229: @item
4230: The @dfn{interpretation semantics} of the word are to do something useful.
4231: @item
4232: The @dfn{compilation semantics} of the word are to do something useful
4233: (and actually the same thing); i.e., it is executed during compilation.
4234: @end itemize
4235:
4236: Marking a word as @dfn{compile-only} prohibits the text interpreter from
4237: performing the interpretation semantics of the word directly; an attempt
4238: to do so will generate an error. It is never necessary to use
4239: @code{compile-only} (and it is not even part of ANS Forth, though it is
4240: provided by many implementations) but it is good etiquette to apply it
4241: to a word that will not behave correctly (and might have unexpected
4242: side-effects) in interpret state. For example, it is only legal to use
4243: the conditional word @code{IF} within a definition. If you forget this
4244: and try to use it elsewhere, the fact that (in Gforth) it is marked as
4245: @code{compile-only} allows the text interpreter to generate a helpful
4246: error message rather than subjecting you to the consequences of your
4247: folly.
4248:
4249: This example shows the difference between an immediate and a
4250: non-immediate word:
4251:
4252: @example
4253: : show-state state @@ . ;
4254: : show-state-now show-state ; immediate
4255: : word1 show-state ;
4256: : word2 show-state-now ;
4257: @end example
4258:
4259: The word @code{immediate} after the definition of @code{show-state-now}
4260: makes that word an immediate word. These definitions introduce a new
4261: word: @code{@@} (pronounced ``fetch''). This word fetches the value of a
4262: variable, and leaves it on the stack. Therefore, the behaviour of
4263: @code{show-state} is to print a number that represents the current value
4264: of @code{state}.
4265:
4266: When you execute @code{word1}, it prints the number 0, indicating that
4267: the system is interpreting. When the text interpreter compiled the
4268: definition of @code{word1}, it encountered @code{show-state} whose
4269: compilation semantics are to append its interpretation semantics to the
4270: current definition. When you execute @code{word1}, it performs the
4271: interpretation semantics of @code{show-state}. At the time that @code{word1}
4272: (and therefore @code{show-state}) are executed, the system is
4273: interpreting.
4274:
4275: When you pressed @key{RET} after entering the definition of @code{word2},
4276: you should have seen the number -1 printed, followed by ``@code{
4277: ok}''. When the text interpreter compiled the definition of
4278: @code{word2}, it encountered @code{show-state-now}, an immediate word,
4279: whose compilation semantics are therefore to perform its interpretation
4280: semantics. It is executed straight away (even before the text
4281: interpreter has moved on to process another group of characters; the
4282: @code{;} in this example). The effect of executing it are to display the
4283: value of @code{state} @i{at the time that the definition of}
4284: @code{word2} @i{is being defined}. Printing -1 demonstrates that the
4285: system is compiling at this time. If you execute @code{word2} it does
4286: nothing at all.
4287:
4288: @cindex @code{."}, how it works
4289: Before leaving the subject of immediate words, consider the behaviour of
4290: @code{."} in the definition of @code{greet}, in the previous
4291: section. This word is both a parsing word and an immediate word. Notice
4292: that there is a space between @code{."} and the start of the text
4293: @code{Hello and welcome}, but that there is no space between the last
4294: letter of @code{welcome} and the @code{"} character. The reason for this
4295: is that @code{."} is a Forth word; it must have a space after it so that
4296: the text interpreter can identify it. The @code{"} is not a Forth word;
4297: it is a @dfn{delimiter}. The examples earlier show that, when the string
4298: is displayed, there is neither a space before the @code{H} nor after the
4299: @code{e}. Since @code{."} is an immediate word, it executes at the time
4300: that @code{greet} is defined. When it executes, its behaviour is to
4301: search forward in the input line looking for the delimiter. When it
4302: finds the delimiter, it updates @code{>IN} to point past the
4303: delimiter. It also compiles some magic code into the definition of
4304: @code{greet}; the xt of a run-time routine that prints a text string. It
4305: compiles the string @code{Hello and welcome} into memory so that it is
4306: available to be printed later. When the text interpreter gains control,
4307: the next word it finds in the input stream is @code{;} and so it
4308: terminates the definition of @code{greet}.
4309:
4310:
4311: @comment ----------------------------------------------
4312: @node Forth is written in Forth, Review - elements of a Forth system, How does that work?, Introduction
4313: @section Forth is written in Forth
4314: @cindex structure of Forth programs
4315:
4316: When you start up a Forth compiler, a large number of definitions
4317: already exist. In Forth, you develop a new application using bottom-up
4318: programming techniques to create new definitions that are defined in
4319: terms of existing definitions. As you create each definition you can
4320: test and debug it interactively.
4321:
4322: If you have tried out the examples in this section, you will probably
4323: have typed them in by hand; when you leave Gforth, your definitions will
4324: be lost. You can avoid this by using a text editor to enter Forth source
4325: code into a file, and then loading code from the file using
4326: @code{include} (@pxref{Forth source files}). A Forth source file is
4327: processed by the text interpreter, just as though you had typed it in by
4328: hand@footnote{Actually, there are some subtle differences -- see
4329: @ref{The Text Interpreter}.}.
4330:
4331: Gforth also supports the traditional Forth alternative to using text
4332: files for program entry (@pxref{Blocks}).
4333:
4334: In common with many, if not most, Forth compilers, most of Gforth is
4335: actually written in Forth. All of the @file{.fs} files in the
4336: installation directory@footnote{For example,
4337: @file{/usr/local/share/gforth...}} are Forth source files, which you can
4338: study to see examples of Forth programming.
4339:
4340: Gforth maintains a history file that records every line that you type to
4341: the text interpreter. This file is preserved between sessions, and is
4342: used to provide a command-line recall facility. If you enter long
4343: definitions by hand, you can use a text editor to paste them out of the
4344: history file into a Forth source file for reuse at a later time
4345: (for more information @pxref{Command-line editing}).
4346:
4347:
4348: @comment ----------------------------------------------
4349: @node Review - elements of a Forth system, Where to go next, Forth is written in Forth, Introduction
4350: @section Review - elements of a Forth system
4351: @cindex elements of a Forth system
4352:
4353: To summarise this chapter:
4354:
4355: @itemize @bullet
4356: @item
4357: Forth programs use @dfn{factoring} to break a problem down into small
4358: fragments called @dfn{words} or @dfn{definitions}.
4359: @item
4360: Forth program development is an interactive process.
4361: @item
4362: The main command loop that accepts input, and controls both
4363: interpretation and compilation, is called the @dfn{text interpreter}
4364: (also known as the @dfn{outer interpreter}).
4365: @item
4366: Forth has a very simple syntax, consisting of words and numbers
4367: separated by spaces or carriage-return characters. Any additional syntax
4368: is imposed by @dfn{parsing words}.
4369: @item
4370: Forth uses a stack to pass parameters between words. As a result, it
4371: uses postfix notation.
4372: @item
4373: To use a word that has previously been defined, the text interpreter
4374: searches for the word in the @dfn{name dictionary}.
4375: @item
4376: Words have @dfn{interpretation semantics} and @dfn{compilation semantics}.
4377: @item
4378: The text interpreter uses the value of @code{state} to select between
4379: the use of the @dfn{interpretation semantics} and the @dfn{compilation
4380: semantics} of a word that it encounters.
4381: @item
4382: The relationship between the @dfn{interpretation semantics} and
4383: @dfn{compilation semantics} for a word
4384: depend upon the way in which the word was defined (for example, whether
4385: it is an @dfn{immediate} word).
4386: @item
4387: Forth definitions can be implemented in Forth (called @dfn{high-level
4388: definitions}) or in some other way (usually a lower-level language and
4389: as a result often called @dfn{low-level definitions}, @dfn{code
4390: definitions} or @dfn{primitives}).
4391: @item
4392: Many Forth systems are implemented mainly in Forth.
4393: @end itemize
4394:
4395:
4396: @comment ----------------------------------------------
4397: @node Where to go next, Exercises, Review - elements of a Forth system, Introduction
4398: @section Where To Go Next
4399: @cindex where to go next
4400:
4401: Amazing as it may seem, if you have read (and understood) this far, you
4402: know almost all the fundamentals about the inner workings of a Forth
4403: system. You certainly know enough to be able to read and understand the
4404: rest of this manual and the ANS Forth document, to learn more about the
4405: facilities that Forth in general and Gforth in particular provide. Even
4406: scarier, you know almost enough to implement your own Forth system.
4407: However, that's not a good idea just yet... better to try writing some
4408: programs in Gforth.
4409:
4410: Forth has such a rich vocabulary that it can be hard to know where to
4411: start in learning it. This section suggests a few sets of words that are
4412: enough to write small but useful programs. Use the word index in this
4413: document to learn more about each word, then try it out and try to write
4414: small definitions using it. Start by experimenting with these words:
4415:
4416: @itemize @bullet
4417: @item
4418: Arithmetic: @code{+ - * / /MOD */ ABS INVERT}
4419: @item
4420: Comparison: @code{MIN MAX =}
4421: @item
4422: Logic: @code{AND OR XOR NOT}
4423: @item
4424: Stack manipulation: @code{DUP DROP SWAP OVER}
4425: @item
4426: Loops and decisions: @code{IF ELSE ENDIF ?DO I LOOP}
4427: @item
4428: Input/Output: @code{. ." EMIT CR KEY}
4429: @item
4430: Defining words: @code{: ; CREATE}
4431: @item
4432: Memory allocation words: @code{ALLOT ,}
4433: @item
4434: Tools: @code{SEE WORDS .S MARKER}
4435: @end itemize
4436:
4437: When you have mastered those, go on to:
4438:
4439: @itemize @bullet
4440: @item
4441: More defining words: @code{VARIABLE CONSTANT VALUE TO CREATE DOES>}
4442: @item
4443: Memory access: @code{@@ !}
4444: @end itemize
4445:
4446: When you have mastered these, there's nothing for it but to read through
4447: the whole of this manual and find out what you've missed.
4448:
4449: @comment ----------------------------------------------
4450: @node Exercises, , Where to go next, Introduction
4451: @section Exercises
4452: @cindex exercises
4453:
4454: TODO: provide a set of programming excercises linked into the stuff done
4455: already and into other sections of the manual. Provide solutions to all
4456: the exercises in a .fs file in the distribution.
4457:
4458: @c Get some inspiration from Starting Forth and Kelly&Spies.
4459:
4460: @c excercises:
4461: @c 1. take inches and convert to feet and inches.
4462: @c 2. take temperature and convert from fahrenheight to celcius;
4463: @c may need to care about symmetric vs floored??
4464: @c 3. take input line and do character substitution
4465: @c to encipher or decipher
4466: @c 4. as above but work on a file for in and out
4467: @c 5. take input line and convert to pig-latin
4468: @c
4469: @c thing of sets of things to exercise then come up with
4470: @c problems that need those things.
4471:
4472:
4473: @c ******************************************************************
4474: @node Words, Error messages, Introduction, Top
4475: @chapter Forth Words
4476: @cindex words
4477:
4478: @menu
4479: * Notation::
4480: * Case insensitivity::
4481: * Comments::
4482: * Boolean Flags::
4483: * Arithmetic::
4484: * Stack Manipulation::
4485: * Memory::
4486: * Control Structures::
4487: * Defining Words::
4488: * Interpretation and Compilation Semantics::
4489: * Tokens for Words::
4490: * Compiling words::
4491: * The Text Interpreter::
4492: * Word Lists::
4493: * Environmental Queries::
4494: * Files::
4495: * Blocks::
4496: * Other I/O::
4497: * Locals::
4498: * Structures::
4499: * Object-oriented Forth::
4500: * Programming Tools::
4501: * Assembler and Code Words::
4502: * Threading Words::
4503: * Passing Commands to the OS::
4504: * Keeping track of Time::
4505: * Miscellaneous Words::
4506: @end menu
4507:
4508: @node Notation, Case insensitivity, Words, Words
4509: @section Notation
4510: @cindex notation of glossary entries
4511: @cindex format of glossary entries
4512: @cindex glossary notation format
4513: @cindex word glossary entry format
4514:
4515: The Forth words are described in this section in the glossary notation
4516: that has become a de-facto standard for Forth texts:
4517:
4518: @format
4519: @i{word} @i{Stack effect} @i{wordset} @i{pronunciation}
4520: @end format
4521: @i{Description}
4522:
4523: @table @var
4524: @item word
4525: The name of the word.
4526:
4527: @item Stack effect
4528: @cindex stack effect
4529: The stack effect is written in the notation @code{@i{before} --
4530: @i{after}}, where @i{before} and @i{after} describe the top of
4531: stack entries before and after the execution of the word. The rest of
4532: the stack is not touched by the word. The top of stack is rightmost,
4533: i.e., a stack sequence is written as it is typed in. Note that Gforth
4534: uses a separate floating point stack, but a unified stack
4535: notation. Also, return stack effects are not shown in @i{stack
4536: effect}, but in @i{Description}. The name of a stack item describes
4537: the type and/or the function of the item. See below for a discussion of
4538: the types.
4539:
4540: All words have two stack effects: A compile-time stack effect and a
4541: run-time stack effect. The compile-time stack-effect of most words is
4542: @i{ -- }. If the compile-time stack-effect of a word deviates from
4543: this standard behaviour, or the word does other unusual things at
4544: compile time, both stack effects are shown; otherwise only the run-time
4545: stack effect is shown.
4546:
4547: @cindex pronounciation of words
4548: @item pronunciation
4549: How the word is pronounced.
4550:
4551: @cindex wordset
4552: @cindex environment wordset
4553: @item wordset
4554: The ANS Forth standard is divided into several word sets. A standard
4555: system need not support all of them. Therefore, in theory, the fewer
4556: word sets your program uses the more portable it will be. However, we
4557: suspect that most ANS Forth systems on personal machines will feature
4558: all word sets. Words that are not defined in ANS Forth have
4559: @code{gforth} or @code{gforth-internal} as word set. @code{gforth}
4560: describes words that will work in future releases of Gforth;
4561: @code{gforth-internal} words are more volatile. Environmental query
4562: strings are also displayed like words; you can recognize them by the
4563: @code{environment} in the word set field.
4564:
4565: @item Description
4566: A description of the behaviour of the word.
4567: @end table
4568:
4569: @cindex types of stack items
4570: @cindex stack item types
4571: The type of a stack item is specified by the character(s) the name
4572: starts with:
4573:
4574: @table @code
4575: @item f
4576: @cindex @code{f}, stack item type
4577: Boolean flags, i.e. @code{false} or @code{true}.
4578: @item c
4579: @cindex @code{c}, stack item type
4580: Char
4581: @item w
4582: @cindex @code{w}, stack item type
4583: Cell, can contain an integer or an address
4584: @item n
4585: @cindex @code{n}, stack item type
4586: signed integer
4587: @item u
4588: @cindex @code{u}, stack item type
4589: unsigned integer
4590: @item d
4591: @cindex @code{d}, stack item type
4592: double sized signed integer
4593: @item ud
4594: @cindex @code{ud}, stack item type
4595: double sized unsigned integer
4596: @item r
4597: @cindex @code{r}, stack item type
4598: Float (on the FP stack)
4599: @item a-
4600: @cindex @code{a_}, stack item type
4601: Cell-aligned address
4602: @item c-
4603: @cindex @code{c_}, stack item type
4604: Char-aligned address (note that a Char may have two bytes in Windows NT)
4605: @item f-
4606: @cindex @code{f_}, stack item type
4607: Float-aligned address
4608: @item df-
4609: @cindex @code{df_}, stack item type
4610: Address aligned for IEEE double precision float
4611: @item sf-
4612: @cindex @code{sf_}, stack item type
4613: Address aligned for IEEE single precision float
4614: @item xt
4615: @cindex @code{xt}, stack item type
4616: Execution token, same size as Cell
4617: @item wid
4618: @cindex @code{wid}, stack item type
4619: Word list ID, same size as Cell
4620: @item ior, wior
4621: @cindex ior type description
4622: @cindex wior type description
4623: I/O result code, cell-sized. In Gforth, you can @code{throw} iors.
4624: @item f83name
4625: @cindex @code{f83name}, stack item type
4626: Pointer to a name structure
4627: @item "
4628: @cindex @code{"}, stack item type
4629: string in the input stream (not on the stack). The terminating character
4630: is a blank by default. If it is not a blank, it is shown in @code{<>}
4631: quotes.
4632: @end table
4633:
4634: @comment ----------------------------------------------
4635: @node Case insensitivity, Comments, Notation, Words
4636: @section Case insensitivity
4637: @cindex case sensitivity
4638: @cindex upper and lower case
4639:
4640: Gforth is case-insensitive; you can enter definitions and invoke
4641: Standard words using upper, lower or mixed case (however,
4642: @pxref{core-idef, Implementation-defined options, Implementation-defined
4643: options}).
4644:
4645: ANS Forth only @i{requires} implementations to recognise Standard words
4646: when they are typed entirely in upper case. Therefore, a Standard
4647: program must use upper case for all Standard words. You can use whatever
4648: case you like for words that you define, but in a Standard program you
4649: have to use the words in the same case that you defined them.
4650:
4651: Gforth supports case sensitivity through @code{table}s (case-sensitive
4652: wordlists, @pxref{Word Lists}).
4653:
4654: Two people have asked how to convert Gforth to be case-sensitive; while
4655: we think this is a bad idea, you can change all wordlists into tables
4656: like this:
4657:
4658: @example
4659: ' table-find forth-wordlist wordlist-map @ !
4660: @end example
4661:
4662: Note that you now have to type the predefined words in the same case
4663: that we defined them, which are varying. You may want to convert them
4664: to your favourite case before doing this operation (I won't explain how,
4665: because if you are even contemplating doing this, you'd better have
4666: enough knowledge of Forth systems to know this already).
4667:
4668: @node Comments, Boolean Flags, Case insensitivity, Words
4669: @section Comments
4670: @cindex comments
4671:
4672: Forth supports two styles of comment; the traditional @i{in-line} comment,
4673: @code{(} and its modern cousin, the @i{comment to end of line}; @code{\}.
4674:
4675:
4676: doc-(
4677: doc-\
4678: doc-\G
4679:
4680:
4681: @node Boolean Flags, Arithmetic, Comments, Words
4682: @section Boolean Flags
4683: @cindex Boolean flags
4684:
4685: A Boolean flag is cell-sized. A cell with all bits clear represents the
4686: flag @code{false} and a flag with all bits set represents the flag
4687: @code{true}. Words that check a flag (for example, @code{IF}) will treat
4688: a cell that has @i{any} bit set as @code{true}.
4689: @c on and off to Memory?
4690: @c true and false to "Bitwise operations" or "Numeric comparison"?
4691:
4692: doc-true
4693: doc-false
4694: doc-on
4695: doc-off
4696:
4697:
4698: @node Arithmetic, Stack Manipulation, Boolean Flags, Words
4699: @section Arithmetic
4700: @cindex arithmetic words
4701:
4702: @cindex division with potentially negative operands
4703: Forth arithmetic is not checked, i.e., you will not hear about integer
4704: overflow on addition or multiplication, you may hear about division by
4705: zero if you are lucky. The operator is written after the operands, but
4706: the operands are still in the original order. I.e., the infix @code{2-1}
4707: corresponds to @code{2 1 -}. Forth offers a variety of division
4708: operators. If you perform division with potentially negative operands,
4709: you do not want to use @code{/} or @code{/mod} with its undefined
4710: behaviour, but rather @code{fm/mod} or @code{sm/mod} (probably the
4711: former, @pxref{Mixed precision}).
4712: @comment TODO discuss the different division forms and the std approach
4713:
4714: @menu
4715: * Single precision::
4716: * Double precision:: Double-cell integer arithmetic
4717: * Bitwise operations::
4718: * Numeric comparison::
4719: * Mixed precision:: Operations with single and double-cell integers
4720: * Floating Point::
4721: @end menu
4722:
4723: @node Single precision, Double precision, Arithmetic, Arithmetic
4724: @subsection Single precision
4725: @cindex single precision arithmetic words
4726:
4727: @c !! cell undefined
4728:
4729: By default, numbers in Forth are single-precision integers that are one
4730: cell in size. They can be signed or unsigned, depending upon how you
4731: treat them. For the rules used by the text interpreter for recognising
4732: single-precision integers see @ref{Number Conversion}.
4733:
4734: These words are all defined for signed operands, but some of them also
4735: work for unsigned numbers: @code{+}, @code{1+}, @code{-}, @code{1-},
4736: @code{*}.
4737:
4738: doc-+
4739: doc-1+
4740: doc--
4741: doc-1-
4742: doc-*
4743: doc-/
4744: doc-mod
4745: doc-/mod
4746: doc-negate
4747: doc-abs
4748: doc-min
4749: doc-max
4750: doc-floored
4751:
4752:
4753: @node Double precision, Bitwise operations, Single precision, Arithmetic
4754: @subsection Double precision
4755: @cindex double precision arithmetic words
4756:
4757: For the rules used by the text interpreter for
4758: recognising double-precision integers, see @ref{Number Conversion}.
4759:
4760: A double precision number is represented by a cell pair, with the most
4761: significant cell at the TOS. It is trivial to convert an unsigned single
4762: to a double: simply push a @code{0} onto the TOS. Since numbers are
4763: represented by Gforth using 2's complement arithmetic, converting a
4764: signed single to a (signed) double requires sign-extension across the
4765: most significant cell. This can be achieved using @code{s>d}. The moral
4766: of the story is that you cannot convert a number without knowing whether
4767: it represents an unsigned or a signed number.
4768:
4769: These words are all defined for signed operands, but some of them also
4770: work for unsigned numbers: @code{d+}, @code{d-}.
4771:
4772: doc-s>d
4773: doc-d>s
4774: doc-d+
4775: doc-d-
4776: doc-dnegate
4777: doc-dabs
4778: doc-dmin
4779: doc-dmax
4780:
4781:
4782: @node Bitwise operations, Numeric comparison, Double precision, Arithmetic
4783: @subsection Bitwise operations
4784: @cindex bitwise operation words
4785:
4786:
4787: doc-and
4788: doc-or
4789: doc-xor
4790: doc-invert
4791: doc-lshift
4792: doc-rshift
4793: doc-2*
4794: doc-d2*
4795: doc-2/
4796: doc-d2/
4797:
4798:
4799: @node Numeric comparison, Mixed precision, Bitwise operations, Arithmetic
4800: @subsection Numeric comparison
4801: @cindex numeric comparison words
4802:
4803: Note that the words that compare for equality (@code{= <> 0= 0<> d= d<>
4804: d0= d0<>}) work for for both signed and unsigned numbers.
4805:
4806: doc-<
4807: doc-<=
4808: doc-<>
4809: doc-=
4810: doc->
4811: doc->=
4812:
4813: doc-0<
4814: doc-0<=
4815: doc-0<>
4816: doc-0=
4817: doc-0>
4818: doc-0>=
4819:
4820: doc-u<
4821: doc-u<=
4822: @c u<> and u= exist but are the same as <> and =
4823: @c doc-u<>
4824: @c doc-u=
4825: doc-u>
4826: doc-u>=
4827:
4828: doc-within
4829:
4830: doc-d<
4831: doc-d<=
4832: doc-d<>
4833: doc-d=
4834: doc-d>
4835: doc-d>=
4836:
4837: doc-d0<
4838: doc-d0<=
4839: doc-d0<>
4840: doc-d0=
4841: doc-d0>
4842: doc-d0>=
4843:
4844: doc-du<
4845: doc-du<=
4846: @c du<> and du= exist but are the same as d<> and d=
4847: @c doc-du<>
4848: @c doc-du=
4849: doc-du>
4850: doc-du>=
4851:
4852:
4853: @node Mixed precision, Floating Point, Numeric comparison, Arithmetic
4854: @subsection Mixed precision
4855: @cindex mixed precision arithmetic words
4856:
4857:
4858: doc-m+
4859: doc-*/
4860: doc-*/mod
4861: doc-m*
4862: doc-um*
4863: doc-m*/
4864: doc-um/mod
4865: doc-fm/mod
4866: doc-sm/rem
4867:
4868:
4869: @node Floating Point, , Mixed precision, Arithmetic
4870: @subsection Floating Point
4871: @cindex floating point arithmetic words
4872:
4873: For the rules used by the text interpreter for
4874: recognising floating-point numbers see @ref{Number Conversion}.
4875:
4876: Gforth has a separate floating point stack, but the documentation uses
4877: the unified notation.@footnote{It's easy to generate the separate
4878: notation from that by just separating the floating-point numbers out:
4879: e.g. @code{( n r1 u r2 -- r3 )} becomes @code{( n u -- ) ( F: r1 r2 --
4880: r3 )}.}
4881:
4882: @cindex floating-point arithmetic, pitfalls
4883: Floating point numbers have a number of unpleasant surprises for the
4884: unwary (e.g., floating point addition is not associative) and even a few
4885: for the wary. You should not use them unless you know what you are doing
4886: or you don't care that the results you get are totally bogus. If you
4887: want to learn about the problems of floating point numbers (and how to
4888: avoid them), you might start with @cite{David Goldberg,
4889: @uref{http://www.validgh.com/goldberg/paper.ps,What Every Computer
4890: Scientist Should Know About Floating-Point Arithmetic}, ACM Computing
4891: Surveys 23(1):5@minus{}48, March 1991}.
4892:
4893:
4894: doc-d>f
4895: doc-f>d
4896: doc-f+
4897: doc-f-
4898: doc-f*
4899: doc-f/
4900: doc-fnegate
4901: doc-fabs
4902: doc-fmax
4903: doc-fmin
4904: doc-floor
4905: doc-fround
4906: doc-f**
4907: doc-fsqrt
4908: doc-fexp
4909: doc-fexpm1
4910: doc-fln
4911: doc-flnp1
4912: doc-flog
4913: doc-falog
4914: doc-f2*
4915: doc-f2/
4916: doc-1/f
4917: doc-precision
4918: doc-set-precision
4919:
4920: @cindex angles in trigonometric operations
4921: @cindex trigonometric operations
4922: Angles in floating point operations are given in radians (a full circle
4923: has 2 pi radians).
4924:
4925: doc-fsin
4926: doc-fcos
4927: doc-fsincos
4928: doc-ftan
4929: doc-fasin
4930: doc-facos
4931: doc-fatan
4932: doc-fatan2
4933: doc-fsinh
4934: doc-fcosh
4935: doc-ftanh
4936: doc-fasinh
4937: doc-facosh
4938: doc-fatanh
4939: doc-pi
4940:
4941: @cindex equality of floats
4942: @cindex floating-point comparisons
4943: One particular problem with floating-point arithmetic is that comparison
4944: for equality often fails when you would expect it to succeed. For this
4945: reason approximate equality is often preferred (but you still have to
4946: know what you are doing). Also note that IEEE NaNs may compare
4947: differently from what you might expect. The comparison words are:
4948:
4949: doc-f~rel
4950: doc-f~abs
4951: doc-f~
4952: doc-f=
4953: doc-f<>
4954:
4955: doc-f<
4956: doc-f<=
4957: doc-f>
4958: doc-f>=
4959:
4960: doc-f0<
4961: doc-f0<=
4962: doc-f0<>
4963: doc-f0=
4964: doc-f0>
4965: doc-f0>=
4966:
4967:
4968: @node Stack Manipulation, Memory, Arithmetic, Words
4969: @section Stack Manipulation
4970: @cindex stack manipulation words
4971:
4972: @cindex floating-point stack in the standard
4973: Gforth maintains a number of separate stacks:
4974:
4975: @cindex data stack
4976: @cindex parameter stack
4977: @itemize @bullet
4978: @item
4979: A data stack (also known as the @dfn{parameter stack}) -- for
4980: characters, cells, addresses, and double cells.
4981:
4982: @cindex floating-point stack
4983: @item
4984: A floating point stack -- for holding floating point (FP) numbers.
4985:
4986: @cindex return stack
4987: @item
4988: A return stack -- for holding the return addresses of colon
4989: definitions and other (non-FP) data.
4990:
4991: @cindex locals stack
4992: @item
4993: A locals stack -- for holding local variables.
4994: @end itemize
4995:
4996: @menu
4997: * Data stack::
4998: * Floating point stack::
4999: * Return stack::
5000: * Locals stack::
5001: * Stack pointer manipulation::
5002: @end menu
5003:
5004: @node Data stack, Floating point stack, Stack Manipulation, Stack Manipulation
5005: @subsection Data stack
5006: @cindex data stack manipulation words
5007: @cindex stack manipulations words, data stack
5008:
5009:
5010: doc-drop
5011: doc-nip
5012: doc-dup
5013: doc-over
5014: doc-tuck
5015: doc-swap
5016: doc-pick
5017: doc-rot
5018: doc--rot
5019: doc-?dup
5020: doc-roll
5021: doc-2drop
5022: doc-2nip
5023: doc-2dup
5024: doc-2over
5025: doc-2tuck
5026: doc-2swap
5027: doc-2rot
5028:
5029:
5030: @node Floating point stack, Return stack, Data stack, Stack Manipulation
5031: @subsection Floating point stack
5032: @cindex floating-point stack manipulation words
5033: @cindex stack manipulation words, floating-point stack
5034:
5035: Whilst every sane Forth has a separate floating-point stack, it is not
5036: strictly required; an ANS Forth system could theoretically keep
5037: floating-point numbers on the data stack. As an additional difficulty,
5038: you don't know how many cells a floating-point number takes. It is
5039: reportedly possible to write words in a way that they work also for a
5040: unified stack model, but we do not recommend trying it. Instead, just
5041: say that your program has an environmental dependency on a separate
5042: floating-point stack.
5043:
5044: doc-floating-stack
5045:
5046: doc-fdrop
5047: doc-fnip
5048: doc-fdup
5049: doc-fover
5050: doc-ftuck
5051: doc-fswap
5052: doc-fpick
5053: doc-frot
5054:
5055:
5056: @node Return stack, Locals stack, Floating point stack, Stack Manipulation
5057: @subsection Return stack
5058: @cindex return stack manipulation words
5059: @cindex stack manipulation words, return stack
5060:
5061: @cindex return stack and locals
5062: @cindex locals and return stack
5063: A Forth system is allowed to keep local variables on the
5064: return stack. This is reasonable, as local variables usually eliminate
5065: the need to use the return stack explicitly. So, if you want to produce
5066: a standard compliant program and you are using local variables in a
5067: word, forget about return stack manipulations in that word (refer to the
5068: standard document for the exact rules).
5069:
5070: doc->r
5071: doc-r>
5072: doc-r@
5073: doc-rdrop
5074: doc-2>r
5075: doc-2r>
5076: doc-2r@
5077: doc-2rdrop
5078:
5079:
5080: @node Locals stack, Stack pointer manipulation, Return stack, Stack Manipulation
5081: @subsection Locals stack
5082:
5083: Gforth uses an extra locals stack. It is described, along with the
5084: reasons for its existence, in @ref{Locals implementation}.
5085:
5086: @node Stack pointer manipulation, , Locals stack, Stack Manipulation
5087: @subsection Stack pointer manipulation
5088: @cindex stack pointer manipulation words
5089:
5090: @c removed s0 r0 l0 -- they are obsolete aliases for sp0 rp0 lp0
5091: doc-sp0
5092: doc-sp@
5093: doc-sp!
5094: doc-fp0
5095: doc-fp@
5096: doc-fp!
5097: doc-rp0
5098: doc-rp@
5099: doc-rp!
5100: doc-lp0
5101: doc-lp@
5102: doc-lp!
5103:
5104:
5105: @node Memory, Control Structures, Stack Manipulation, Words
5106: @section Memory
5107: @cindex memory words
5108:
5109: @menu
5110: * Memory model::
5111: * Dictionary allocation::
5112: * Heap Allocation::
5113: * Memory Access::
5114: * Address arithmetic::
5115: * Memory Blocks::
5116: @end menu
5117:
5118: In addition to the standard Forth memory allocation words, there is also
5119: a @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
5120: garbage collector}.
5121:
5122: @node Memory model, Dictionary allocation, Memory, Memory
5123: @subsection ANS Forth and Gforth memory models
5124:
5125: @c The ANS Forth description is a mess (e.g., is the heap part of
5126: @c the dictionary?), so let's not stick to closely with it.
5127:
5128: ANS Forth considers a Forth system as consisting of several address
5129: spaces, of which only @dfn{data space} is managed and accessible with
5130: the memory words. Memory not necessarily in data space includes the
5131: stacks, the code (called code space) and the headers (called name
5132: space). In Gforth everything is in data space, but the code for the
5133: primitives is usually read-only.
5134:
5135: Data space is divided into a number of areas: The (data space portion of
5136: the) dictionary@footnote{Sometimes, the term @dfn{dictionary} is used to
5137: refer to the search data structure embodied in word lists and headers,
5138: because it is used for looking up names, just as you would in a
5139: conventional dictionary.}, the heap, and a number of system-allocated
5140: buffers.
5141:
5142: @cindex address arithmetic restrictions, ANS vs. Gforth
5143: @cindex contiguous regions, ANS vs. Gforth
5144: In ANS Forth data space is also divided into contiguous regions. You
5145: can only use address arithmetic within a contiguous region, not between
5146: them. Usually each allocation gives you one contiguous region, but the
5147: dictionary allocation words have additional rules (@pxref{Dictionary
5148: allocation}).
5149:
5150: Gforth provides one big address space, and address arithmetic can be
5151: performed between any addresses. However, in the dictionary headers or
5152: code are interleaved with data, so almost the only contiguous data space
5153: regions there are those described by ANS Forth as contiguous; but you
5154: can be sure that the dictionary is allocated towards increasing
5155: addresses even between contiguous regions. The memory order of
5156: allocations in the heap is platform-dependent (and possibly different
5157: from one run to the next).
5158:
5159:
5160: @node Dictionary allocation, Heap Allocation, Memory model, Memory
5161: @subsection Dictionary allocation
5162: @cindex reserving data space
5163: @cindex data space - reserving some
5164:
5165: Dictionary allocation is a stack-oriented allocation scheme, i.e., if
5166: you want to deallocate X, you also deallocate everything
5167: allocated after X.
5168:
5169: @cindex contiguous regions in dictionary allocation
5170: The allocations using the words below are contiguous and grow the region
5171: towards increasing addresses. Other words that allocate dictionary
5172: memory of any kind (i.e., defining words including @code{:noname}) end
5173: the contiguous region and start a new one.
5174:
5175: In ANS Forth only @code{create}d words are guaranteed to produce an
5176: address that is the start of the following contiguous region. In
5177: particular, the cell allocated by @code{variable} is not guaranteed to
5178: be contiguous with following @code{allot}ed memory.
5179:
5180: You can deallocate memory by using @code{allot} with a negative argument
5181: (with some restrictions, see @code{allot}). For larger deallocations use
5182: @code{marker}.
5183:
5184:
5185: doc-here
5186: doc-unused
5187: doc-allot
5188: doc-c,
5189: doc-f,
5190: doc-,
5191: doc-2,
5192:
5193: Memory accesses have to be aligned (@pxref{Address arithmetic}). So of
5194: course you should allocate memory in an aligned way, too. I.e., before
5195: allocating allocating a cell, @code{here} must be cell-aligned, etc.
5196: The words below align @code{here} if it is not already. Basically it is
5197: only already aligned for a type, if the last allocation was a multiple
5198: of the size of this type and if @code{here} was aligned for this type
5199: before.
5200:
5201: After freshly @code{create}ing a word, @code{here} is @code{align}ed in
5202: ANS Forth (@code{maxalign}ed in Gforth).
5203:
5204: doc-align
5205: doc-falign
5206: doc-sfalign
5207: doc-dfalign
5208: doc-maxalign
5209: doc-cfalign
5210:
5211:
5212: @node Heap Allocation, Memory Access, Dictionary allocation, Memory
5213: @subsection Heap allocation
5214: @cindex heap allocation
5215: @cindex dynamic allocation of memory
5216: @cindex memory-allocation word set
5217:
5218: @cindex contiguous regions and heap allocation
5219: Heap allocation supports deallocation of allocated memory in any
5220: order. Dictionary allocation is not affected by it (i.e., it does not
5221: end a contiguous region). In Gforth, these words are implemented using
5222: the standard C library calls malloc(), free() and resize().
5223:
5224: The memory region produced by one invocation of @code{allocate} or
5225: @code{resize} is internally contiguous. There is no contiguity between
5226: such a region and any other region (including others allocated from the
5227: heap).
5228:
5229: doc-allocate
5230: doc-free
5231: doc-resize
5232:
5233:
5234: @node Memory Access, Address arithmetic, Heap Allocation, Memory
5235: @subsection Memory Access
5236: @cindex memory access words
5237:
5238: doc-@
5239: doc-!
5240: doc-+!
5241: doc-c@
5242: doc-c!
5243: doc-2@
5244: doc-2!
5245: doc-f@
5246: doc-f!
5247: doc-sf@
5248: doc-sf!
5249: doc-df@
5250: doc-df!
5251:
5252:
5253: @node Address arithmetic, Memory Blocks, Memory Access, Memory
5254: @subsection Address arithmetic
5255: @cindex address arithmetic words
5256:
5257: Address arithmetic is the foundation on which you can build data
5258: structures like arrays, records (@pxref{Structures}) and objects
5259: (@pxref{Object-oriented Forth}).
5260:
5261: @cindex address unit
5262: @cindex au (address unit)
5263: ANS Forth does not specify the sizes of the data types. Instead, it
5264: offers a number of words for computing sizes and doing address
5265: arithmetic. Address arithmetic is performed in terms of address units
5266: (aus); on most systems the address unit is one byte. Note that a
5267: character may have more than one au, so @code{chars} is no noop (on
5268: platforms where it is a noop, it compiles to nothing).
5269:
5270: The basic address arithmetic words are @code{+} and @code{-}. E.g., if
5271: you have the address of a cell, perform @code{1 cells +}, and you will
5272: have the address of the next cell.
5273:
5274: @cindex contiguous regions and address arithmetic
5275: In ANS Forth you can perform address arithmetic only within a contiguous
5276: region, i.e., if you have an address into one region, you can only add
5277: and subtract such that the result is still within the region; you can
5278: only subtract or compare addresses from within the same contiguous
5279: region. Reasons: several contiguous regions can be arranged in memory
5280: in any way; on segmented systems addresses may have unusual
5281: representations, such that address arithmetic only works within a
5282: region. Gforth provides a few more guarantees (linear address space,
5283: dictionary grows upwards), but in general I have found it easy to stay
5284: within contiguous regions (exception: computing and comparing to the
5285: address just beyond the end of an array).
5286:
5287: @cindex alignment of addresses for types
5288: ANS Forth also defines words for aligning addresses for specific
5289: types. Many computers require that accesses to specific data types
5290: must only occur at specific addresses; e.g., that cells may only be
5291: accessed at addresses divisible by 4. Even if a machine allows unaligned
5292: accesses, it can usually perform aligned accesses faster.
5293:
5294: For the performance-conscious: alignment operations are usually only
5295: necessary during the definition of a data structure, not during the
5296: (more frequent) accesses to it.
5297:
5298: ANS Forth defines no words for character-aligning addresses. This is not
5299: an oversight, but reflects the fact that addresses that are not
5300: char-aligned have no use in the standard and therefore will not be
5301: created.
5302:
5303: @cindex @code{CREATE} and alignment
5304: ANS Forth guarantees that addresses returned by @code{CREATE}d words
5305: are cell-aligned; in addition, Gforth guarantees that these addresses
5306: are aligned for all purposes.
5307:
5308: Note that the ANS Forth word @code{char} has nothing to do with address
5309: arithmetic.
5310:
5311:
5312: doc-chars
5313: doc-char+
5314: doc-cells
5315: doc-cell+
5316: doc-cell
5317: doc-aligned
5318: doc-floats
5319: doc-float+
5320: doc-float
5321: doc-faligned
5322: doc-sfloats
5323: doc-sfloat+
5324: doc-sfaligned
5325: doc-dfloats
5326: doc-dfloat+
5327: doc-dfaligned
5328: doc-maxaligned
5329: doc-cfaligned
5330: doc-address-unit-bits
5331:
5332:
5333: @node Memory Blocks, , Address arithmetic, Memory
5334: @subsection Memory Blocks
5335: @cindex memory block words
5336: @cindex character strings - moving and copying
5337:
5338: Memory blocks often represent character strings; For ways of storing
5339: character strings in memory see @ref{String Formats}. For other
5340: string-processing words see @ref{Displaying characters and strings}.
5341:
5342: A few of these words work on address unit blocks. In that case, you
5343: usually have to insert @code{CHARS} before the word when working on
5344: character strings. Most words work on character blocks, and expect a
5345: char-aligned address.
5346:
5347: When copying characters between overlapping memory regions, use
5348: @code{chars move} or choose carefully between @code{cmove} and
5349: @code{cmove>}.
5350:
5351: doc-move
5352: doc-erase
5353: doc-cmove
5354: doc-cmove>
5355: doc-fill
5356: doc-blank
5357: doc-compare
5358: doc-search
5359: doc--trailing
5360: doc-/string
5361: doc-bounds
5362:
5363: @comment TODO examples
5364:
5365:
5366: @node Control Structures, Defining Words, Memory, Words
5367: @section Control Structures
5368: @cindex control structures
5369:
5370: Control structures in Forth cannot be used interpretively, only in a
5371: colon definition@footnote{To be precise, they have no interpretation
5372: semantics (@pxref{Interpretation and Compilation Semantics}).}. We do
5373: not like this limitation, but have not seen a satisfying way around it
5374: yet, although many schemes have been proposed.
5375:
5376: @menu
5377: * Selection:: IF ... ELSE ... ENDIF
5378: * Simple Loops:: BEGIN ...
5379: * Counted Loops:: DO
5380: * Arbitrary control structures::
5381: * Calls and returns::
5382: * Exception Handling::
5383: @end menu
5384:
5385: @node Selection, Simple Loops, Control Structures, Control Structures
5386: @subsection Selection
5387: @cindex selection control structures
5388: @cindex control structures for selection
5389:
5390: @cindex @code{IF} control structure
5391: @example
5392: @i{flag}
5393: IF
5394: @i{code}
5395: ENDIF
5396: @end example
5397: @noindent
5398:
5399: If @i{flag} is non-zero (as far as @code{IF} etc. are concerned, a cell
5400: with any bit set represents truth) @i{code} is executed.
5401:
5402: @example
5403: @i{flag}
5404: IF
5405: @i{code1}
5406: ELSE
5407: @i{code2}
5408: ENDIF
5409: @end example
5410:
5411: If @var{flag} is true, @i{code1} is executed, otherwise @i{code2} is
5412: executed.
5413:
5414: You can use @code{THEN} instead of @code{ENDIF}. Indeed, @code{THEN} is
5415: standard, and @code{ENDIF} is not, although it is quite popular. We
5416: recommend using @code{ENDIF}, because it is less confusing for people
5417: who also know other languages (and is not prone to reinforcing negative
5418: prejudices against Forth in these people). Adding @code{ENDIF} to a
5419: system that only supplies @code{THEN} is simple:
5420: @example
5421: : ENDIF POSTPONE then ; immediate
5422: @end example
5423:
5424: [According to @cite{Webster's New Encyclopedic Dictionary}, @dfn{then
5425: (adv.)} has the following meanings:
5426: @quotation
5427: ... 2b: following next after in order ... 3d: as a necessary consequence
5428: (if you were there, then you saw them).
5429: @end quotation
5430: Forth's @code{THEN} has the meaning 2b, whereas @code{THEN} in Pascal
5431: and many other programming languages has the meaning 3d.]
5432:
5433: Gforth also provides the words @code{?DUP-IF} and @code{?DUP-0=-IF}, so
5434: you can avoid using @code{?dup}. Using these alternatives is also more
5435: efficient than using @code{?dup}. Definitions in ANS Forth
5436: for @code{ENDIF}, @code{?DUP-IF} and @code{?DUP-0=-IF} are provided in
5437: @file{compat/control.fs}.
5438:
5439: @cindex @code{CASE} control structure
5440: @example
5441: @i{n}
5442: CASE
5443: @i{n1} OF @i{code1} ENDOF
5444: @i{n2} OF @i{code2} ENDOF
5445: @dots{}
5446: ( n ) @i{default-code} ( n )
5447: ENDCASE
5448: @end example
5449:
5450: Executes the first @i{codei}, where the @i{ni} is equal to @i{n}. If no
5451: @i{ni} matches, the optional @i{default-code} is executed. The optional
5452: default case can be added by simply writing the code after the last
5453: @code{ENDOF}. It may use @i{n}, which is on top of the stack, but must
5454: not consume it.
5455:
5456: @progstyle
5457: To keep the code understandable, you should ensure that on all paths
5458: through a selection construct the stack is changed in the same way
5459: (wrt. number and types of stack items consumed and pushed).
5460:
5461: @node Simple Loops, Counted Loops, Selection, Control Structures
5462: @subsection Simple Loops
5463: @cindex simple loops
5464: @cindex loops without count
5465:
5466: @cindex @code{WHILE} loop
5467: @example
5468: BEGIN
5469: @i{code1}
5470: @i{flag}
5471: WHILE
5472: @i{code2}
5473: REPEAT
5474: @end example
5475:
5476: @i{code1} is executed and @i{flag} is computed. If it is true,
5477: @i{code2} is executed and the loop is restarted; If @i{flag} is
5478: false, execution continues after the @code{REPEAT}.
5479:
5480: @cindex @code{UNTIL} loop
5481: @example
5482: BEGIN
5483: @i{code}
5484: @i{flag}
5485: UNTIL
5486: @end example
5487:
5488: @i{code} is executed. The loop is restarted if @code{flag} is false.
5489:
5490: @progstyle
5491: To keep the code understandable, a complete iteration of the loop should
5492: not change the number and types of the items on the stacks.
5493:
5494: @cindex endless loop
5495: @cindex loops, endless
5496: @example
5497: BEGIN
5498: @i{code}
5499: AGAIN
5500: @end example
5501:
5502: This is an endless loop.
5503:
5504: @node Counted Loops, Arbitrary control structures, Simple Loops, Control Structures
5505: @subsection Counted Loops
5506: @cindex counted loops
5507: @cindex loops, counted
5508: @cindex @code{DO} loops
5509:
5510: The basic counted loop is:
5511: @example
5512: @i{limit} @i{start}
5513: ?DO
5514: @i{body}
5515: LOOP
5516: @end example
5517:
5518: This performs one iteration for every integer, starting from @i{start}
5519: and up to, but excluding @i{limit}. The counter, or @i{index}, can be
5520: accessed with @code{i}. For example, the loop:
5521: @example
5522: 10 0 ?DO
5523: i .
5524: LOOP
5525: @end example
5526: @noindent
5527: prints @code{0 1 2 3 4 5 6 7 8 9}
5528:
5529: The index of the innermost loop can be accessed with @code{i}, the index
5530: of the next loop with @code{j}, and the index of the third loop with
5531: @code{k}.
5532:
5533:
5534: doc-i
5535: doc-j
5536: doc-k
5537:
5538:
5539: The loop control data are kept on the return stack, so there are some
5540: restrictions on mixing return stack accesses and counted loop words. In
5541: particuler, if you put values on the return stack outside the loop, you
5542: cannot read them inside the loop@footnote{well, not in a way that is
5543: portable.}. If you put values on the return stack within a loop, you
5544: have to remove them before the end of the loop and before accessing the
5545: index of the loop.
5546:
5547: There are several variations on the counted loop:
5548:
5549: @itemize @bullet
5550: @item
5551: @code{LEAVE} leaves the innermost counted loop immediately; execution
5552: continues after the associated @code{LOOP} or @code{NEXT}. For example:
5553:
5554: @example
5555: 10 0 ?DO i DUP . 3 = IF LEAVE THEN LOOP
5556: @end example
5557: prints @code{0 1 2 3}
5558:
5559:
5560: @item
5561: @code{UNLOOP} prepares for an abnormal loop exit, e.g., via
5562: @code{EXIT}. @code{UNLOOP} removes the loop control parameters from the
5563: return stack so @code{EXIT} can get to its return address. For example:
5564:
5565: @example
5566: : demo 10 0 ?DO i DUP . 3 = IF UNLOOP EXIT THEN LOOP ." Done" ;
5567: @end example
5568: prints @code{0 1 2 3}
5569:
5570:
5571: @item
5572: If @i{start} is greater than @i{limit}, a @code{?DO} loop is entered
5573: (and @code{LOOP} iterates until they become equal by wrap-around
5574: arithmetic). This behaviour is usually not what you want. Therefore,
5575: Gforth offers @code{+DO} and @code{U+DO} (as replacements for
5576: @code{?DO}), which do not enter the loop if @i{start} is greater than
5577: @i{limit}; @code{+DO} is for signed loop parameters, @code{U+DO} for
5578: unsigned loop parameters.
5579:
5580: @item
5581: @code{?DO} can be replaced by @code{DO}. @code{DO} always enters
5582: the loop, independent of the loop parameters. Do not use @code{DO}, even
5583: if you know that the loop is entered in any case. Such knowledge tends
5584: to become invalid during maintenance of a program, and then the
5585: @code{DO} will make trouble.
5586:
5587: @item
5588: @code{LOOP} can be replaced with @code{@i{n} +LOOP}; this updates the
5589: index by @i{n} instead of by 1. The loop is terminated when the border
5590: between @i{limit-1} and @i{limit} is crossed. E.g.:
5591:
5592: @example
5593: 4 0 +DO i . 2 +LOOP
5594: @end example
5595: @noindent
5596: prints @code{0 2}
5597:
5598: @example
5599: 4 1 +DO i . 2 +LOOP
5600: @end example
5601: @noindent
5602: prints @code{1 3}
5603:
5604: @item
5605: @cindex negative increment for counted loops
5606: @cindex counted loops with negative increment
5607: The behaviour of @code{@i{n} +LOOP} is peculiar when @i{n} is negative:
5608:
5609: @example
5610: -1 0 ?DO i . -1 +LOOP
5611: @end example
5612: @noindent
5613: prints @code{0 -1}
5614:
5615: @example
5616: 0 0 ?DO i . -1 +LOOP
5617: @end example
5618: prints nothing.
5619:
5620: Therefore we recommend avoiding @code{@i{n} +LOOP} with negative
5621: @i{n}. One alternative is @code{@i{u} -LOOP}, which reduces the
5622: index by @i{u} each iteration. The loop is terminated when the border
5623: between @i{limit+1} and @i{limit} is crossed. Gforth also provides
5624: @code{-DO} and @code{U-DO} for down-counting loops. E.g.:
5625:
5626: @example
5627: -2 0 -DO i . 1 -LOOP
5628: @end example
5629: @noindent
5630: prints @code{0 -1}
5631:
5632: @example
5633: -1 0 -DO i . 1 -LOOP
5634: @end example
5635: @noindent
5636: prints @code{0}
5637:
5638: @example
5639: 0 0 -DO i . 1 -LOOP
5640: @end example
5641: @noindent
5642: prints nothing.
5643:
5644: @end itemize
5645:
5646: Unfortunately, @code{+DO}, @code{U+DO}, @code{-DO}, @code{U-DO} and
5647: @code{-LOOP} are not defined in ANS Forth. However, an implementation
5648: for these words that uses only standard words is provided in
5649: @file{compat/loops.fs}.
5650:
5651:
5652: @cindex @code{FOR} loops
5653: Another counted loop is:
5654: @example
5655: @i{n}
5656: FOR
5657: @i{body}
5658: NEXT
5659: @end example
5660: This is the preferred loop of native code compiler writers who are too
5661: lazy to optimize @code{?DO} loops properly. This loop structure is not
5662: defined in ANS Forth. In Gforth, this loop iterates @i{n+1} times;
5663: @code{i} produces values starting with @i{n} and ending with 0. Other
5664: Forth systems may behave differently, even if they support @code{FOR}
5665: loops. To avoid problems, don't use @code{FOR} loops.
5666:
5667: @node Arbitrary control structures, Calls and returns, Counted Loops, Control Structures
5668: @subsection Arbitrary control structures
5669: @cindex control structures, user-defined
5670:
5671: @cindex control-flow stack
5672: ANS Forth permits and supports using control structures in a non-nested
5673: way. Information about incomplete control structures is stored on the
5674: control-flow stack. This stack may be implemented on the Forth data
5675: stack, and this is what we have done in Gforth.
5676:
5677: @cindex @code{orig}, control-flow stack item
5678: @cindex @code{dest}, control-flow stack item
5679: An @i{orig} entry represents an unresolved forward branch, a @i{dest}
5680: entry represents a backward branch target. A few words are the basis for
5681: building any control structure possible (except control structures that
5682: need storage, like calls, coroutines, and backtracking).
5683:
5684:
5685: doc-if
5686: doc-ahead
5687: doc-then
5688: doc-begin
5689: doc-until
5690: doc-again
5691: doc-cs-pick
5692: doc-cs-roll
5693:
5694:
5695: The Standard words @code{CS-PICK} and @code{CS-ROLL} allow you to
5696: manipulate the control-flow stack in a portable way. Without them, you
5697: would need to know how many stack items are occupied by a control-flow
5698: entry (many systems use one cell. In Gforth they currently take three,
5699: but this may change in the future).
5700:
5701: Some standard control structure words are built from these words:
5702:
5703:
5704: doc-else
5705: doc-while
5706: doc-repeat
5707:
5708:
5709: @noindent
5710: Gforth adds some more control-structure words:
5711:
5712:
5713: doc-endif
5714: doc-?dup-if
5715: doc-?dup-0=-if
5716:
5717:
5718: @noindent
5719: Counted loop words constitute a separate group of words:
5720:
5721:
5722: doc-?do
5723: doc-+do
5724: doc-u+do
5725: doc--do
5726: doc-u-do
5727: doc-do
5728: doc-for
5729: doc-loop
5730: doc-+loop
5731: doc--loop
5732: doc-next
5733: doc-leave
5734: doc-?leave
5735: doc-unloop
5736: doc-done
5737:
5738:
5739: The standard does not allow using @code{CS-PICK} and @code{CS-ROLL} on
5740: @i{do-sys}. Gforth allows it, but it's your job to ensure that for
5741: every @code{?DO} etc. there is exactly one @code{UNLOOP} on any path
5742: through the definition (@code{LOOP} etc. compile an @code{UNLOOP} on the
5743: fall-through path). Also, you have to ensure that all @code{LEAVE}s are
5744: resolved (by using one of the loop-ending words or @code{DONE}).
5745:
5746: @noindent
5747: Another group of control structure words are:
5748:
5749:
5750: doc-case
5751: doc-endcase
5752: doc-of
5753: doc-endof
5754:
5755:
5756: @i{case-sys} and @i{of-sys} cannot be processed using @code{CS-PICK} and
5757: @code{CS-ROLL}.
5758:
5759: @subsubsection Programming Style
5760: @cindex control structures programming style
5761: @cindex programming style, arbitrary control structures
5762:
5763: In order to ensure readability we recommend that you do not create
5764: arbitrary control structures directly, but define new control structure
5765: words for the control structure you want and use these words in your
5766: program. For example, instead of writing:
5767:
5768: @example
5769: BEGIN
5770: ...
5771: IF [ 1 CS-ROLL ]
5772: ...
5773: AGAIN THEN
5774: @end example
5775:
5776: @noindent
5777: we recommend defining control structure words, e.g.,
5778:
5779: @example
5780: : WHILE ( DEST -- ORIG DEST )
5781: POSTPONE IF
5782: 1 CS-ROLL ; immediate
5783:
5784: : REPEAT ( orig dest -- )
5785: POSTPONE AGAIN
5786: POSTPONE THEN ; immediate
5787: @end example
5788:
5789: @noindent
5790: and then using these to create the control structure:
5791:
5792: @example
5793: BEGIN
5794: ...
5795: WHILE
5796: ...
5797: REPEAT
5798: @end example
5799:
5800: That's much easier to read, isn't it? Of course, @code{REPEAT} and
5801: @code{WHILE} are predefined, so in this example it would not be
5802: necessary to define them.
5803:
5804: @node Calls and returns, Exception Handling, Arbitrary control structures, Control Structures
5805: @subsection Calls and returns
5806: @cindex calling a definition
5807: @cindex returning from a definition
5808:
5809: @cindex recursive definitions
5810: A definition can be called simply be writing the name of the definition
5811: to be called. Normally a definition is invisible during its own
5812: definition. If you want to write a directly recursive definition, you
5813: can use @code{recursive} to make the current definition visible, or
5814: @code{recurse} to call the current definition directly.
5815:
5816:
5817: doc-recursive
5818: doc-recurse
5819:
5820:
5821: @comment TODO add example of the two recursion methods
5822: @quotation
5823: @progstyle
5824: I prefer using @code{recursive} to @code{recurse}, because calling the
5825: definition by name is more descriptive (if the name is well-chosen) than
5826: the somewhat cryptic @code{recurse}. E.g., in a quicksort
5827: implementation, it is much better to read (and think) ``now sort the
5828: partitions'' than to read ``now do a recursive call''.
5829: @end quotation
5830:
5831: For mutual recursion, use @code{Defer}red words, like this:
5832:
5833: @example
5834: Defer foo
5835:
5836: : bar ( ... -- ... )
5837: ... foo ... ;
5838:
5839: :noname ( ... -- ... )
5840: ... bar ... ;
5841: IS foo
5842: @end example
5843:
5844: Deferred words are discussed in more detail in @ref{Deferred words}.
5845:
5846: The current definition returns control to the calling definition when
5847: the end of the definition is reached or @code{EXIT} is encountered.
5848:
5849: doc-exit
5850: doc-;s
5851:
5852:
5853: @node Exception Handling, , Calls and returns, Control Structures
5854: @subsection Exception Handling
5855: @cindex exceptions
5856:
5857: @c quit is a very bad idea for error handling,
5858: @c because it does not translate into a THROW
5859: @c it also does not belong into this chapter
5860:
5861: If a word detects an error condition that it cannot handle, it can
5862: @code{throw} an exception. In the simplest case, this will terminate
5863: your program, and report an appropriate error.
5864:
5865: doc-throw
5866:
5867: @code{Throw} consumes a cell-sized error number on the stack. There are
5868: some predefined error numbers in ANS Forth (see @file{errors.fs}). In
5869: Gforth (and most other systems) you can use the iors produced by various
5870: words as error numbers (e.g., a typical use of @code{allocate} is
5871: @code{allocate throw}). Gforth also provides the word @code{exception}
5872: to define your own error numbers (with decent error reporting); an ANS
5873: Forth version of this word (but without the error messages) is available
5874: in @code{compat/except.fs}. And finally, you can use your own error
5875: numbers (anything outside the range -4095..0), but won't get nice error
5876: messages, only numbers. For example, try:
5877:
5878: @example
5879: -10 throw \ ANS defined
5880: -267 throw \ system defined
5881: s" my error" exception throw \ user defined
5882: 7 throw \ arbitrary number
5883: @end example
5884:
5885: doc---exception-exception
5886:
5887: A common idiom to @code{THROW} a specific error if a flag is true is
5888: this:
5889:
5890: @example
5891: @code{( flag ) 0<> @i{errno} and throw}
5892: @end example
5893:
5894: Your program can provide exception handlers to catch exceptions. An
5895: exception handler can be used to correct the problem, or to clean up
5896: some data structures and just throw the exception to the next exception
5897: handler. Note that @code{throw} jumps to the dynamically innermost
5898: exception handler. The system's exception handler is outermost, and just
5899: prints an error and restarts command-line interpretation (or, in batch
5900: mode (i.e., while processing the shell command line), leaves Gforth).
5901:
5902: The ANS Forth way to catch exceptions is @code{catch}:
5903:
5904: doc-catch
5905:
5906: The most common use of exception handlers is to clean up the state when
5907: an error happens. E.g.,
5908:
5909: @example
5910: base @ >r hex \ actually the hex should be inside foo, or we h
5911: ['] foo catch ( nerror|0 )
5912: r> base !
5913: ( nerror|0 ) throw \ pass it on
5914: @end example
5915:
5916: A use of @code{catch} for handling the error @code{myerror} might look
5917: like this:
5918:
5919: @example
5920: ['] foo catch
5921: CASE
5922: myerror OF ... ( do something about it ) ENDOF
5923: dup throw \ default: pass other errors on, do nothing on non-errors
5924: ENDCASE
5925: @end example
5926:
5927: Having to wrap the code into a separate word is often cumbersome,
5928: therefore Gforth provides an alternative syntax:
5929:
5930: @example
5931: TRY
5932: @i{code1}
5933: RECOVER \ optional
5934: @i{code2} \ optional
5935: ENDTRY
5936: @end example
5937:
5938: This performs @i{Code1}. If @i{code1} completes normally, execution
5939: continues after the @code{endtry}. If @i{Code1} throws, the stacks are
5940: reset to the state during @code{try}, the throw value is pushed on the
5941: data stack, and execution constinues at @i{code2}, and finally falls
5942: through the @code{endtry} into the following code. If there is no
5943: @code{recover} clause, this works like an empty recover clause.
5944:
5945: doc-try
5946: doc-recover
5947: doc-endtry
5948:
5949: The cleanup example from above in this syntax:
5950:
5951: @example
5952: base @ >r TRY
5953: hex foo \ now the hex is placed correctly
5954: 0 \ value for throw
5955: ENDTRY
5956: r> base ! throw
5957: @end example
5958:
5959: And here's the error handling example:
5960:
5961: @example
5962: TRY
5963: foo
5964: RECOVER
5965: CASE
5966: myerror OF ... ( do something about it ) ENDOF
5967: throw \ pass other errors on
5968: ENDCASE
5969: ENDTRY
5970: @end example
5971:
5972: @progstyle
5973: As usual, you should ensure that the stack depth is statically known at
5974: the end: either after the @code{throw} for passing on errors, or after
5975: the @code{ENDTRY} (or, if you use @code{catch}, after the end of the
5976: selection construct for handling the error).
5977:
5978: There are two alternatives to @code{throw}: @code{Abort"} is conditional
5979: and you can provide an error message. @code{Abort} just produces an
5980: ``Aborted'' error.
5981:
5982: The problem with these words is that exception handlers cannot
5983: differentiate between different @code{abort"}s; they just look like
5984: @code{-2 throw} to them (the error message cannot be accessed by
5985: standard programs). Similar @code{abort} looks like @code{-1 throw} to
5986: exception handlers.
5987:
5988: doc-abort"
5989: doc-abort
5990:
5991:
5992:
5993: @c -------------------------------------------------------------
5994: @node Defining Words, Interpretation and Compilation Semantics, Control Structures, Words
5995: @section Defining Words
5996: @cindex defining words
5997:
5998: Defining words are used to extend Forth by creating new entries in the dictionary.
5999:
6000: @menu
6001: * CREATE::
6002: * Variables:: Variables and user variables
6003: * Constants::
6004: * Values:: Initialised variables
6005: * Colon Definitions::
6006: * Anonymous Definitions:: Definitions without names
6007: * Supplying names:: Passing definition names as strings
6008: * User-defined Defining Words::
6009: * Deferred words:: Allow forward references
6010: * Aliases::
6011: @end menu
6012:
6013: @node CREATE, Variables, Defining Words, Defining Words
6014: @subsection @code{CREATE}
6015: @cindex simple defining words
6016: @cindex defining words, simple
6017:
6018: Defining words are used to create new entries in the dictionary. The
6019: simplest defining word is @code{CREATE}. @code{CREATE} is used like
6020: this:
6021:
6022: @example
6023: CREATE new-word1
6024: @end example
6025:
6026: @code{CREATE} is a parsing word, i.e., it takes an argument from the
6027: input stream (@code{new-word1} in our example). It generates a
6028: dictionary entry for @code{new-word1}. When @code{new-word1} is
6029: executed, all that it does is leave an address on the stack. The address
6030: represents the value of the data space pointer (@code{HERE}) at the time
6031: that @code{new-word1} was defined. Therefore, @code{CREATE} is a way of
6032: associating a name with the address of a region of memory.
6033:
6034: doc-create
6035:
6036: Note that in ANS Forth guarantees only for @code{create} that its body
6037: is in dictionary data space (i.e., where @code{here}, @code{allot}
6038: etc. work, @pxref{Dictionary allocation}). Also, in ANS Forth only
6039: @code{create}d words can be modified with @code{does>}
6040: (@pxref{User-defined Defining Words}). And in ANS Forth @code{>body}
6041: can only be applied to @code{create}d words.
6042:
6043: By extending this example to reserve some memory in data space, we end
6044: up with something like a @i{variable}. Here are two different ways to do
6045: it:
6046:
6047: @example
6048: CREATE new-word2 1 cells allot \ reserve 1 cell - initial value undefined
6049: CREATE new-word3 4 , \ reserve 1 cell and initialise it (to 4)
6050: @end example
6051:
6052: The variable can be examined and modified using @code{@@} (``fetch'') and
6053: @code{!} (``store'') like this:
6054:
6055: @example
6056: new-word2 @@ . \ get address, fetch from it and display
6057: 1234 new-word2 ! \ new value, get address, store to it
6058: @end example
6059:
6060: @cindex arrays
6061: A similar mechanism can be used to create arrays. For example, an
6062: 80-character text input buffer:
6063:
6064: @example
6065: CREATE text-buf 80 chars allot
6066:
6067: text-buf 0 chars c@@ \ the 1st character (offset 0)
6068: text-buf 3 chars c@@ \ the 4th character (offset 3)
6069: @end example
6070:
6071: You can build arbitrarily complex data structures by allocating
6072: appropriate areas of memory. For further discussions of this, and to
6073: learn about some Gforth tools that make it easier,
6074: @xref{Structures}.
6075:
6076:
6077: @node Variables, Constants, CREATE, Defining Words
6078: @subsection Variables
6079: @cindex variables
6080:
6081: The previous section showed how a sequence of commands could be used to
6082: generate a variable. As a final refinement, the whole code sequence can
6083: be wrapped up in a defining word (pre-empting the subject of the next
6084: section), making it easier to create new variables:
6085:
6086: @example
6087: : myvariableX ( "name" -- a-addr ) CREATE 1 cells allot ;
6088: : myvariable0 ( "name" -- a-addr ) CREATE 0 , ;
6089:
6090: myvariableX foo \ variable foo starts off with an unknown value
6091: myvariable0 joe \ whilst joe is initialised to 0
6092:
6093: 45 3 * foo ! \ set foo to 135
6094: 1234 joe ! \ set joe to 1234
6095: 3 joe +! \ increment joe by 3.. to 1237
6096: @end example
6097:
6098: Not surprisingly, there is no need to define @code{myvariable}, since
6099: Forth already has a definition @code{Variable}. ANS Forth does not
6100: guarantee that a @code{Variable} is initialised when it is created
6101: (i.e., it may behave like @code{myvariableX}). In contrast, Gforth's
6102: @code{Variable} initialises the variable to 0 (i.e., it behaves exactly
6103: like @code{myvariable0}). Forth also provides @code{2Variable} and
6104: @code{fvariable} for double and floating-point variables, respectively
6105: -- they are initialised to 0. and 0e in Gforth. If you use a @code{Variable} to
6106: store a boolean, you can use @code{on} and @code{off} to toggle its
6107: state.
6108:
6109: doc-variable
6110: doc-2variable
6111: doc-fvariable
6112:
6113: @cindex user variables
6114: @cindex user space
6115: The defining word @code{User} behaves in the same way as @code{Variable}.
6116: The difference is that it reserves space in @i{user (data) space} rather
6117: than normal data space. In a Forth system that has a multi-tasker, each
6118: task has its own set of user variables.
6119:
6120: doc-user
6121: @c doc-udp
6122: @c doc-uallot
6123:
6124: @comment TODO is that stuff about user variables strictly correct? Is it
6125: @comment just terminal tasks that have user variables?
6126: @comment should document tasker.fs (with some examples) elsewhere
6127: @comment in this manual, then expand on user space and user variables.
6128:
6129: @node Constants, Values, Variables, Defining Words
6130: @subsection Constants
6131: @cindex constants
6132:
6133: @code{Constant} allows you to declare a fixed value and refer to it by
6134: name. For example:
6135:
6136: @example
6137: 12 Constant INCHES-PER-FOOT
6138: 3E+08 fconstant SPEED-O-LIGHT
6139: @end example
6140:
6141: A @code{Variable} can be both read and written, so its run-time
6142: behaviour is to supply an address through which its current value can be
6143: manipulated. In contrast, the value of a @code{Constant} cannot be
6144: changed once it has been declared@footnote{Well, often it can be -- but
6145: not in a Standard, portable way. It's safer to use a @code{Value} (read
6146: on).} so it's not necessary to supply the address -- it is more
6147: efficient to return the value of the constant directly. That's exactly
6148: what happens; the run-time effect of a constant is to put its value on
6149: the top of the stack (You can find one
6150: way of implementing @code{Constant} in @ref{User-defined Defining Words}).
6151:
6152: Forth also provides @code{2Constant} and @code{fconstant} for defining
6153: double and floating-point constants, respectively.
6154:
6155: doc-constant
6156: doc-2constant
6157: doc-fconstant
6158:
6159: @c that's too deep, and it's not necessarily true for all ANS Forths. - anton
6160: @c nac-> How could that not be true in an ANS Forth? You can't define a
6161: @c constant, use it and then delete the definition of the constant..
6162:
6163: @c anton->An ANS Forth system can compile a constant to a literal; On
6164: @c decompilation you would see only the number, just as if it had been used
6165: @c in the first place. The word will stay, of course, but it will only be
6166: @c used by the text interpreter (no run-time duties, except when it is
6167: @c POSTPONEd or somesuch).
6168:
6169: @c nac:
6170: @c I agree that it's rather deep, but IMO it is an important difference
6171: @c relative to other programming languages.. often it's annoying: it
6172: @c certainly changes my programming style relative to C.
6173:
6174: @c anton: In what way?
6175:
6176: Constants in Forth behave differently from their equivalents in other
6177: programming languages. In other languages, a constant (such as an EQU in
6178: assembler or a #define in C) only exists at compile-time; in the
6179: executable program the constant has been translated into an absolute
6180: number and, unless you are using a symbolic debugger, it's impossible to
6181: know what abstract thing that number represents. In Forth a constant has
6182: an entry in the header space and remains there after the code that uses
6183: it has been defined. In fact, it must remain in the dictionary since it
6184: has run-time duties to perform. For example:
6185:
6186: @example
6187: 12 Constant INCHES-PER-FOOT
6188: : FEET-TO-INCHES ( n1 -- n2 ) INCHES-PER-FOOT * ;
6189: @end example
6190:
6191: @cindex in-lining of constants
6192: When @code{FEET-TO-INCHES} is executed, it will in turn execute the xt
6193: associated with the constant @code{INCHES-PER-FOOT}. If you use
6194: @code{see} to decompile the definition of @code{FEET-TO-INCHES}, you can
6195: see that it makes a call to @code{INCHES-PER-FOOT}. Some Forth compilers
6196: attempt to optimise constants by in-lining them where they are used. You
6197: can force Gforth to in-line a constant like this:
6198:
6199: @example
6200: : FEET-TO-INCHES ( n1 -- n2 ) [ INCHES-PER-FOOT ] LITERAL * ;
6201: @end example
6202:
6203: If you use @code{see} to decompile @i{this} version of
6204: @code{FEET-TO-INCHES}, you can see that @code{INCHES-PER-FOOT} is no
6205: longer present. To understand how this works, read
6206: @ref{Interpret/Compile states}, and @ref{Literals}.
6207:
6208: In-lining constants in this way might improve execution time
6209: fractionally, and can ensure that a constant is now only referenced at
6210: compile-time. However, the definition of the constant still remains in
6211: the dictionary. Some Forth compilers provide a mechanism for controlling
6212: a second dictionary for holding transient words such that this second
6213: dictionary can be deleted later in order to recover memory
6214: space. However, there is no standard way of doing this.
6215:
6216:
6217: @node Values, Colon Definitions, Constants, Defining Words
6218: @subsection Values
6219: @cindex values
6220:
6221: A @code{Value} behaves like a @code{Constant}, but it can be changed.
6222: @code{TO} is a parsing word that changes a @code{Values}. In Gforth
6223: (not in ANS Forth) you can access (and change) a @code{value} also with
6224: @code{>body}.
6225:
6226: Here are some
6227: examples:
6228:
6229: @example
6230: 12 Value APPLES \ Define APPLES with an initial value of 12
6231: 34 TO APPLES \ Change the value of APPLES. TO is a parsing word
6232: 1 ' APPLES >body +! \ Increment APPLES. Non-standard usage.
6233: APPLES \ puts 35 on the top of the stack.
6234: @end example
6235:
6236: doc-value
6237: doc-to
6238:
6239:
6240:
6241: @node Colon Definitions, Anonymous Definitions, Values, Defining Words
6242: @subsection Colon Definitions
6243: @cindex colon definitions
6244:
6245: @example
6246: : name ( ... -- ... )
6247: word1 word2 word3 ;
6248: @end example
6249:
6250: @noindent
6251: Creates a word called @code{name} that, upon execution, executes
6252: @code{word1 word2 word3}. @code{name} is a @dfn{(colon) definition}.
6253:
6254: The explanation above is somewhat superficial. For simple examples of
6255: colon definitions see @ref{Your first definition}. For an in-depth
6256: discussion of some of the issues involved, @xref{Interpretation and
6257: Compilation Semantics}.
6258:
6259: doc-:
6260: doc-;
6261:
6262:
6263: @node Anonymous Definitions, Supplying names, Colon Definitions, Defining Words
6264: @subsection Anonymous Definitions
6265: @cindex colon definitions
6266: @cindex defining words without name
6267:
6268: Sometimes you want to define an @dfn{anonymous word}; a word without a
6269: name. You can do this with:
6270:
6271: doc-:noname
6272:
6273: This leaves the execution token for the word on the stack after the
6274: closing @code{;}. Here's an example in which a deferred word is
6275: initialised with an @code{xt} from an anonymous colon definition:
6276:
6277: @example
6278: Defer deferred
6279: :noname ( ... -- ... )
6280: ... ;
6281: IS deferred
6282: @end example
6283:
6284: @noindent
6285: Gforth provides an alternative way of doing this, using two separate
6286: words:
6287:
6288: doc-noname
6289: @cindex execution token of last defined word
6290: doc-lastxt
6291:
6292: @noindent
6293: The previous example can be rewritten using @code{noname} and
6294: @code{lastxt}:
6295:
6296: @example
6297: Defer deferred
6298: noname : ( ... -- ... )
6299: ... ;
6300: lastxt IS deferred
6301: @end example
6302:
6303: @noindent
6304: @code{noname} works with any defining word, not just @code{:}.
6305:
6306: @code{lastxt} also works when the last word was not defined as
6307: @code{noname}. It does not work for combined words, though. It also has
6308: the useful property that is is valid as soon as the header for a
6309: definition has been built. Thus:
6310:
6311: @example
6312: lastxt . : foo [ lastxt . ] ; ' foo .
6313: @end example
6314:
6315: @noindent
6316: prints 3 numbers; the last two are the same.
6317:
6318: @node Supplying names, User-defined Defining Words, Anonymous Definitions, Defining Words
6319: @subsection Supplying the name of a defined word
6320: @cindex names for defined words
6321: @cindex defining words, name given in a string
6322:
6323: By default, a defining word takes the name for the defined word from the
6324: input stream. Sometimes you want to supply the name from a string. You
6325: can do this with:
6326:
6327: doc-nextname
6328:
6329: For example:
6330:
6331: @example
6332: s" foo" nextname create
6333: @end example
6334:
6335: @noindent
6336: is equivalent to:
6337:
6338: @example
6339: create foo
6340: @end example
6341:
6342: @noindent
6343: @code{nextname} works with any defining word.
6344:
6345:
6346: @node User-defined Defining Words, Deferred words, Supplying names, Defining Words
6347: @subsection User-defined Defining Words
6348: @cindex user-defined defining words
6349: @cindex defining words, user-defined
6350:
6351: You can create a new defining word by wrapping defining-time code around
6352: an existing defining word and putting the sequence in a colon
6353: definition.
6354:
6355: @c anton: This example is very complex and leads in a quite different
6356: @c direction from the CREATE-DOES> stuff that follows. It should probably
6357: @c be done elsewhere, or as a subsubsection of this subsection (or as a
6358: @c subsection of Defining Words)
6359:
6360: For example, suppose that you have a word @code{stats} that
6361: gathers statistics about colon definitions given the @i{xt} of the
6362: definition, and you want every colon definition in your application to
6363: make a call to @code{stats}. You can define and use a new version of
6364: @code{:} like this:
6365:
6366: @example
6367: : stats ( xt -- ) DUP ." (Gathering statistics for " . ." )"
6368: ... ; \ other code
6369:
6370: : my: : lastxt postpone literal ['] stats compile, ;
6371:
6372: my: foo + - ;
6373: @end example
6374:
6375: When @code{foo} is defined using @code{my:} these steps occur:
6376:
6377: @itemize @bullet
6378: @item
6379: @code{my:} is executed.
6380: @item
6381: The @code{:} within the definition (the one between @code{my:} and
6382: @code{lastxt}) is executed, and does just what it always does; it parses
6383: the input stream for a name, builds a dictionary header for the name
6384: @code{foo} and switches @code{state} from interpret to compile.
6385: @item
6386: The word @code{lastxt} is executed. It puts the @i{xt} for the word that is
6387: being defined -- @code{foo} -- onto the stack.
6388: @item
6389: The code that was produced by @code{postpone literal} is executed; this
6390: causes the value on the stack to be compiled as a literal in the code
6391: area of @code{foo}.
6392: @item
6393: The code @code{['] stats} compiles a literal into the definition of
6394: @code{my:}. When @code{compile,} is executed, that literal -- the
6395: execution token for @code{stats} -- is layed down in the code area of
6396: @code{foo} , following the literal@footnote{Strictly speaking, the
6397: mechanism that @code{compile,} uses to convert an @i{xt} into something
6398: in the code area is implementation-dependent. A threaded implementation
6399: might spit out the execution token directly whilst another
6400: implementation might spit out a native code sequence.}.
6401: @item
6402: At this point, the execution of @code{my:} is complete, and control
6403: returns to the text interpreter. The text interpreter is in compile
6404: state, so subsequent text @code{+ -} is compiled into the definition of
6405: @code{foo} and the @code{;} terminates the definition as always.
6406: @end itemize
6407:
6408: You can use @code{see} to decompile a word that was defined using
6409: @code{my:} and see how it is different from a normal @code{:}
6410: definition. For example:
6411:
6412: @example
6413: : bar + - ; \ like foo but using : rather than my:
6414: see bar
6415: : bar
6416: + - ;
6417: see foo
6418: : foo
6419: 107645672 stats + - ;
6420:
6421: \ use ' stats . to show that 107645672 is the xt for stats
6422: @end example
6423:
6424: You can use techniques like this to make new defining words in terms of
6425: @i{any} existing defining word.
6426:
6427:
6428: @cindex defining defining words
6429: @cindex @code{CREATE} ... @code{DOES>}
6430: If you want the words defined with your defining words to behave
6431: differently from words defined with standard defining words, you can
6432: write your defining word like this:
6433:
6434: @example
6435: : def-word ( "name" -- )
6436: CREATE @i{code1}
6437: DOES> ( ... -- ... )
6438: @i{code2} ;
6439:
6440: def-word name
6441: @end example
6442:
6443: @cindex child words
6444: This fragment defines a @dfn{defining word} @code{def-word} and then
6445: executes it. When @code{def-word} executes, it @code{CREATE}s a new
6446: word, @code{name}, and executes the code @i{code1}. The code @i{code2}
6447: is not executed at this time. The word @code{name} is sometimes called a
6448: @dfn{child} of @code{def-word}.
6449:
6450: When you execute @code{name}, the address of the body of @code{name} is
6451: put on the data stack and @i{code2} is executed (the address of the body
6452: of @code{name} is the address @code{HERE} returns immediately after the
6453: @code{CREATE}, i.e., the address a @code{create}d word returns by
6454: default).
6455:
6456: @c anton:
6457: @c www.dictionary.com says:
6458: @c at·a·vism: 1.The reappearance of a characteristic in an organism after
6459: @c several generations of absence, usually caused by the chance
6460: @c recombination of genes. 2.An individual or a part that exhibits
6461: @c atavism. Also called throwback. 3.The return of a trait or recurrence
6462: @c of previous behavior after a period of absence.
6463: @c
6464: @c Doesn't seem to fit.
6465:
6466: @c @cindex atavism in child words
6467: You can use @code{def-word} to define a set of child words that behave
6468: similarly; they all have a common run-time behaviour determined by
6469: @i{code2}. Typically, the @i{code1} sequence builds a data area in the
6470: body of the child word. The structure of the data is common to all
6471: children of @code{def-word}, but the data values are specific -- and
6472: private -- to each child word. When a child word is executed, the
6473: address of its private data area is passed as a parameter on TOS to be
6474: used and manipulated@footnote{It is legitimate both to read and write to
6475: this data area.} by @i{code2}.
6476:
6477: The two fragments of code that make up the defining words act (are
6478: executed) at two completely separate times:
6479:
6480: @itemize @bullet
6481: @item
6482: At @i{define time}, the defining word executes @i{code1} to generate a
6483: child word
6484: @item
6485: At @i{child execution time}, when a child word is invoked, @i{code2}
6486: is executed, using parameters (data) that are private and specific to
6487: the child word.
6488: @end itemize
6489:
6490: Another way of understanding the behaviour of @code{def-word} and
6491: @code{name} is to say that, if you make the following definitions:
6492: @example
6493: : def-word1 ( "name" -- )
6494: CREATE @i{code1} ;
6495:
6496: : action1 ( ... -- ... )
6497: @i{code2} ;
6498:
6499: def-word1 name1
6500: @end example
6501:
6502: @noindent
6503: Then using @code{name1 action1} is equivalent to using @code{name}.
6504:
6505: The classic example is that you can define @code{CONSTANT} in this way:
6506:
6507: @example
6508: : CONSTANT ( w "name" -- )
6509: CREATE ,
6510: DOES> ( -- w )
6511: @@ ;
6512: @end example
6513:
6514: @comment There is a beautiful description of how this works and what
6515: @comment it does in the Forthwrite 100th edition.. as well as an elegant
6516: @comment commentary on the Counting Fruits problem.
6517:
6518: When you create a constant with @code{5 CONSTANT five}, a set of
6519: define-time actions take place; first a new word @code{five} is created,
6520: then the value 5 is laid down in the body of @code{five} with
6521: @code{,}. When @code{five} is executed, the address of the body is put on
6522: the stack, and @code{@@} retrieves the value 5. The word @code{five} has
6523: no code of its own; it simply contains a data field and a pointer to the
6524: code that follows @code{DOES>} in its defining word. That makes words
6525: created in this way very compact.
6526:
6527: The final example in this section is intended to remind you that space
6528: reserved in @code{CREATE}d words is @i{data} space and therefore can be
6529: both read and written by a Standard program@footnote{Exercise: use this
6530: example as a starting point for your own implementation of @code{Value}
6531: and @code{TO} -- if you get stuck, investigate the behaviour of @code{'} and
6532: @code{[']}.}:
6533:
6534: @example
6535: : foo ( "name" -- )
6536: CREATE -1 ,
6537: DOES> ( -- )
6538: @@ . ;
6539:
6540: foo first-word
6541: foo second-word
6542:
6543: 123 ' first-word >BODY !
6544: @end example
6545:
6546: If @code{first-word} had been a @code{CREATE}d word, we could simply
6547: have executed it to get the address of its data field. However, since it
6548: was defined to have @code{DOES>} actions, its execution semantics are to
6549: perform those @code{DOES>} actions. To get the address of its data field
6550: it's necessary to use @code{'} to get its xt, then @code{>BODY} to
6551: translate the xt into the address of the data field. When you execute
6552: @code{first-word}, it will display @code{123}. When you execute
6553: @code{second-word} it will display @code{-1}.
6554:
6555: @cindex stack effect of @code{DOES>}-parts
6556: @cindex @code{DOES>}-parts, stack effect
6557: In the examples above the stack comment after the @code{DOES>} specifies
6558: the stack effect of the defined words, not the stack effect of the
6559: following code (the following code expects the address of the body on
6560: the top of stack, which is not reflected in the stack comment). This is
6561: the convention that I use and recommend (it clashes a bit with using
6562: locals declarations for stack effect specification, though).
6563:
6564: @menu
6565: * CREATE..DOES> applications::
6566: * CREATE..DOES> details::
6567: * Advanced does> usage example::
6568: @end menu
6569:
6570: @node CREATE..DOES> applications, CREATE..DOES> details, User-defined Defining Words, User-defined Defining Words
6571: @subsubsection Applications of @code{CREATE..DOES>}
6572: @cindex @code{CREATE} ... @code{DOES>}, applications
6573:
6574: You may wonder how to use this feature. Here are some usage patterns:
6575:
6576: @cindex factoring similar colon definitions
6577: When you see a sequence of code occurring several times, and you can
6578: identify a meaning, you will factor it out as a colon definition. When
6579: you see similar colon definitions, you can factor them using
6580: @code{CREATE..DOES>}. E.g., an assembler usually defines several words
6581: that look very similar:
6582: @example
6583: : ori, ( reg-target reg-source n -- )
6584: 0 asm-reg-reg-imm ;
6585: : andi, ( reg-target reg-source n -- )
6586: 1 asm-reg-reg-imm ;
6587: @end example
6588:
6589: @noindent
6590: This could be factored with:
6591: @example
6592: : reg-reg-imm ( op-code -- )
6593: CREATE ,
6594: DOES> ( reg-target reg-source n -- )
6595: @@ asm-reg-reg-imm ;
6596:
6597: 0 reg-reg-imm ori,
6598: 1 reg-reg-imm andi,
6599: @end example
6600:
6601: @cindex currying
6602: Another view of @code{CREATE..DOES>} is to consider it as a crude way to
6603: supply a part of the parameters for a word (known as @dfn{currying} in
6604: the functional language community). E.g., @code{+} needs two
6605: parameters. Creating versions of @code{+} with one parameter fixed can
6606: be done like this:
6607:
6608: @example
6609: : curry+ ( n1 "name" -- )
6610: CREATE ,
6611: DOES> ( n2 -- n1+n2 )
6612: @@ + ;
6613:
6614: 3 curry+ 3+
6615: -2 curry+ 2-
6616: @end example
6617:
6618: @node CREATE..DOES> details, Advanced does> usage example, CREATE..DOES> applications, User-defined Defining Words
6619: @subsubsection The gory details of @code{CREATE..DOES>}
6620: @cindex @code{CREATE} ... @code{DOES>}, details
6621:
6622: doc-does>
6623:
6624: @cindex @code{DOES>} in a separate definition
6625: This means that you need not use @code{CREATE} and @code{DOES>} in the
6626: same definition; you can put the @code{DOES>}-part in a separate
6627: definition. This allows us to, e.g., select among different @code{DOES>}-parts:
6628: @example
6629: : does1
6630: DOES> ( ... -- ... )
6631: ... ;
6632:
6633: : does2
6634: DOES> ( ... -- ... )
6635: ... ;
6636:
6637: : def-word ( ... -- ... )
6638: create ...
6639: IF
6640: does1
6641: ELSE
6642: does2
6643: ENDIF ;
6644: @end example
6645:
6646: In this example, the selection of whether to use @code{does1} or
6647: @code{does2} is made at definition-time; at the time that the child word is
6648: @code{CREATE}d.
6649:
6650: @cindex @code{DOES>} in interpretation state
6651: In a standard program you can apply a @code{DOES>}-part only if the last
6652: word was defined with @code{CREATE}. In Gforth, the @code{DOES>}-part
6653: will override the behaviour of the last word defined in any case. In a
6654: standard program, you can use @code{DOES>} only in a colon
6655: definition. In Gforth, you can also use it in interpretation state, in a
6656: kind of one-shot mode; for example:
6657: @example
6658: CREATE name ( ... -- ... )
6659: @i{initialization}
6660: DOES>
6661: @i{code} ;
6662: @end example
6663:
6664: @noindent
6665: is equivalent to the standard:
6666: @example
6667: :noname
6668: DOES>
6669: @i{code} ;
6670: CREATE name EXECUTE ( ... -- ... )
6671: @i{initialization}
6672: @end example
6673:
6674: doc->body
6675:
6676: @node Advanced does> usage example, , CREATE..DOES> details, User-defined Defining Words
6677: @subsubsection Advanced does> usage example
6678:
6679: The MIPS disassembler (@file{arch/mips/disasm.fs}) contains many words
6680: for disassembling instructions, that follow a very repetetive scheme:
6681:
6682: @example
6683: :noname @var{disasm-operands} s" @var{inst-name}" type ;
6684: @var{entry-num} cells @var{table} + !
6685: @end example
6686:
6687: Of course, this inspires the idea to factor out the commonalities to
6688: allow a definition like
6689:
6690: @example
6691: @var{disasm-operands} @var{entry-num} @var{table} define-inst @var{inst-name}
6692: @end example
6693:
6694: The parameters @var{disasm-operands} and @var{table} are usually
6695: correlated. Moreover, before I wrote the disassembler, there already
6696: existed code that defines instructions like this:
6697:
6698: @example
6699: @var{entry-num} @var{inst-format} @var{inst-name}
6700: @end example
6701:
6702: This code comes from the assembler and resides in
6703: @file{arch/mips/insts.fs}.
6704:
6705: So I had to define the @var{inst-format} words that performed the scheme
6706: above when executed. At first I chose to use run-time code-generation:
6707:
6708: @example
6709: : @var{inst-format} ( entry-num "name" -- ; compiled code: addr w -- )
6710: :noname Postpone @var{disasm-operands}
6711: name Postpone sliteral Postpone type Postpone ;
6712: swap cells @var{table} + ! ;
6713: @end example
6714:
6715: Note that this supplies the other two parameters of the scheme above.
6716:
6717: An alternative would have been to write this using
6718: @code{create}/@code{does>}:
6719:
6720: @example
6721: : @var{inst-format} ( entry-num "name" -- )
6722: here name string, ( entry-num c-addr ) \ parse and save "name"
6723: noname create , ( entry-num )
6724: lastxt swap cells @var{table} + !
6725: does> ( addr w -- )
6726: \ disassemble instruction w at addr
6727: @@ >r
6728: @var{disasm-operands}
6729: r> count type ;
6730: @end example
6731:
6732: Somehow the first solution is simpler, mainly because it's simpler to
6733: shift a string from definition-time to use-time with @code{sliteral}
6734: than with @code{string,} and friends.
6735:
6736: I wrote a lot of words following this scheme and soon thought about
6737: factoring out the commonalities among them. Note that this uses a
6738: two-level defining word, i.e., a word that defines ordinary defining
6739: words.
6740:
6741: This time a solution involving @code{postpone} and friends seemed more
6742: difficult (try it as an exercise), so I decided to use a
6743: @code{create}/@code{does>} word; since I was already at it, I also used
6744: @code{create}/@code{does>} for the lower level (try using
6745: @code{postpone} etc. as an exercise), resulting in the following
6746: definition:
6747:
6748: @example
6749: : define-format ( disasm-xt table-xt -- )
6750: \ define an instruction format that uses disasm-xt for
6751: \ disassembling and enters the defined instructions into table
6752: \ table-xt
6753: create 2,
6754: does> ( u "inst" -- )
6755: \ defines an anonymous word for disassembling instruction inst,
6756: \ and enters it as u-th entry into table-xt
6757: 2@@ swap here name string, ( u table-xt disasm-xt c-addr ) \ remember string
6758: noname create 2, \ define anonymous word
6759: execute lastxt swap ! \ enter xt of defined word into table-xt
6760: does> ( addr w -- )
6761: \ disassemble instruction w at addr
6762: 2@@ >r ( addr w disasm-xt R: c-addr )
6763: execute ( R: c-addr ) \ disassemble operands
6764: r> count type ; \ print name
6765: @end example
6766:
6767: Note that the tables here (in contrast to above) do the @code{cells +}
6768: by themselves (that's why you have to pass an xt). This word is used in
6769: the following way:
6770:
6771: @example
6772: ' @var{disasm-operands} ' @var{table} define-format @var{inst-format}
6773: @end example
6774:
6775: As shown above, the defined instruction format is then used like this:
6776:
6777: @example
6778: @var{entry-num} @var{inst-format} @var{inst-name}
6779: @end example
6780:
6781: In terms of currying, this kind of two-level defining word provides the
6782: parameters in three stages: first @var{disasm-operands} and @var{table},
6783: then @var{entry-num} and @var{inst-name}, finally @code{addr w}, i.e.,
6784: the instruction to be disassembled.
6785:
6786: Of course this did not quite fit all the instruction format names used
6787: in @file{insts.fs}, so I had to define a few wrappers that conditioned
6788: the parameters into the right form.
6789:
6790: If you have trouble following this section, don't worry. First, this is
6791: involved and takes time (and probably some playing around) to
6792: understand; second, this is the first two-level
6793: @code{create}/@code{does>} word I have written in seventeen years of
6794: Forth; and if I did not have @file{insts.fs} to start with, I may well
6795: have elected to use just a one-level defining word (with some repeating
6796: of parameters when using the defining word). So it is not necessary to
6797: understand this, but it may improve your understanding of Forth.
6798:
6799:
6800: @node Deferred words, Aliases, User-defined Defining Words, Defining Words
6801: @subsection Deferred words
6802: @cindex deferred words
6803:
6804: The defining word @code{Defer} allows you to define a word by name
6805: without defining its behaviour; the definition of its behaviour is
6806: deferred. Here are two situation where this can be useful:
6807:
6808: @itemize @bullet
6809: @item
6810: Where you want to allow the behaviour of a word to be altered later, and
6811: for all precompiled references to the word to change when its behaviour
6812: is changed.
6813: @item
6814: For mutual recursion; @xref{Calls and returns}.
6815: @end itemize
6816:
6817: In the following example, @code{foo} always invokes the version of
6818: @code{greet} that prints ``@code{Good morning}'' whilst @code{bar}
6819: always invokes the version that prints ``@code{Hello}''. There is no way
6820: of getting @code{foo} to use the later version without re-ordering the
6821: source code and recompiling it.
6822:
6823: @example
6824: : greet ." Good morning" ;
6825: : foo ... greet ... ;
6826: : greet ." Hello" ;
6827: : bar ... greet ... ;
6828: @end example
6829:
6830: This problem can be solved by defining @code{greet} as a @code{Defer}red
6831: word. The behaviour of a @code{Defer}red word can be defined and
6832: redefined at any time by using @code{IS} to associate the xt of a
6833: previously-defined word with it. The previous example becomes:
6834:
6835: @example
6836: Defer greet ( -- )
6837: : foo ... greet ... ;
6838: : bar ... greet ... ;
6839: : greet1 ( -- ) ." Good morning" ;
6840: : greet2 ( -- ) ." Hello" ;
6841: ' greet2 <IS> greet \ make greet behave like greet2
6842: @end example
6843:
6844: @progstyle
6845: You should write a stack comment for every deferred word, and put only
6846: XTs into deferred words that conform to this stack effect. Otherwise
6847: it's too difficult to use the deferred word.
6848:
6849: A deferred word can be used to improve the statistics-gathering example
6850: from @ref{User-defined Defining Words}; rather than edit the
6851: application's source code to change every @code{:} to a @code{my:}, do
6852: this:
6853:
6854: @example
6855: : real: : ; \ retain access to the original
6856: defer : \ redefine as a deferred word
6857: ' my: <IS> : \ use special version of :
6858: \
6859: \ load application here
6860: \
6861: ' real: <IS> : \ go back to the original
6862: @end example
6863:
6864:
6865: One thing to note is that @code{<IS>} consumes its name when it is
6866: executed. If you want to specify the name at compile time, use
6867: @code{[IS]}:
6868:
6869: @example
6870: : set-greet ( xt -- )
6871: [IS] greet ;
6872:
6873: ' greet1 set-greet
6874: @end example
6875:
6876: A deferred word can only inherit execution semantics from the xt
6877: (because that is all that an xt can represent -- for more discussion of
6878: this @pxref{Tokens for Words}); by default it will have default
6879: interpretation and compilation semantics deriving from this execution
6880: semantics. However, you can change the interpretation and compilation
6881: semantics of the deferred word in the usual ways:
6882:
6883: @example
6884: : bar .... ; compile-only
6885: Defer fred immediate
6886: Defer jim
6887:
6888: ' bar <IS> jim \ jim has default semantics
6889: ' bar <IS> fred \ fred is immediate
6890: @end example
6891:
6892: doc-defer
6893: doc-<is>
6894: doc-[is]
6895: doc-is
6896: @comment TODO document these: what's defers [is]
6897: doc-what's
6898: doc-defers
6899:
6900: @c Use @code{words-deferred} to see a list of deferred words.
6901:
6902: Definitions in ANS Forth for @code{defer}, @code{<is>} and @code{[is]}
6903: are provided in @file{compat/defer.fs}.
6904:
6905:
6906: @node Aliases, , Deferred words, Defining Words
6907: @subsection Aliases
6908: @cindex aliases
6909:
6910: The defining word @code{Alias} allows you to define a word by name that
6911: has the same behaviour as some other word. Here are two situation where
6912: this can be useful:
6913:
6914: @itemize @bullet
6915: @item
6916: When you want access to a word's definition from a different word list
6917: (for an example of this, see the definition of the @code{Root} word list
6918: in the Gforth source).
6919: @item
6920: When you want to create a synonym; a definition that can be known by
6921: either of two names (for example, @code{THEN} and @code{ENDIF} are
6922: aliases).
6923: @end itemize
6924:
6925: Like deferred words, an alias has default compilation and interpretation
6926: semantics at the beginning (not the modifications of the other word),
6927: but you can change them in the usual ways (@code{immediate},
6928: @code{compile-only}). For example:
6929:
6930: @example
6931: : foo ... ; immediate
6932:
6933: ' foo Alias bar \ bar is not an immediate word
6934: ' foo Alias fooby immediate \ fooby is an immediate word
6935: @end example
6936:
6937: Words that are aliases have the same xt, different headers in the
6938: dictionary, and consequently different name tokens (@pxref{Tokens for
6939: Words}) and possibly different immediate flags. An alias can only have
6940: default or immediate compilation semantics; you can define aliases for
6941: combined words with @code{interpret/compile:} -- see @ref{Combined words}.
6942:
6943: doc-alias
6944:
6945:
6946: @node Interpretation and Compilation Semantics, Tokens for Words, Defining Words, Words
6947: @section Interpretation and Compilation Semantics
6948: @cindex semantics, interpretation and compilation
6949:
6950: @c !! state and ' are used without explanation
6951: @c example for immediate/compile-only? or is the tutorial enough
6952:
6953: @cindex interpretation semantics
6954: The @dfn{interpretation semantics} of a (named) word are what the text
6955: interpreter does when it encounters the word in interpret state. It also
6956: appears in some other contexts, e.g., the execution token returned by
6957: @code{' @i{word}} identifies the interpretation semantics of @i{word}
6958: (in other words, @code{' @i{word} execute} is equivalent to
6959: interpret-state text interpretation of @code{@i{word}}).
6960:
6961: @cindex compilation semantics
6962: The @dfn{compilation semantics} of a (named) word are what the text
6963: interpreter does when it encounters the word in compile state. It also
6964: appears in other contexts, e.g, @code{POSTPONE @i{word}}
6965: compiles@footnote{In standard terminology, ``appends to the current
6966: definition''.} the compilation semantics of @i{word}.
6967:
6968: @cindex execution semantics
6969: The standard also talks about @dfn{execution semantics}. They are used
6970: only for defining the interpretation and compilation semantics of many
6971: words. By default, the interpretation semantics of a word are to
6972: @code{execute} its execution semantics, and the compilation semantics of
6973: a word are to @code{compile,} its execution semantics.@footnote{In
6974: standard terminology: The default interpretation semantics are its
6975: execution semantics; the default compilation semantics are to append its
6976: execution semantics to the execution semantics of the current
6977: definition.}
6978:
6979: Unnamed words (@pxref{Anonymous Definitions}) cannot be encountered by
6980: the text interpreter, ticked, or @code{postpone}d, so they have no
6981: interpretation or compilation semantics. Their behaviour is represented
6982: by their XT (@pxref{Tokens for Words}), and we call it execution
6983: semantics, too.
6984:
6985: @comment TODO expand, make it co-operate with new sections on text interpreter.
6986:
6987: @cindex immediate words
6988: @cindex compile-only words
6989: You can change the semantics of the most-recently defined word:
6990:
6991:
6992: doc-immediate
6993: doc-compile-only
6994: doc-restrict
6995:
6996: By convention, words with non-default compilation semantics (e.g.,
6997: immediate words) often have names surrounded with brackets (e.g.,
6998: @code{[']}, @pxref{Execution token}).
6999:
7000: Note that ticking (@code{'}) a compile-only word gives an error
7001: (``Interpreting a compile-only word'').
7002:
7003: @menu
7004: * Combined words::
7005: @end menu
7006:
7007:
7008: @node Combined words, , Interpretation and Compilation Semantics, Interpretation and Compilation Semantics
7009: @subsection Combined Words
7010: @cindex combined words
7011:
7012: Gforth allows you to define @dfn{combined words} -- words that have an
7013: arbitrary combination of interpretation and compilation semantics.
7014:
7015: doc-interpret/compile:
7016:
7017: This feature was introduced for implementing @code{TO} and @code{S"}. I
7018: recommend that you do not define such words, as cute as they may be:
7019: they make it hard to get at both parts of the word in some contexts.
7020: E.g., assume you want to get an execution token for the compilation
7021: part. Instead, define two words, one that embodies the interpretation
7022: part, and one that embodies the compilation part. Once you have done
7023: that, you can define a combined word with @code{interpret/compile:} for
7024: the convenience of your users.
7025:
7026: You might try to use this feature to provide an optimizing
7027: implementation of the default compilation semantics of a word. For
7028: example, by defining:
7029: @example
7030: :noname
7031: foo bar ;
7032: :noname
7033: POSTPONE foo POSTPONE bar ;
7034: interpret/compile: opti-foobar
7035: @end example
7036:
7037: @noindent
7038: as an optimizing version of:
7039:
7040: @example
7041: : foobar
7042: foo bar ;
7043: @end example
7044:
7045: Unfortunately, this does not work correctly with @code{[compile]},
7046: because @code{[compile]} assumes that the compilation semantics of all
7047: @code{interpret/compile:} words are non-default. I.e., @code{[compile]
7048: opti-foobar} would compile compilation semantics, whereas
7049: @code{[compile] foobar} would compile interpretation semantics.
7050:
7051: @cindex state-smart words (are a bad idea)
7052: @anchor{state-smartness}
7053: Some people try to use @dfn{state-smart} words to emulate the feature provided
7054: by @code{interpret/compile:} (words are state-smart if they check
7055: @code{STATE} during execution). E.g., they would try to code
7056: @code{foobar} like this:
7057:
7058: @example
7059: : foobar
7060: STATE @@
7061: IF ( compilation state )
7062: POSTPONE foo POSTPONE bar
7063: ELSE
7064: foo bar
7065: ENDIF ; immediate
7066: @end example
7067:
7068: Although this works if @code{foobar} is only processed by the text
7069: interpreter, it does not work in other contexts (like @code{'} or
7070: @code{POSTPONE}). E.g., @code{' foobar} will produce an execution token
7071: for a state-smart word, not for the interpretation semantics of the
7072: original @code{foobar}; when you execute this execution token (directly
7073: with @code{EXECUTE} or indirectly through @code{COMPILE,}) in compile
7074: state, the result will not be what you expected (i.e., it will not
7075: perform @code{foo bar}). State-smart words are a bad idea. Simply don't
7076: write them@footnote{For a more detailed discussion of this topic, see
7077: M. Anton Ertl,
7078: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl98.ps.gz,@code{State}-smartness---Why
7079: it is Evil and How to Exorcise it}}, EuroForth '98.}!
7080:
7081: @cindex defining words with arbitrary semantics combinations
7082: It is also possible to write defining words that define words with
7083: arbitrary combinations of interpretation and compilation semantics. In
7084: general, they look like this:
7085:
7086: @example
7087: : def-word
7088: create-interpret/compile
7089: @i{code1}
7090: interpretation>
7091: @i{code2}
7092: <interpretation
7093: compilation>
7094: @i{code3}
7095: <compilation ;
7096: @end example
7097:
7098: For a @i{word} defined with @code{def-word}, the interpretation
7099: semantics are to push the address of the body of @i{word} and perform
7100: @i{code2}, and the compilation semantics are to push the address of
7101: the body of @i{word} and perform @i{code3}. E.g., @code{constant}
7102: can also be defined like this (except that the defined constants don't
7103: behave correctly when @code{[compile]}d):
7104:
7105: @example
7106: : constant ( n "name" -- )
7107: create-interpret/compile
7108: ,
7109: interpretation> ( -- n )
7110: @@
7111: <interpretation
7112: compilation> ( compilation. -- ; run-time. -- n )
7113: @@ postpone literal
7114: <compilation ;
7115: @end example
7116:
7117:
7118: doc-create-interpret/compile
7119: doc-interpretation>
7120: doc-<interpretation
7121: doc-compilation>
7122: doc-<compilation
7123:
7124:
7125: Words defined with @code{interpret/compile:} and
7126: @code{create-interpret/compile} have an extended header structure that
7127: differs from other words; however, unless you try to access them with
7128: plain address arithmetic, you should not notice this. Words for
7129: accessing the header structure usually know how to deal with this; e.g.,
7130: @code{'} @i{word} @code{>body} also gives you the body of a word created
7131: with @code{create-interpret/compile}.
7132:
7133:
7134: @c -------------------------------------------------------------
7135: @node Tokens for Words, Compiling words, Interpretation and Compilation Semantics, Words
7136: @section Tokens for Words
7137: @cindex tokens for words
7138:
7139: This section describes the creation and use of tokens that represent
7140: words.
7141:
7142: @menu
7143: * Execution token:: represents execution/interpretation semantics
7144: * Compilation token:: represents compilation semantics
7145: * Name token:: represents named words
7146: @end menu
7147:
7148: @node Execution token, Compilation token, Tokens for Words, Tokens for Words
7149: @subsection Execution token
7150:
7151: @cindex xt
7152: @cindex execution token
7153: An @dfn{execution token} (@i{XT}) represents some behaviour of a word.
7154: You can use @code{execute} to invoke this behaviour.
7155:
7156: @cindex tick (')
7157: You can use @code{'} to get an execution token that represents the
7158: interpretation semantics of a named word:
7159:
7160: @example
7161: 5 ' .
7162: execute
7163: @end example
7164:
7165: doc-'
7166:
7167: @code{'} parses at run-time; there is also a word @code{[']} that parses
7168: when it is compiled, and compiles the resulting XT:
7169:
7170: @example
7171: : foo ['] . execute ;
7172: 5 foo
7173: : bar ' execute ; \ by contrast,
7174: 5 bar . \ ' parses "." when bar executes
7175: @end example
7176:
7177: doc-[']
7178:
7179: If you want the execution token of @i{word}, write @code{['] @i{word}}
7180: in compiled code and @code{' @i{word}} in interpreted code. Gforth's
7181: @code{'} and @code{[']} behave somewhat unusually by complaining about
7182: compile-only words (because these words have no interpretation
7183: semantics). You might get what you want by using @code{COMP' @i{word}
7184: DROP} or @code{[COMP'] @i{word} DROP} (for details @pxref{Compilation
7185: token}).
7186:
7187: Another way to get an XT is @code{:noname} or @code{lastxt}
7188: (@pxref{Anonymous Definitions}). For anonymous words this gives an xt
7189: for the only behaviour the word has (the execution semantics). For
7190: named words, @code{lastxt} produces an XT for the same behaviour it
7191: would produce if the word was defined anonymously.
7192:
7193: @example
7194: :noname ." hello" ;
7195: execute
7196: @end example
7197:
7198: An XT occupies one cell and can be manipulated like any other cell.
7199:
7200: @cindex code field address
7201: @cindex CFA
7202: In ANS Forth the XT is just an abstract data type (i.e., defined by the
7203: operations that produce or consume it). For old hands: In Gforth, the
7204: XT is implemented as a code field address (CFA).
7205:
7206: doc-execute
7207: doc-perform
7208:
7209: @node Compilation token, Name token, Execution token, Tokens for Words
7210: @subsection Compilation token
7211:
7212: @cindex compilation token
7213: @cindex CT (compilation token)
7214: Gforth represents the compilation semantics of a named word by a
7215: @dfn{compilation token} consisting of two cells: @i{w xt}. The top cell
7216: @i{xt} is an execution token. The compilation semantics represented by
7217: the compilation token can be performed with @code{execute}, which
7218: consumes the whole compilation token, with an additional stack effect
7219: determined by the represented compilation semantics.
7220:
7221: At present, the @i{w} part of a compilation token is an execution token,
7222: and the @i{xt} part represents either @code{execute} or
7223: @code{compile,}@footnote{Depending upon the compilation semantics of the
7224: word. If the word has default compilation semantics, the @i{xt} will
7225: represent @code{compile,}. Otherwise (e.g., for immediate words), the
7226: @i{xt} will represent @code{execute}.}. However, don't rely on that
7227: knowledge, unless necessary; future versions of Gforth may introduce
7228: unusual compilation tokens (e.g., a compilation token that represents
7229: the compilation semantics of a literal).
7230:
7231: You can perform the compilation semantics represented by the compilation
7232: token with @code{execute}. You can compile the compilation semantics
7233: with @code{postpone,}. I.e., @code{COMP' @i{word} postpone,} is
7234: equivalent to @code{postpone @i{word}}.
7235:
7236: doc-[comp']
7237: doc-comp'
7238: doc-postpone,
7239:
7240: @node Name token, , Compilation token, Tokens for Words
7241: @subsection Name token
7242:
7243: @cindex name token
7244: @cindex name field address
7245: @cindex NFA
7246: Gforth represents named words by the @dfn{name token}, (@i{nt}). In
7247: Gforth, the abstract data type @emph{name token} is implemented as a
7248: name field address (NFA).
7249:
7250: doc-find-name
7251: doc-name>int
7252: doc-name?int
7253: doc-name>comp
7254: doc-name>string
7255:
7256: @c ----------------------------------------------------------
7257: @node Compiling words, The Text Interpreter, Tokens for Words, Words
7258: @section Compiling words
7259: @cindex compiling words
7260: @cindex macros
7261:
7262: In contrast to most other languages, Forth has no strict boundary
7263: between compilation and run-time. E.g., you can run arbitrary code
7264: between defining words (or for computing data used by defining words
7265: like @code{constant}). Moreover, @code{Immediate} (@pxref{Interpretation
7266: and Compilation Semantics} and @code{[}...@code{]} (see below) allow
7267: running arbitrary code while compiling a colon definition (exception:
7268: you must not allot dictionary space).
7269:
7270: @menu
7271: * Literals:: Compiling data values
7272: * Macros:: Compiling words
7273: @end menu
7274:
7275: @node Literals, Macros, Compiling words, Compiling words
7276: @subsection Literals
7277: @cindex Literals
7278:
7279: The simplest and most frequent example is to compute a literal during
7280: compilation. E.g., the following definition prints an array of strings,
7281: one string per line:
7282:
7283: @example
7284: : .strings ( addr u -- ) \ gforth
7285: 2* cells bounds U+DO
7286: cr i 2@@ type
7287: 2 cells +LOOP ;
7288: @end example
7289:
7290: With a simple-minded compiler like Gforth's, this computes @code{2
7291: cells} on every loop iteration. You can compute this value once and for
7292: all at compile time and compile it into the definition like this:
7293:
7294: @example
7295: : .strings ( addr u -- ) \ gforth
7296: 2* cells bounds U+DO
7297: cr i 2@@ type
7298: [ 2 cells ] literal +LOOP ;
7299: @end example
7300:
7301: @code{[} switches the text interpreter to interpret state (you will get
7302: an @code{ok} prompt if you type this example interactively and insert a
7303: newline between @code{[} and @code{]}), so it performs the
7304: interpretation semantics of @code{2 cells}; this computes a number.
7305: @code{]} switches the text interpreter back into compile state. It then
7306: performs @code{Literal}'s compilation semantics, which are to compile
7307: this number into the current word. You can decompile the word with
7308: @code{see .strings} to see the effect on the compiled code.
7309:
7310: You can also optimize the @code{2* cells} into @code{[ 2 cells ] literal
7311: *} in this way.
7312:
7313: doc-[
7314: doc-]
7315: doc-literal
7316: doc-]L
7317:
7318: There are also words for compiling other data types than single cells as
7319: literals:
7320:
7321: doc-2literal
7322: doc-fliteral
7323: doc-sliteral
7324:
7325: @cindex colon-sys, passing data across @code{:}
7326: @cindex @code{:}, passing data across
7327: You might be tempted to pass data from outside a colon definition to the
7328: inside on the data stack. This does not work, because @code{:} puhes a
7329: colon-sys, making stuff below unaccessible. E.g., this does not work:
7330:
7331: @example
7332: 5 : foo literal ; \ error: "unstructured"
7333: @end example
7334:
7335: Instead, you have to pass the value in some other way, e.g., through a
7336: variable:
7337:
7338: @example
7339: variable temp
7340: 5 temp !
7341: : foo [ temp @@ ] literal ;
7342: @end example
7343:
7344:
7345: @node Macros, , Literals, Compiling words
7346: @subsection Macros
7347: @cindex Macros
7348: @cindex compiling compilation semantics
7349:
7350: @code{Literal} and friends compile data values into the current
7351: definition. You can also write words that compile other words into the
7352: current definition. E.g.,
7353:
7354: @example
7355: : compile-+ ( -- ) \ compiled code: ( n1 n2 -- n )
7356: POSTPONE + ;
7357:
7358: : foo ( n1 n2 -- n )
7359: [ compile-+ ] ;
7360: 1 2 foo .
7361: @end example
7362:
7363: This is equivalent to @code{: foo + ;} (@code{see foo} to check this).
7364: What happens in this example? @code{Postpone} compiles the compilation
7365: semantics of @code{+} into @code{compile-+}; later the text interpreter
7366: executes @code{compile-+} and thus the compilation semantics of +, which
7367: compile (the execution semantics of) @code{+} into
7368: @code{foo}.@footnote{A recent RFI answer requires that compiling words
7369: should only be executed in compile state, so this example is not
7370: guaranteed to work on all standard systems, but on any decent system it
7371: will work.}
7372:
7373: doc-postpone
7374: doc-[compile]
7375:
7376: Compiling words like @code{compile-+} are usually immediate (or similar)
7377: so you do not have to switch to interpret state to execute them;
7378: mopifying the last example accordingly produces:
7379:
7380: @example
7381: : [compile-+] ( compilation: --; interpretation: -- )
7382: \ compiled code: ( n1 n2 -- n )
7383: POSTPONE + ; immediate
7384:
7385: : foo ( n1 n2 -- n )
7386: [compile-+] ;
7387: 1 2 foo .
7388: @end example
7389:
7390: Immediate compiling words are similar to macros in other languages (in
7391: particular, Lisp). The important differences to macros in, e.g., C are:
7392:
7393: @itemize @bullet
7394:
7395: @item
7396: You use the same language for defining and processing macros, not a
7397: separate preprocessing language and processor.
7398:
7399: @item
7400: Consequently, the full power of Forth is available in macro definitions.
7401: E.g., you can perform arbitrarily complex computations, or generate
7402: different code conditionally or in a loop (e.g., @pxref{Advanced macros
7403: Tutorial}). This power is very useful when writing a parser generators
7404: or other code-generating software.
7405:
7406: @item
7407: Macros defined using @code{postpone} etc. deal with the language at a
7408: higher level than strings; name binding happens at macro definition
7409: time, so you can avoid the pitfalls of name collisions that can happen
7410: in C macros. Of course, Forth is a liberal language and also allows to
7411: shoot yourself in the foot with text-interpreted macros like
7412:
7413: @example
7414: : [compile-+] s" +" evaluate ; immediate
7415: @end example
7416:
7417: Apart from binding the name at macro use time, using @code{evaluate}
7418: also makes your definition @code{state}-smart (@pxref{state-smartness}).
7419: @end itemize
7420:
7421: You may want the macro to compile a number into a word. The word to do
7422: it is @code{literal}, but you have to @code{postpone} it, so its
7423: compilation semantics take effect when the macro is executed, not when
7424: it is compiled:
7425:
7426: @example
7427: : [compile-5] ( -- ) \ compiled code: ( -- n )
7428: 5 POSTPONE literal ; immediate
7429:
7430: : foo [compile-5] ;
7431: foo .
7432: @end example
7433:
7434: You may want to pass parameters to a macro, that the macro should
7435: compile into the current definition. If the parameter is a number, then
7436: you can use @code{postpone literal} (similar for other values).
7437:
7438: If you want to pass a word that is to be compiled, the usual way is to
7439: pass an execution token and @code{compile,} it:
7440:
7441: @example
7442: : twice1 ( xt -- ) \ compiled code: ... -- ...
7443: dup compile, compile, ;
7444:
7445: : 2+ ( n1 -- n2 )
7446: [ ' 1+ twice1 ] ;
7447: @end example
7448:
7449: doc-compile,
7450:
7451: An alternative available in Gforth, that allows you to pass compile-only
7452: words as parameters is to use the compilation token (@pxref{Compilation
7453: token}). The same example in this technique:
7454:
7455: @example
7456: : twice ( ... ct -- ... ) \ compiled code: ... -- ...
7457: 2dup 2>r execute 2r> execute ;
7458:
7459: : 2+ ( n1 -- n2 )
7460: [ comp' 1+ twice ] ;
7461: @end example
7462:
7463: In the example above @code{2>r} and @code{2r>} ensure that @code{twice}
7464: works even if the executed compilation semantics has an effect on the
7465: data stack.
7466:
7467: You can also define complete definitions with these words; this provides
7468: an alternative to using @code{does>} (@pxref{User-defined Defining
7469: Words}). E.g., instead of
7470:
7471: @example
7472: : curry+ ( n1 "name" -- )
7473: CREATE ,
7474: DOES> ( n2 -- n1+n2 )
7475: @@ + ;
7476: @end example
7477:
7478: you could define
7479:
7480: @example
7481: : curry+ ( n1 "name" -- )
7482: \ name execution: ( n2 -- n1+n2 )
7483: >r : r> POSTPONE literal POSTPONE + POSTPONE ; ;
7484:
7485: -3 curry+ 3-
7486: see 3-
7487: @end example
7488:
7489: The sequence @code{>r : r>} is necessary, because @code{:} puts a
7490: colon-sys on the data stack that makes everything below it unaccessible.
7491:
7492: This way of writing defining words is sometimes more, sometimes less
7493: convenient than using @code{does>} (@pxref{Advanced does> usage
7494: example}). One advantage of this method is that it can be optimized
7495: better, because the compiler knows that the value compiled with
7496: @code{literal} is fixed, whereas the data associated with a
7497: @code{create}d word can be changed.
7498:
7499: @c ----------------------------------------------------------
7500: @node The Text Interpreter, Word Lists, Compiling words, Words
7501: @section The Text Interpreter
7502: @cindex interpreter - outer
7503: @cindex text interpreter
7504: @cindex outer interpreter
7505:
7506: @c Should we really describe all these ugly details? IMO the text
7507: @c interpreter should be much cleaner, but that may not be possible within
7508: @c ANS Forth. - anton
7509: @c nac-> I wanted to explain how it works to show how you can exploit
7510: @c it in your own programs. When I was writing a cross-compiler, figuring out
7511: @c some of these gory details was very helpful to me. None of the textbooks
7512: @c I've seen cover it, and the most modern Forth textbook -- Forth Inc's,
7513: @c seems to positively avoid going into too much detail for some of
7514: @c the internals.
7515:
7516: @c anton: ok. I wonder, though, if this is the right place; for some stuff
7517: @c it is; for the ugly details, I would prefer another place. I wonder
7518: @c whether we should have a chapter before "Words" that describes some
7519: @c basic concepts referred to in words, and a chapter after "Words" that
7520: @c describes implementation details.
7521:
7522: The text interpreter@footnote{This is an expanded version of the
7523: material in @ref{Introducing the Text Interpreter}.} is an endless loop
7524: that processes input from the current input device. It is also called
7525: the outer interpreter, in contrast to the inner interpreter
7526: (@pxref{Engine}) which executes the compiled Forth code on interpretive
7527: implementations.
7528:
7529: @cindex interpret state
7530: @cindex compile state
7531: The text interpreter operates in one of two states: @dfn{interpret
7532: state} and @dfn{compile state}. The current state is defined by the
7533: aptly-named variable @code{state}.
7534:
7535: This section starts by describing how the text interpreter behaves when
7536: it is in interpret state, processing input from the user input device --
7537: the keyboard. This is the mode that a Forth system is in after it starts
7538: up.
7539:
7540: @cindex input buffer
7541: @cindex terminal input buffer
7542: The text interpreter works from an area of memory called the @dfn{input
7543: buffer}@footnote{When the text interpreter is processing input from the
7544: keyboard, this area of memory is called the @dfn{terminal input buffer}
7545: (TIB) and is addressed by the (obsolescent) words @code{TIB} and
7546: @code{#TIB}.}, which stores your keyboard input when you press the
7547: @key{RET} key. Starting at the beginning of the input buffer, it skips
7548: leading spaces (called @dfn{delimiters}) then parses a string (a
7549: sequence of non-space characters) until it reaches either a space
7550: character or the end of the buffer. Having parsed a string, it makes two
7551: attempts to process it:
7552:
7553: @cindex dictionary
7554: @itemize @bullet
7555: @item
7556: It looks for the string in a @dfn{dictionary} of definitions. If the
7557: string is found, the string names a @dfn{definition} (also known as a
7558: @dfn{word}) and the dictionary search returns information that allows
7559: the text interpreter to perform the word's @dfn{interpretation
7560: semantics}. In most cases, this simply means that the word will be
7561: executed.
7562: @item
7563: If the string is not found in the dictionary, the text interpreter
7564: attempts to treat it as a number, using the rules described in
7565: @ref{Number Conversion}. If the string represents a legal number in the
7566: current radix, the number is pushed onto a parameter stack (the data
7567: stack for integers, the floating-point stack for floating-point
7568: numbers).
7569: @end itemize
7570:
7571: If both attempts fail, or if the word is found in the dictionary but has
7572: no interpretation semantics@footnote{This happens if the word was
7573: defined as @code{COMPILE-ONLY}.} the text interpreter discards the
7574: remainder of the input buffer, issues an error message and waits for
7575: more input. If one of the attempts succeeds, the text interpreter
7576: repeats the parsing process until the whole of the input buffer has been
7577: processed, at which point it prints the status message ``@code{ ok}''
7578: and waits for more input.
7579:
7580: @c anton: this should be in the input stream subsection (or below it)
7581:
7582: @cindex parse area
7583: The text interpreter keeps track of its position in the input buffer by
7584: updating a variable called @code{>IN} (pronounced ``to-in''). The value
7585: of @code{>IN} starts out as 0, indicating an offset of 0 from the start
7586: of the input buffer. The region from offset @code{>IN @@} to the end of
7587: the input buffer is called the @dfn{parse area}@footnote{In other words,
7588: the text interpreter processes the contents of the input buffer by
7589: parsing strings from the parse area until the parse area is empty.}.
7590: This example shows how @code{>IN} changes as the text interpreter parses
7591: the input buffer:
7592:
7593: @example
7594: : remaining >IN @@ SOURCE 2 PICK - -ROT + SWAP
7595: CR ." ->" TYPE ." <-" ; IMMEDIATE
7596:
7597: 1 2 3 remaining + remaining .
7598:
7599: : foo 1 2 3 remaining SWAP remaining ;
7600: @end example
7601:
7602: @noindent
7603: The result is:
7604:
7605: @example
7606: ->+ remaining .<-
7607: ->.<-5 ok
7608:
7609: ->SWAP remaining ;-<
7610: ->;<- ok
7611: @end example
7612:
7613: @cindex parsing words
7614: The value of @code{>IN} can also be modified by a word in the input
7615: buffer that is executed by the text interpreter. This means that a word
7616: can ``trick'' the text interpreter into either skipping a section of the
7617: input buffer@footnote{This is how parsing words work.} or into parsing a
7618: section twice. For example:
7619:
7620: @example
7621: : lat ." <<foo>>" ;
7622: : flat ." <<bar>>" >IN DUP @@ 3 - SWAP ! ;
7623: @end example
7624:
7625: @noindent
7626: When @code{flat} is executed, this output is produced@footnote{Exercise
7627: for the reader: what would happen if the @code{3} were replaced with
7628: @code{4}?}:
7629:
7630: @example
7631: <<bar>><<foo>>
7632: @end example
7633:
7634: This technique can be used to work around some of the interoperability
7635: problems of parsing words. Of course, it's better to avoid parsing
7636: words where possible.
7637:
7638: @noindent
7639: Two important notes about the behaviour of the text interpreter:
7640:
7641: @itemize @bullet
7642: @item
7643: It processes each input string to completion before parsing additional
7644: characters from the input buffer.
7645: @item
7646: It treats the input buffer as a read-only region (and so must your code).
7647: @end itemize
7648:
7649: @noindent
7650: When the text interpreter is in compile state, its behaviour changes in
7651: these ways:
7652:
7653: @itemize @bullet
7654: @item
7655: If a parsed string is found in the dictionary, the text interpreter will
7656: perform the word's @dfn{compilation semantics}. In most cases, this
7657: simply means that the execution semantics of the word will be appended
7658: to the current definition.
7659: @item
7660: When a number is encountered, it is compiled into the current definition
7661: (as a literal) rather than being pushed onto a parameter stack.
7662: @item
7663: If an error occurs, @code{state} is modified to put the text interpreter
7664: back into interpret state.
7665: @item
7666: Each time a line is entered from the keyboard, Gforth prints
7667: ``@code{ compiled}'' rather than `` @code{ok}''.
7668: @end itemize
7669:
7670: @cindex text interpreter - input sources
7671: When the text interpreter is using an input device other than the
7672: keyboard, its behaviour changes in these ways:
7673:
7674: @itemize @bullet
7675: @item
7676: When the parse area is empty, the text interpreter attempts to refill
7677: the input buffer from the input source. When the input source is
7678: exhausted, the input source is set back to the previous input source.
7679: @item
7680: It doesn't print out ``@code{ ok}'' or ``@code{ compiled}'' messages each
7681: time the parse area is emptied.
7682: @item
7683: If an error occurs, the input source is set back to the user input
7684: device.
7685: @end itemize
7686:
7687: You can read about this in more detail in @ref{Input Sources}.
7688:
7689: doc->in
7690: doc-source
7691:
7692: doc-tib
7693: doc-#tib
7694:
7695:
7696: @menu
7697: * Input Sources::
7698: * Number Conversion::
7699: * Interpret/Compile states::
7700: * Interpreter Directives::
7701: @end menu
7702:
7703: @node Input Sources, Number Conversion, The Text Interpreter, The Text Interpreter
7704: @subsection Input Sources
7705: @cindex input sources
7706: @cindex text interpreter - input sources
7707:
7708: By default, the text interpreter processes input from the user input
7709: device (the keyboard) when Forth starts up. The text interpreter can
7710: process input from any of these sources:
7711:
7712: @itemize @bullet
7713: @item
7714: The user input device -- the keyboard.
7715: @item
7716: A file, using the words described in @ref{Forth source files}.
7717: @item
7718: A block, using the words described in @ref{Blocks}.
7719: @item
7720: A text string, using @code{evaluate}.
7721: @end itemize
7722:
7723: A program can identify the current input device from the values of
7724: @code{source-id} and @code{blk}.
7725:
7726:
7727: doc-source-id
7728: doc-blk
7729:
7730: doc-save-input
7731: doc-restore-input
7732:
7733: doc-evaluate
7734:
7735:
7736:
7737: @node Number Conversion, Interpret/Compile states, Input Sources, The Text Interpreter
7738: @subsection Number Conversion
7739: @cindex number conversion
7740: @cindex double-cell numbers, input format
7741: @cindex input format for double-cell numbers
7742: @cindex single-cell numbers, input format
7743: @cindex input format for single-cell numbers
7744: @cindex floating-point numbers, input format
7745: @cindex input format for floating-point numbers
7746:
7747: This section describes the rules that the text interpreter uses when it
7748: tries to convert a string into a number.
7749:
7750: Let <digit> represent any character that is a legal digit in the current
7751: number base@footnote{For example, 0-9 when the number base is decimal or
7752: 0-9, A-F when the number base is hexadecimal.}.
7753:
7754: Let <decimal digit> represent any character in the range 0-9.
7755:
7756: Let @{@i{a b}@} represent the @i{optional} presence of any of the characters
7757: in the braces (@i{a} or @i{b} or neither).
7758:
7759: Let * represent any number of instances of the previous character
7760: (including none).
7761:
7762: Let any other character represent itself.
7763:
7764: @noindent
7765: Now, the conversion rules are:
7766:
7767: @itemize @bullet
7768: @item
7769: A string of the form <digit><digit>* is treated as a single-precision
7770: (cell-sized) positive integer. Examples are 0 123 6784532 32343212343456 42
7771: @item
7772: A string of the form -<digit><digit>* is treated as a single-precision
7773: (cell-sized) negative integer, and is represented using 2's-complement
7774: arithmetic. Examples are -45 -5681 -0
7775: @item
7776: A string of the form <digit><digit>*.<digit>* is treated as a double-precision
7777: (double-cell-sized) positive integer. Examples are 3465. 3.465 34.65
7778: (all three of these represent the same number).
7779: @item
7780: A string of the form -<digit><digit>*.<digit>* is treated as a
7781: double-precision (double-cell-sized) negative integer, and is
7782: represented using 2's-complement arithmetic. Examples are -3465. -3.465
7783: -34.65 (all three of these represent the same number).
7784: @item
7785: A string of the form @{+ -@}<decimal digit>@{.@}<decimal digit>*@{e
7786: E@}@{+ -@}<decimal digit><decimal digit>* is treated as a floating-point
7787: number. Examples are 1e 1e0 1.e 1.e0 +1e+0 (which all represent the same
7788: number) +12.E-4
7789: @end itemize
7790:
7791: By default, the number base used for integer number conversion is given
7792: by the contents of the variable @code{base}. Note that a lot of
7793: confusion can result from unexpected values of @code{base}. If you
7794: change @code{base} anywhere, make sure to save the old value and restore
7795: it afterwards. In general I recommend keeping @code{base} decimal, and
7796: using the prefixes described below for the popular non-decimal bases.
7797:
7798: doc-dpl
7799: doc-base
7800: doc-hex
7801: doc-decimal
7802:
7803:
7804: @cindex '-prefix for character strings
7805: @cindex &-prefix for decimal numbers
7806: @cindex %-prefix for binary numbers
7807: @cindex $-prefix for hexadecimal numbers
7808: Gforth allows you to override the value of @code{base} by using a
7809: prefix@footnote{Some Forth implementations provide a similar scheme by
7810: implementing @code{$} etc. as parsing words that process the subsequent
7811: number in the input stream and push it onto the stack. For example, see
7812: @cite{Number Conversion and Literals}, by Wil Baden; Forth Dimensions
7813: 20(3) pages 26--27. In such implementations, unlike in Gforth, a space
7814: is required between the prefix and the number.} before the first digit
7815: of an (integer) number. Four prefixes are supported:
7816:
7817: @itemize @bullet
7818: @item
7819: @code{&} -- decimal
7820: @item
7821: @code{%} -- binary
7822: @item
7823: @code{$} -- hexadecimal
7824: @item
7825: @code{'} -- base @code{max-char+1}
7826: @end itemize
7827:
7828: Here are some examples, with the equivalent decimal number shown after
7829: in braces:
7830:
7831: -$41 (-65), %1001101 (205), %1001.0001 (145 - a double-precision number),
7832: 'AB (16706; ascii A is 65, ascii B is 66, number is 65*256 + 66),
7833: 'ab (24930; ascii a is 97, ascii B is 98, number is 97*256 + 98),
7834: &905 (905), $abc (2478), $ABC (2478).
7835:
7836: @cindex number conversion - traps for the unwary
7837: @noindent
7838: Number conversion has a number of traps for the unwary:
7839:
7840: @itemize @bullet
7841: @item
7842: You cannot determine the current number base using the code sequence
7843: @code{base @@ .} -- the number base is always 10 in the current number
7844: base. Instead, use something like @code{base @@ dec.}
7845: @item
7846: If the number base is set to a value greater than 14 (for example,
7847: hexadecimal), the number 123E4 is ambiguous; the conversion rules allow
7848: it to be intepreted as either a single-precision integer or a
7849: floating-point number (Gforth treats it as an integer). The ambiguity
7850: can be resolved by explicitly stating the sign of the mantissa and/or
7851: exponent: 123E+4 or +123E4 -- if the number base is decimal, no
7852: ambiguity arises; either representation will be treated as a
7853: floating-point number.
7854: @item
7855: There is a word @code{bin} but it does @i{not} set the number base!
7856: It is used to specify file types.
7857: @item
7858: ANS Forth requires the @code{.} of a double-precision number to be the
7859: final character in the string. Gforth allows the @code{.} to be
7860: anywhere after the first digit.
7861: @item
7862: The number conversion process does not check for overflow.
7863: @item
7864: In an ANS Forth program @code{base} is required to be decimal when
7865: converting floating-point numbers. In Gforth, number conversion to
7866: floating-point numbers always uses base &10, irrespective of the value
7867: of @code{base}.
7868: @end itemize
7869:
7870: You can read numbers into your programs with the words described in
7871: @ref{Input}.
7872:
7873: @node Interpret/Compile states, Interpreter Directives, Number Conversion, The Text Interpreter
7874: @subsection Interpret/Compile states
7875: @cindex Interpret/Compile states
7876:
7877: A standard program is not permitted to change @code{state}
7878: explicitly. However, it can change @code{state} implicitly, using the
7879: words @code{[} and @code{]}. When @code{[} is executed it switches
7880: @code{state} to interpret state, and therefore the text interpreter
7881: starts interpreting. When @code{]} is executed it switches @code{state}
7882: to compile state and therefore the text interpreter starts
7883: compiling. The most common usage for these words is for switching into
7884: interpret state and back from within a colon definition; this technique
7885: can be used to compile a literal (for an example, @pxref{Literals}) or
7886: for conditional compilation (for an example, @pxref{Interpreter
7887: Directives}).
7888:
7889:
7890: @c This is a bad example: It's non-standard, and it's not necessary.
7891: @c However, I can't think of a good example for switching into compile
7892: @c state when there is no current word (@code{state}-smart words are not a
7893: @c good reason). So maybe we should use an example for switching into
7894: @c interpret @code{state} in a colon def. - anton
7895: @c nac-> I agree. I started out by putting in the example, then realised
7896: @c that it was non-ANS, so wrote more words around it. I hope this
7897: @c re-written version is acceptable to you. I do want to keep the example
7898: @c as it is helpful for showing what is and what is not portable, particularly
7899: @c where it outlaws a style in common use.
7900:
7901: @c anton: it's more important to show what's portable. After we have done
7902: @c that, we can also show what's not. In any case, I have written a
7903: @c section Compiling Words which also deals with [ ].
7904:
7905: @code{[} and @code{]} also give you the ability to switch into compile
7906: state and back, but we cannot think of any useful Standard application
7907: for this ability. Pre-ANS Forth textbooks have examples like this:
7908:
7909: @example
7910: : AA ." this is A" ;
7911: : BB ." this is B" ;
7912: : CC ." this is C" ;
7913:
7914: create table ] aa bb cc [
7915:
7916: : go ( n -- ) \ n is offset into table.. 0 for 1st entry
7917: cells table + @ execute ;
7918: @end example
7919:
7920: This example builds a jump table; @code{0 go} will display ``@code{this
7921: is A}''. Using @code{[} and @code{]} in this example is equivalent to
7922: defining @code{table} like this:
7923:
7924: @example
7925: create table ' aa COMPILE, ' bb COMPILE, ' cc COMPILE,
7926: @end example
7927:
7928: The problem with this code is that the definition of @code{table} is not
7929: portable -- it @i{compile}s execution tokens into code space. Whilst it
7930: @i{may} work on systems where code space and data space co-incide, the
7931: Standard only allows data space to be assigned for a @code{CREATE}d
7932: word. In addition, the Standard only allows @code{@@} to access data
7933: space, whilst this example is using it to access code space. The only
7934: portable, Standard way to build this table is to build it in data space,
7935: like this:
7936:
7937: @example
7938: create table ' aa , ' bb , ' cc ,
7939: @end example
7940:
7941: doc-state
7942:
7943:
7944: @node Interpreter Directives, , Interpret/Compile states, The Text Interpreter
7945: @subsection Interpreter Directives
7946: @cindex interpreter directives
7947: @cindex conditional compilation
7948:
7949: These words are usually used in interpret state; typically to control
7950: which parts of a source file are processed by the text
7951: interpreter. There are only a few ANS Forth Standard words, but Gforth
7952: supplements these with a rich set of immediate control structure words
7953: to compensate for the fact that the non-immediate versions can only be
7954: used in compile state (@pxref{Control Structures}). Typical usages:
7955:
7956: @example
7957: FALSE Constant HAVE-ASSEMBLER
7958: .
7959: .
7960: HAVE-ASSEMBLER [IF]
7961: : ASSEMBLER-FEATURE
7962: ...
7963: ;
7964: [ENDIF]
7965: .
7966: .
7967: : SEE
7968: ... \ general-purpose SEE code
7969: [ HAVE-ASSEMBLER [IF] ]
7970: ... \ assembler-specific SEE code
7971: [ [ENDIF] ]
7972: ;
7973: @end example
7974:
7975:
7976: doc-[IF]
7977: doc-[ELSE]
7978: doc-[THEN]
7979: doc-[ENDIF]
7980:
7981: doc-[IFDEF]
7982: doc-[IFUNDEF]
7983:
7984: doc-[?DO]
7985: doc-[DO]
7986: doc-[FOR]
7987: doc-[LOOP]
7988: doc-[+LOOP]
7989: doc-[NEXT]
7990:
7991: doc-[BEGIN]
7992: doc-[UNTIL]
7993: doc-[AGAIN]
7994: doc-[WHILE]
7995: doc-[REPEAT]
7996:
7997:
7998: @c -------------------------------------------------------------
7999: @node Word Lists, Environmental Queries, The Text Interpreter, Words
8000: @section Word Lists
8001: @cindex word lists
8002: @cindex header space
8003:
8004: A wordlist is a list of named words; you can add new words and look up
8005: words by name (and you can remove words in a restricted way with
8006: markers). Every named (and @code{reveal}ed) word is in one wordlist.
8007:
8008: @cindex search order stack
8009: The text interpreter searches the wordlists present in the search order
8010: (a stack of wordlists), from the top to the bottom. Within each
8011: wordlist, the search starts conceptually at the newest word; i.e., if
8012: two words in a wordlist have the same name, the newer word is found.
8013:
8014: @cindex compilation word list
8015: New words are added to the @dfn{compilation wordlist} (aka current
8016: wordlist).
8017:
8018: @cindex wid
8019: A word list is identified by a cell-sized word list identifier (@i{wid})
8020: in much the same way as a file is identified by a file handle. The
8021: numerical value of the wid has no (portable) meaning, and might change
8022: from session to session.
8023:
8024: The ANS Forth ``Search order'' word set is intended to provide a set of
8025: low-level tools that allow various different schemes to be
8026: implemented. Gforth also provides @code{vocabulary}, a traditional Forth
8027: word. @file{compat/vocabulary.fs} provides an implementation in ANS
8028: Forth.
8029:
8030: @comment TODO: locals section refers to here, saying that every word list (aka
8031: @comment vocabulary) has its own methods for searching etc. Need to document that.
8032: @c anton: but better in a separate subsection on wordlist internals
8033:
8034: @comment TODO: document markers, reveal, tables, mappedwordlist
8035:
8036: @comment the gforthman- prefix is used to pick out the true definition of a
8037: @comment word from the source files, rather than some alias.
8038:
8039: doc-forth-wordlist
8040: doc-definitions
8041: doc-get-current
8042: doc-set-current
8043: doc-get-order
8044: doc---gforthman-set-order
8045: doc-wordlist
8046: doc-table
8047: doc->order
8048: doc-previous
8049: doc-also
8050: doc---gforthman-forth
8051: doc-only
8052: doc---gforthman-order
8053:
8054: doc-find
8055: doc-search-wordlist
8056:
8057: doc-words
8058: doc-vlist
8059: @c doc-words-deferred
8060:
8061: @c doc-mappedwordlist @c map-structure undefined, implemantation-specific
8062: doc-root
8063: doc-vocabulary
8064: doc-seal
8065: doc-vocs
8066: doc-current
8067: doc-context
8068:
8069:
8070: @menu
8071: * Vocabularies::
8072: * Why use word lists?::
8073: * Word list example::
8074: @end menu
8075:
8076: @node Vocabularies, Why use word lists?, Word Lists, Word Lists
8077: @subsection Vocabularies
8078: @cindex Vocabularies, detailed explanation
8079:
8080: Here is an example of creating and using a new wordlist using ANS
8081: Forth words:
8082:
8083: @example
8084: wordlist constant my-new-words-wordlist
8085: : my-new-words get-order nip my-new-words-wordlist swap set-order ;
8086:
8087: \ add it to the search order
8088: also my-new-words
8089:
8090: \ alternatively, add it to the search order and make it
8091: \ the compilation word list
8092: also my-new-words definitions
8093: \ type "order" to see the problem
8094: @end example
8095:
8096: The problem with this example is that @code{order} has no way to
8097: associate the name @code{my-new-words} with the wid of the word list (in
8098: Gforth, @code{order} and @code{vocs} will display @code{???} for a wid
8099: that has no associated name). There is no Standard way of associating a
8100: name with a wid.
8101:
8102: In Gforth, this example can be re-coded using @code{vocabulary}, which
8103: associates a name with a wid:
8104:
8105: @example
8106: vocabulary my-new-words
8107:
8108: \ add it to the search order
8109: also my-new-words
8110:
8111: \ alternatively, add it to the search order and make it
8112: \ the compilation word list
8113: my-new-words definitions
8114: \ type "order" to see that the problem is solved
8115: @end example
8116:
8117:
8118: @node Why use word lists?, Word list example, Vocabularies, Word Lists
8119: @subsection Why use word lists?
8120: @cindex word lists - why use them?
8121:
8122: Here are some reasons why people use wordlists:
8123:
8124: @itemize @bullet
8125:
8126: @c anton: Gforth's hashing implementation makes the search speed
8127: @c independent from the number of words. But it is linear with the number
8128: @c of wordlists that have to be searched, so in effect using more wordlists
8129: @c actually slows down compilation.
8130:
8131: @c @item
8132: @c To improve compilation speed by reducing the number of header space
8133: @c entries that must be searched. This is achieved by creating a new
8134: @c word list that contains all of the definitions that are used in the
8135: @c definition of a Forth system but which would not usually be used by
8136: @c programs running on that system. That word list would be on the search
8137: @c list when the Forth system was compiled but would be removed from the
8138: @c search list for normal operation. This can be a useful technique for
8139: @c low-performance systems (for example, 8-bit processors in embedded
8140: @c systems) but is unlikely to be necessary in high-performance desktop
8141: @c systems.
8142:
8143: @item
8144: To prevent a set of words from being used outside the context in which
8145: they are valid. Two classic examples of this are an integrated editor
8146: (all of the edit commands are defined in a separate word list; the
8147: search order is set to the editor word list when the editor is invoked;
8148: the old search order is restored when the editor is terminated) and an
8149: integrated assembler (the op-codes for the machine are defined in a
8150: separate word list which is used when a @code{CODE} word is defined).
8151:
8152: @item
8153: To organize the words of an application or library into a user-visible
8154: set (in @code{forth-wordlist} or some other common wordlist) and a set
8155: of helper words used just for the implementation (hidden in a separate
8156: wordlist). This keeps @code{words}' output smaller, separates
8157: implementation and interface, and reduces the chance of name conflicts
8158: within the common wordlist.
8159:
8160: @item
8161: To prevent a name-space clash between multiple definitions with the same
8162: name. For example, when building a cross-compiler you might have a word
8163: @code{IF} that generates conditional code for your target system. By
8164: placing this definition in a different word list you can control whether
8165: the host system's @code{IF} or the target system's @code{IF} get used in
8166: any particular context by controlling the order of the word lists on the
8167: search order stack.
8168:
8169: @end itemize
8170:
8171: The downsides of using wordlists are:
8172:
8173: @itemize
8174:
8175: @item
8176: Debugging becomes more cumbersome.
8177:
8178: @item
8179: Name conflicts worked around with wordlists are still there, and you
8180: have to arrange the search order carefully to get the desired results;
8181: if you forget to do that, you get hard-to-find errors (as in any case
8182: where you read the code differently from the compiler; @code{see} can
8183: help seeing which of several possible words the name resolves to in such
8184: cases). @code{See} displays just the name of the words, not what
8185: wordlist they belong to, so it might be misleading. Using unique names
8186: is a better approach to avoid name conflicts.
8187:
8188: @item
8189: You have to explicitly undo any changes to the search order. In many
8190: cases it would be more convenient if this happened implicitly. Gforth
8191: currently does not provide such a feature, but it may do so in the
8192: future.
8193: @end itemize
8194:
8195:
8196: @node Word list example, , Why use word lists?, Word Lists
8197: @subsection Word list example
8198: @cindex word lists - example
8199:
8200: The following example is from the
8201: @uref{http://www.complang.tuwien.ac.at/forth/garbage-collection.zip,
8202: garbage collector} and uses wordlists to separate public words from
8203: helper words:
8204:
8205: @example
8206: get-current ( wid )
8207: vocabulary garbage-collector also garbage-collector definitions
8208: ... \ define helper words
8209: ( wid ) set-current \ restore original (i.e., public) compilation wordlist
8210: ... \ define the public (i.e., API) words
8211: \ they can refer to the helper words
8212: previous \ restore original search order (helper words become invisible)
8213: @end example
8214:
8215: @c -------------------------------------------------------------
8216: @node Environmental Queries, Files, Word Lists, Words
8217: @section Environmental Queries
8218: @cindex environmental queries
8219:
8220: ANS Forth introduced the idea of ``environmental queries'' as a way
8221: for a program running on a system to determine certain characteristics of the system.
8222: The Standard specifies a number of strings that might be recognised by a system.
8223:
8224: The Standard requires that the header space used for environmental queries
8225: be distinct from the header space used for definitions.
8226:
8227: Typically, environmental queries are supported by creating a set of
8228: definitions in a word list that is @i{only} used during environmental
8229: queries; that is what Gforth does. There is no Standard way of adding
8230: definitions to the set of recognised environmental queries, but any
8231: implementation that supports the loading of optional word sets must have
8232: some mechanism for doing this (after loading the word set, the
8233: associated environmental query string must return @code{true}). In
8234: Gforth, the word list used to honour environmental queries can be
8235: manipulated just like any other word list.
8236:
8237:
8238: doc-environment?
8239: doc-environment-wordlist
8240:
8241: doc-gforth
8242: doc-os-class
8243:
8244:
8245: Note that, whilst the documentation for (e.g.) @code{gforth} shows it
8246: returning two items on the stack, querying it using @code{environment?}
8247: will return an additional item; the @code{true} flag that shows that the
8248: string was recognised.
8249:
8250: @comment TODO Document the standard strings or note where they are documented herein
8251:
8252: Here are some examples of using environmental queries:
8253:
8254: @example
8255: s" address-unit-bits" environment? 0=
8256: [IF]
8257: cr .( environmental attribute address-units-bits unknown... ) cr
8258: [ELSE]
8259: drop \ ensure balanced stack effect
8260: [THEN]
8261:
8262: \ this might occur in the prelude of a standard program that uses THROW
8263: s" exception" environment? [IF]
8264: 0= [IF]
8265: : throw abort" exception thrown" ;
8266: [THEN]
8267: [ELSE] \ we don't know, so make sure
8268: : throw abort" exception thrown" ;
8269: [THEN]
8270:
8271: s" gforth" environment? [IF] .( Gforth version ) TYPE
8272: [ELSE] .( Not Gforth..) [THEN]
8273:
8274: \ a program using v*
8275: s" gforth" environment? [IF]
8276: s" 0.5.0" compare 0< [IF] \ v* is a primitive since 0.5.0
8277: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8278: >r swap 2swap swap 0e r> 0 ?DO
8279: dup f@ over + 2swap dup f@ f* f+ over + 2swap
8280: LOOP
8281: 2drop 2drop ;
8282: [THEN]
8283: [ELSE] \
8284: : v* ( f_addr1 nstride1 f_addr2 nstride2 ucount -- r )
8285: ...
8286: [THEN]
8287: @end example
8288:
8289: Here is an example of adding a definition to the environment word list:
8290:
8291: @example
8292: get-current environment-wordlist set-current
8293: true constant block
8294: true constant block-ext
8295: set-current
8296: @end example
8297:
8298: You can see what definitions are in the environment word list like this:
8299:
8300: @example
8301: environment-wordlist >order words previous
8302: @end example
8303:
8304:
8305: @c -------------------------------------------------------------
8306: @node Files, Blocks, Environmental Queries, Words
8307: @section Files
8308: @cindex files
8309: @cindex I/O - file-handling
8310:
8311: Gforth provides facilities for accessing files that are stored in the
8312: host operating system's file-system. Files that are processed by Gforth
8313: can be divided into two categories:
8314:
8315: @itemize @bullet
8316: @item
8317: Files that are processed by the Text Interpreter (@dfn{Forth source files}).
8318: @item
8319: Files that are processed by some other program (@dfn{general files}).
8320: @end itemize
8321:
8322: @menu
8323: * Forth source files::
8324: * General files::
8325: * Search Paths::
8326: @end menu
8327:
8328: @c -------------------------------------------------------------
8329: @node Forth source files, General files, Files, Files
8330: @subsection Forth source files
8331: @cindex including files
8332: @cindex Forth source files
8333:
8334: The simplest way to interpret the contents of a file is to use one of
8335: these two formats:
8336:
8337: @example
8338: include mysource.fs
8339: s" mysource.fs" included
8340: @end example
8341:
8342: You usually want to include a file only if it is not included already
8343: (by, say, another source file). In that case, you can use one of these
8344: three formats:
8345:
8346: @example
8347: require mysource.fs
8348: needs mysource.fs
8349: s" mysource.fs" required
8350: @end example
8351:
8352: @cindex stack effect of included files
8353: @cindex including files, stack effect
8354: It is good practice to write your source files such that interpreting them
8355: does not change the stack. Source files designed in this way can be used with
8356: @code{required} and friends without complications. For example:
8357:
8358: @example
8359: 1024 require foo.fs drop
8360: @end example
8361:
8362: Here you want to pass the argument 1024 (e.g., a buffer size) to
8363: @file{foo.fs}. Interpreting @file{foo.fs} has the stack effect ( n -- n
8364: ), which allows its use with @code{require}. Of course with such
8365: parameters to required files, you have to ensure that the first
8366: @code{require} fits for all uses (i.e., @code{require} it early in the
8367: master load file).
8368:
8369: doc-include-file
8370: doc-included
8371: doc-included?
8372: doc-include
8373: doc-required
8374: doc-require
8375: doc-needs
8376: @c doc-init-included-files @c internal
8377: doc-sourcefilename
8378: doc-sourceline#
8379:
8380: A definition in ANS Forth for @code{required} is provided in
8381: @file{compat/required.fs}.
8382:
8383: @c -------------------------------------------------------------
8384: @node General files, Search Paths, Forth source files, Files
8385: @subsection General files
8386: @cindex general files
8387: @cindex file-handling
8388:
8389: Files are opened/created by name and type. The following file access
8390: methods (FAMs) are recognised:
8391:
8392: @cindex fam (file access method)
8393: doc-r/o
8394: doc-r/w
8395: doc-w/o
8396: doc-bin
8397:
8398:
8399: When a file is opened/created, it returns a file identifier,
8400: @i{wfileid} that is used for all other file commands. All file
8401: commands also return a status value, @i{wior}, that is 0 for a
8402: successful operation and an implementation-defined non-zero value in the
8403: case of an error.
8404:
8405:
8406: doc-open-file
8407: doc-create-file
8408:
8409: doc-close-file
8410: doc-delete-file
8411: doc-rename-file
8412: doc-read-file
8413: doc-read-line
8414: doc-write-file
8415: doc-write-line
8416: doc-emit-file
8417: doc-flush-file
8418:
8419: doc-file-status
8420: doc-file-position
8421: doc-reposition-file
8422: doc-file-size
8423: doc-resize-file
8424:
8425:
8426: @c ---------------------------------------------------------
8427: @node Search Paths, , General files, Files
8428: @subsection Search Paths
8429: @cindex path for @code{included}
8430: @cindex file search path
8431: @cindex @code{include} search path
8432: @cindex search path for files
8433:
8434: If you specify an absolute filename (i.e., a filename starting with
8435: @file{/} or @file{~}, or with @file{:} in the second position (as in
8436: @samp{C:...})) for @code{included} and friends, that file is included
8437: just as you would expect.
8438:
8439: If the filename starts with @file{./}, this refers to the directory that
8440: the present file was @code{included} from. This allows files to include
8441: other files relative to their own position (irrespective of the current
8442: working directory or the absolute position). This feature is essential
8443: for libraries consisting of several files, where a file may include
8444: other files from the library. It corresponds to @code{#include "..."}
8445: in C. If the current input source is not a file, @file{.} refers to the
8446: directory of the innermost file being included, or, if there is no file
8447: being included, to the current working directory.
8448:
8449: For relative filenames (not starting with @file{./}), Gforth uses a
8450: search path similar to Forth's search order (@pxref{Word Lists}). It
8451: tries to find the given filename in the directories present in the path,
8452: and includes the first one it finds. There are separate search paths for
8453: Forth source files and general files. If the search path contains the
8454: directory @file{.}, this refers to the directory of the current file, or
8455: the working directory, as if the file had been specified with @file{./}.
8456:
8457: Use @file{~+} to refer to the current working directory (as in the
8458: @code{bash}).
8459:
8460: @c anton: fold the following subsubsections into this subsection?
8461:
8462: @menu
8463: * Source Search Paths::
8464: * General Search Paths::
8465: @end menu
8466:
8467: @c ---------------------------------------------------------
8468: @node Source Search Paths, General Search Paths, Search Paths, Search Paths
8469: @subsubsection Source Search Paths
8470: @cindex search path control, source files
8471:
8472: The search path is initialized when you start Gforth (@pxref{Invoking
8473: Gforth}). You can display it and change it using @code{fpath} in
8474: combination with the general path handling words.
8475:
8476: doc-fpath
8477: @c the functionality of the following words is easily available through
8478: @c fpath and the general path words. The may go away.
8479: @c doc-.fpath
8480: @c doc-fpath+
8481: @c doc-fpath=
8482: @c doc-open-fpath-file
8483:
8484: @noindent
8485: Here is an example of using @code{fpath} and @code{require}:
8486:
8487: @example
8488: fpath path= /usr/lib/forth/|./
8489: require timer.fs
8490: @end example
8491:
8492:
8493: @c ---------------------------------------------------------
8494: @node General Search Paths, , Source Search Paths, Search Paths
8495: @subsubsection General Search Paths
8496: @cindex search path control, source files
8497:
8498: Your application may need to search files in several directories, like
8499: @code{included} does. To facilitate this, Gforth allows you to define
8500: and use your own search paths, by providing generic equivalents of the
8501: Forth search path words:
8502:
8503: doc-open-path-file
8504: doc-path-allot
8505: doc-clear-path
8506: doc-also-path
8507: doc-.path
8508: doc-path+
8509: doc-path=
8510:
8511: @c anton: better define a word for it, say "path-allot ( ucount -- path-addr )
8512:
8513: Here's an example of creating an empty search path:
8514: @c
8515: @example
8516: create mypath 500 path-allot \ maximum length 500 chars (is checked)
8517: @end example
8518:
8519: @c -------------------------------------------------------------
8520: @node Blocks, Other I/O, Files, Words
8521: @section Blocks
8522: @cindex I/O - blocks
8523: @cindex blocks
8524:
8525: When you run Gforth on a modern desk-top computer, it runs under the
8526: control of an operating system which provides certain services. One of
8527: these services is @var{file services}, which allows Forth source code
8528: and data to be stored in files and read into Gforth (@pxref{Files}).
8529:
8530: Traditionally, Forth has been an important programming language on
8531: systems where it has interfaced directly to the underlying hardware with
8532: no intervening operating system. Forth provides a mechanism, called
8533: @dfn{blocks}, for accessing mass storage on such systems.
8534:
8535: A block is a 1024-byte data area, which can be used to hold data or
8536: Forth source code. No structure is imposed on the contents of the
8537: block. A block is identified by its number; blocks are numbered
8538: contiguously from 1 to an implementation-defined maximum.
8539:
8540: A typical system that used blocks but no operating system might use a
8541: single floppy-disk drive for mass storage, with the disks formatted to
8542: provide 256-byte sectors. Blocks would be implemented by assigning the
8543: first four sectors of the disk to block 1, the second four sectors to
8544: block 2 and so on, up to the limit of the capacity of the disk. The disk
8545: would not contain any file system information, just the set of blocks.
8546:
8547: @cindex blocks file
8548: On systems that do provide file services, blocks are typically
8549: implemented by storing a sequence of blocks within a single @dfn{blocks
8550: file}. The size of the blocks file will be an exact multiple of 1024
8551: bytes, corresponding to the number of blocks it contains. This is the
8552: mechanism that Gforth uses.
8553:
8554: @cindex @file{blocks.fb}
8555: Only one blocks file can be open at a time. If you use block words without
8556: having specified a blocks file, Gforth defaults to the blocks file
8557: @file{blocks.fb}. Gforth uses the Forth search path when attempting to
8558: locate a blocks file (@pxref{Source Search Paths}).
8559:
8560: @cindex block buffers
8561: When you read and write blocks under program control, Gforth uses a
8562: number of @dfn{block buffers} as intermediate storage. These buffers are
8563: not used when you use @code{load} to interpret the contents of a block.
8564:
8565: The behaviour of the block buffers is analagous to that of a cache.
8566: Each block buffer has three states:
8567:
8568: @itemize @bullet
8569: @item
8570: Unassigned
8571: @item
8572: Assigned-clean
8573: @item
8574: Assigned-dirty
8575: @end itemize
8576:
8577: Initially, all block buffers are @i{unassigned}. In order to access a
8578: block, the block (specified by its block number) must be assigned to a
8579: block buffer.
8580:
8581: The assignment of a block to a block buffer is performed by @code{block}
8582: or @code{buffer}. Use @code{block} when you wish to modify the existing
8583: contents of a block. Use @code{buffer} when you don't care about the
8584: existing contents of the block@footnote{The ANS Forth definition of
8585: @code{buffer} is intended not to cause disk I/O; if the data associated
8586: with the particular block is already stored in a block buffer due to an
8587: earlier @code{block} command, @code{buffer} will return that block
8588: buffer and the existing contents of the block will be
8589: available. Otherwise, @code{buffer} will simply assign a new, empty
8590: block buffer for the block.}.
8591:
8592: Once a block has been assigned to a block buffer using @code{block} or
8593: @code{buffer}, that block buffer becomes the @i{current block
8594: buffer}. Data may only be manipulated (read or written) within the
8595: current block buffer.
8596:
8597: When the contents of the current block buffer has been modified it is
8598: necessary, @emph{before calling @code{block} or @code{buffer} again}, to
8599: either abandon the changes (by doing nothing) or mark the block as
8600: changed (assigned-dirty), using @code{update}. Using @code{update} does
8601: not change the blocks file; it simply changes a block buffer's state to
8602: @i{assigned-dirty}. The block will be written implicitly when it's
8603: buffer is needed for another block, or explicitly by @code{flush} or
8604: @code{save-buffers}.
8605:
8606: word @code{Flush} writes all @i{assigned-dirty} blocks back to the
8607: blocks file on disk. Leaving Gforth with @code{bye} also performs a
8608: @code{flush}.
8609:
8610: In Gforth, @code{block} and @code{buffer} use a @i{direct-mapped}
8611: algorithm to assign a block buffer to a block. That means that any
8612: particular block can only be assigned to one specific block buffer,
8613: called (for the particular operation) the @i{victim buffer}. If the
8614: victim buffer is @i{unassigned} or @i{assigned-clean} it is allocated to
8615: the new block immediately. If it is @i{assigned-dirty} its current
8616: contents are written back to the blocks file on disk before it is
8617: allocated to the new block.
8618:
8619: Although no structure is imposed on the contents of a block, it is
8620: traditional to display the contents as 16 lines each of 64 characters. A
8621: block provides a single, continuous stream of input (for example, it
8622: acts as a single parse area) -- there are no end-of-line characters
8623: within a block, and no end-of-file character at the end of a
8624: block. There are two consequences of this:
8625:
8626: @itemize @bullet
8627: @item
8628: The last character of one line wraps straight into the first character
8629: of the following line
8630: @item
8631: The word @code{\} -- comment to end of line -- requires special
8632: treatment; in the context of a block it causes all characters until the
8633: end of the current 64-character ``line'' to be ignored.
8634: @end itemize
8635:
8636: In Gforth, when you use @code{block} with a non-existent block number,
8637: the current blocks file will be extended to the appropriate size and the
8638: block buffer will be initialised with spaces.
8639:
8640: Gforth includes a simple block editor (type @code{use blocked.fb 0 list}
8641: for details) but doesn't encourage the use of blocks; the mechanism is
8642: only provided for backward compatibility -- ANS Forth requires blocks to
8643: be available when files are.
8644:
8645: Common techniques that are used when working with blocks include:
8646:
8647: @itemize @bullet
8648: @item
8649: A screen editor that allows you to edit blocks without leaving the Forth
8650: environment.
8651: @item
8652: Shadow screens; where every code block has an associated block
8653: containing comments (for example: code in odd block numbers, comments in
8654: even block numbers). Typically, the block editor provides a convenient
8655: mechanism to toggle between code and comments.
8656: @item
8657: Load blocks; a single block (typically block 1) contains a number of
8658: @code{thru} commands which @code{load} the whole of the application.
8659: @end itemize
8660:
8661: See Frank Sergeant's Pygmy Forth to see just how well blocks can be
8662: integrated into a Forth programming environment.
8663:
8664: @comment TODO what about errors on open-blocks?
8665:
8666: doc-open-blocks
8667: doc-use
8668: doc-block-offset
8669: doc-get-block-fid
8670: doc-block-position
8671:
8672: doc-list
8673: doc-scr
8674:
8675: doc---gforthman-block
8676: doc-buffer
8677:
8678: doc-empty-buffers
8679: doc-empty-buffer
8680: doc-update
8681: doc-updated?
8682: doc-save-buffers
8683: doc-save-buffer
8684: doc-flush
8685:
8686: doc-load
8687: doc-thru
8688: doc-+load
8689: doc-+thru
8690: doc---gforthman--->
8691: doc-block-included
8692:
8693:
8694: @c -------------------------------------------------------------
8695: @node Other I/O, Locals, Blocks, Words
8696: @section Other I/O
8697: @cindex I/O - keyboard and display
8698:
8699: @menu
8700: * Simple numeric output:: Predefined formats
8701: * Formatted numeric output:: Formatted (pictured) output
8702: * String Formats:: How Forth stores strings in memory
8703: * Displaying characters and strings:: Other stuff
8704: * Input:: Input
8705: @end menu
8706:
8707: @node Simple numeric output, Formatted numeric output, Other I/O, Other I/O
8708: @subsection Simple numeric output
8709: @cindex numeric output - simple/free-format
8710:
8711: The simplest output functions are those that display numbers from the
8712: data or floating-point stacks. Floating-point output is always displayed
8713: using base 10. Numbers displayed from the data stack use the value stored
8714: in @code{base}.
8715:
8716:
8717: doc-.
8718: doc-dec.
8719: doc-hex.
8720: doc-u.
8721: doc-.r
8722: doc-u.r
8723: doc-d.
8724: doc-ud.
8725: doc-d.r
8726: doc-ud.r
8727: doc-f.
8728: doc-fe.
8729: doc-fs.
8730:
8731:
8732: Examples of printing the number 1234.5678E23 in the different floating-point output
8733: formats are shown below:
8734:
8735: @example
8736: f. 123456779999999000000000000.
8737: fe. 123.456779999999E24
8738: fs. 1.23456779999999E26
8739: @end example
8740:
8741:
8742: @node Formatted numeric output, String Formats, Simple numeric output, Other I/O
8743: @subsection Formatted numeric output
8744: @cindex formatted numeric output
8745: @cindex pictured numeric output
8746: @cindex numeric output - formatted
8747:
8748: Forth traditionally uses a technique called @dfn{pictured numeric
8749: output} for formatted printing of integers. In this technique, digits
8750: are extracted from the number (using the current output radix defined by
8751: @code{base}), converted to ASCII codes and appended to a string that is
8752: built in a scratch-pad area of memory (@pxref{core-idef,
8753: Implementation-defined options, Implementation-defined
8754: options}). Arbitrary characters can be appended to the string during the
8755: extraction process. The completed string is specified by an address
8756: and length and can be manipulated (@code{TYPE}ed, copied, modified)
8757: under program control.
8758:
8759: All of the integer output words described in the previous section
8760: (@pxref{Simple numeric output}) are implemented in Gforth using pictured
8761: numeric output.
8762:
8763: Three important things to remember about pictured numeric output:
8764:
8765: @itemize @bullet
8766: @item
8767: It always operates on double-precision numbers; to display a
8768: single-precision number, convert it first (for ways of doing this
8769: @pxref{Double precision}).
8770: @item
8771: It always treats the double-precision number as though it were
8772: unsigned. The examples below show ways of printing signed numbers.
8773: @item
8774: The string is built up from right to left; least significant digit first.
8775: @end itemize
8776:
8777:
8778: doc-<#
8779: doc-<<#
8780: doc-#
8781: doc-#s
8782: doc-hold
8783: doc-sign
8784: doc-#>
8785: doc-#>>
8786:
8787: doc-represent
8788:
8789:
8790: @noindent
8791: Here are some examples of using pictured numeric output:
8792:
8793: @example
8794: : my-u. ( u -- )
8795: \ Simplest use of pns.. behaves like Standard u.
8796: 0 \ convert to unsigned double
8797: <<# \ start conversion
8798: #s \ convert all digits
8799: #> \ complete conversion
8800: TYPE SPACE \ display, with trailing space
8801: #>> ; \ release hold area
8802:
8803: : cents-only ( u -- )
8804: 0 \ convert to unsigned double
8805: <<# \ start conversion
8806: # # \ convert two least-significant digits
8807: #> \ complete conversion, discard other digits
8808: TYPE SPACE \ display, with trailing space
8809: #>> ; \ release hold area
8810:
8811: : dollars-and-cents ( u -- )
8812: 0 \ convert to unsigned double
8813: <<# \ start conversion
8814: # # \ convert two least-significant digits
8815: [char] . hold \ insert decimal point
8816: #s \ convert remaining digits
8817: [char] $ hold \ append currency symbol
8818: #> \ complete conversion
8819: TYPE SPACE \ display, with trailing space
8820: #>> ; \ release hold area
8821:
8822: : my-. ( n -- )
8823: \ handling negatives.. behaves like Standard .
8824: s>d \ convert to signed double
8825: swap over dabs \ leave sign byte followed by unsigned double
8826: <<# \ start conversion
8827: #s \ convert all digits
8828: rot sign \ get at sign byte, append "-" if needed
8829: #> \ complete conversion
8830: TYPE SPACE \ display, with trailing space
8831: #>> ; \ release hold area
8832:
8833: : account. ( n -- )
8834: \ accountants don't like minus signs, they use parentheses
8835: \ for negative numbers
8836: s>d \ convert to signed double
8837: swap over dabs \ leave sign byte followed by unsigned double
8838: <<# \ start conversion
8839: 2 pick \ get copy of sign byte
8840: 0< IF [char] ) hold THEN \ right-most character of output
8841: #s \ convert all digits
8842: rot \ get at sign byte
8843: 0< IF [char] ( hold THEN
8844: #> \ complete conversion
8845: TYPE SPACE \ display, with trailing space
8846: #>> ; \ release hold area
8847:
8848: @end example
8849:
8850: Here are some examples of using these words:
8851:
8852: @example
8853: 1 my-u. 1
8854: hex -1 my-u. decimal FFFFFFFF
8855: 1 cents-only 01
8856: 1234 cents-only 34
8857: 2 dollars-and-cents $0.02
8858: 1234 dollars-and-cents $12.34
8859: 123 my-. 123
8860: -123 my. -123
8861: 123 account. 123
8862: -456 account. (456)
8863: @end example
8864:
8865:
8866: @node String Formats, Displaying characters and strings, Formatted numeric output, Other I/O
8867: @subsection String Formats
8868: @cindex strings - see character strings
8869: @cindex character strings - formats
8870: @cindex I/O - see character strings
8871: @cindex counted strings
8872:
8873: @c anton: this does not really belong here; maybe the memory section,
8874: @c or the principles chapter
8875:
8876: Forth commonly uses two different methods for representing character
8877: strings:
8878:
8879: @itemize @bullet
8880: @item
8881: @cindex address of counted string
8882: @cindex counted string
8883: As a @dfn{counted string}, represented by a @i{c-addr}. The char
8884: addressed by @i{c-addr} contains a character-count, @i{n}, of the
8885: string and the string occupies the subsequent @i{n} char addresses in
8886: memory.
8887: @item
8888: As cell pair on the stack; @i{c-addr u}, where @i{u} is the length
8889: of the string in characters, and @i{c-addr} is the address of the
8890: first byte of the string.
8891: @end itemize
8892:
8893: ANS Forth encourages the use of the second format when representing
8894: strings.
8895:
8896:
8897: doc-count
8898:
8899:
8900: For words that move, copy and search for strings see @ref{Memory
8901: Blocks}. For words that display characters and strings see
8902: @ref{Displaying characters and strings}.
8903:
8904: @node Displaying characters and strings, Input, String Formats, Other I/O
8905: @subsection Displaying characters and strings
8906: @cindex characters - compiling and displaying
8907: @cindex character strings - compiling and displaying
8908:
8909: This section starts with a glossary of Forth words and ends with a set
8910: of examples.
8911:
8912:
8913: doc-bl
8914: doc-space
8915: doc-spaces
8916: doc-emit
8917: doc-toupper
8918: doc-."
8919: doc-.(
8920: doc-type
8921: doc-typewhite
8922: doc-cr
8923: @cindex cursor control
8924: doc-at-xy
8925: doc-page
8926: doc-s"
8927: doc-c"
8928: doc-char
8929: doc-[char]
8930:
8931:
8932: @noindent
8933: As an example, consider the following text, stored in a file @file{test.fs}:
8934:
8935: @example
8936: .( text-1)
8937: : my-word
8938: ." text-2" cr
8939: .( text-3)
8940: ;
8941:
8942: ." text-4"
8943:
8944: : my-char
8945: [char] ALPHABET emit
8946: char emit
8947: ;
8948: @end example
8949:
8950: When you load this code into Gforth, the following output is generated:
8951:
8952: @example
8953: @kbd{include test.fs @key{RET}} text-1text-3text-4 ok
8954: @end example
8955:
8956: @itemize @bullet
8957: @item
8958: Messages @code{text-1} and @code{text-3} are displayed because @code{.(}
8959: is an immediate word; it behaves in the same way whether it is used inside
8960: or outside a colon definition.
8961: @item
8962: Message @code{text-4} is displayed because of Gforth's added interpretation
8963: semantics for @code{."}.
8964: @item
8965: Message @code{text-2} is @i{not} displayed, because the text interpreter
8966: performs the compilation semantics for @code{."} within the definition of
8967: @code{my-word}.
8968: @end itemize
8969:
8970: Here are some examples of executing @code{my-word} and @code{my-char}:
8971:
8972: @example
8973: @kbd{my-word @key{RET}} text-2
8974: ok
8975: @kbd{my-char fred @key{RET}} Af ok
8976: @kbd{my-char jim @key{RET}} Aj ok
8977: @end example
8978:
8979: @itemize @bullet
8980: @item
8981: Message @code{text-2} is displayed because of the run-time behaviour of
8982: @code{."}.
8983: @item
8984: @code{[char]} compiles the ``A'' from ``ALPHABET'' and puts its display code
8985: on the stack at run-time. @code{emit} always displays the character
8986: when @code{my-char} is executed.
8987: @item
8988: @code{char} parses a string at run-time and the second @code{emit} displays
8989: the first character of the string.
8990: @item
8991: If you type @code{see my-char} you can see that @code{[char]} discarded
8992: the text ``LPHABET'' and only compiled the display code for ``A'' into the
8993: definition of @code{my-char}.
8994: @end itemize
8995:
8996:
8997:
8998: @node Input, , Displaying characters and strings, Other I/O
8999: @subsection Input
9000: @cindex input
9001: @cindex I/O - see input
9002: @cindex parsing a string
9003:
9004: For ways of storing character strings in memory see @ref{String Formats}.
9005:
9006: @comment TODO examples for >number >float accept key key? pad parse word refill
9007: @comment then index them
9008:
9009:
9010: doc-key
9011: doc-key?
9012: doc-ekey
9013: doc-ekey?
9014: doc-ekey>char
9015: doc->number
9016: doc->float
9017: doc-accept
9018: doc-pad
9019: @c anton: these belong in the input stream section
9020: doc-parse
9021: doc-word
9022: doc-sword
9023: doc-name
9024: doc-refill
9025: @comment obsolescent words..
9026: doc-convert
9027: doc-query
9028: doc-expect
9029: doc-span
9030:
9031:
9032: @c -------------------------------------------------------------
9033: @node Locals, Structures, Other I/O, Words
9034: @section Locals
9035: @cindex locals
9036:
9037: Local variables can make Forth programming more enjoyable and Forth
9038: programs easier to read. Unfortunately, the locals of ANS Forth are
9039: laden with restrictions. Therefore, we provide not only the ANS Forth
9040: locals wordset, but also our own, more powerful locals wordset (we
9041: implemented the ANS Forth locals wordset through our locals wordset).
9042:
9043: The ideas in this section have also been published in M. Anton Ertl,
9044: @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl94l.ps.gz,
9045: Automatic Scoping of Local Variables}}, EuroForth '94.
9046:
9047: @menu
9048: * Gforth locals::
9049: * ANS Forth locals::
9050: @end menu
9051:
9052: @node Gforth locals, ANS Forth locals, Locals, Locals
9053: @subsection Gforth locals
9054: @cindex Gforth locals
9055: @cindex locals, Gforth style
9056:
9057: Locals can be defined with
9058:
9059: @example
9060: @{ local1 local2 ... -- comment @}
9061: @end example
9062: or
9063: @example
9064: @{ local1 local2 ... @}
9065: @end example
9066:
9067: E.g.,
9068: @example
9069: : max @{ n1 n2 -- n3 @}
9070: n1 n2 > if
9071: n1
9072: else
9073: n2
9074: endif ;
9075: @end example
9076:
9077: The similarity of locals definitions with stack comments is intended. A
9078: locals definition often replaces the stack comment of a word. The order
9079: of the locals corresponds to the order in a stack comment and everything
9080: after the @code{--} is really a comment.
9081:
9082: This similarity has one disadvantage: It is too easy to confuse locals
9083: declarations with stack comments, causing bugs and making them hard to
9084: find. However, this problem can be avoided by appropriate coding
9085: conventions: Do not use both notations in the same program. If you do,
9086: they should be distinguished using additional means, e.g. by position.
9087:
9088: @cindex types of locals
9089: @cindex locals types
9090: The name of the local may be preceded by a type specifier, e.g.,
9091: @code{F:} for a floating point value:
9092:
9093: @example
9094: : CX* @{ F: Ar F: Ai F: Br F: Bi -- Cr Ci @}
9095: \ complex multiplication
9096: Ar Br f* Ai Bi f* f-
9097: Ar Bi f* Ai Br f* f+ ;
9098: @end example
9099:
9100: @cindex flavours of locals
9101: @cindex locals flavours
9102: @cindex value-flavoured locals
9103: @cindex variable-flavoured locals
9104: Gforth currently supports cells (@code{W:}, @code{W^}), doubles
9105: (@code{D:}, @code{D^}), floats (@code{F:}, @code{F^}) and characters
9106: (@code{C:}, @code{C^}) in two flavours: a value-flavoured local (defined
9107: with @code{W:}, @code{D:} etc.) produces its value and can be changed
9108: with @code{TO}. A variable-flavoured local (defined with @code{W^} etc.)
9109: produces its address (which becomes invalid when the variable's scope is
9110: left). E.g., the standard word @code{emit} can be defined in terms of
9111: @code{type} like this:
9112:
9113: @example
9114: : emit @{ C^ char* -- @}
9115: char* 1 type ;
9116: @end example
9117:
9118: @cindex default type of locals
9119: @cindex locals, default type
9120: A local without type specifier is a @code{W:} local. Both flavours of
9121: locals are initialized with values from the data or FP stack.
9122:
9123: Currently there is no way to define locals with user-defined data
9124: structures, but we are working on it.
9125:
9126: Gforth allows defining locals everywhere in a colon definition. This
9127: poses the following questions:
9128:
9129: @menu
9130: * Where are locals visible by name?::
9131: * How long do locals live?::
9132: * Locals programming style::
9133: * Locals implementation::
9134: @end menu
9135:
9136: @node Where are locals visible by name?, How long do locals live?, Gforth locals, Gforth locals
9137: @subsubsection Where are locals visible by name?
9138: @cindex locals visibility
9139: @cindex visibility of locals
9140: @cindex scope of locals
9141:
9142: Basically, the answer is that locals are visible where you would expect
9143: it in block-structured languages, and sometimes a little longer. If you
9144: want to restrict the scope of a local, enclose its definition in
9145: @code{SCOPE}...@code{ENDSCOPE}.
9146:
9147:
9148: doc-scope
9149: doc-endscope
9150:
9151:
9152: These words behave like control structure words, so you can use them
9153: with @code{CS-PICK} and @code{CS-ROLL} to restrict the scope in
9154: arbitrary ways.
9155:
9156: If you want a more exact answer to the visibility question, here's the
9157: basic principle: A local is visible in all places that can only be
9158: reached through the definition of the local@footnote{In compiler
9159: construction terminology, all places dominated by the definition of the
9160: local.}. In other words, it is not visible in places that can be reached
9161: without going through the definition of the local. E.g., locals defined
9162: in @code{IF}...@code{ENDIF} are visible until the @code{ENDIF}, locals
9163: defined in @code{BEGIN}...@code{UNTIL} are visible after the
9164: @code{UNTIL} (until, e.g., a subsequent @code{ENDSCOPE}).
9165:
9166: The reasoning behind this solution is: We want to have the locals
9167: visible as long as it is meaningful. The user can always make the
9168: visibility shorter by using explicit scoping. In a place that can
9169: only be reached through the definition of a local, the meaning of a
9170: local name is clear. In other places it is not: How is the local
9171: initialized at the control flow path that does not contain the
9172: definition? Which local is meant, if the same name is defined twice in
9173: two independent control flow paths?
9174:
9175: This should be enough detail for nearly all users, so you can skip the
9176: rest of this section. If you really must know all the gory details and
9177: options, read on.
9178:
9179: In order to implement this rule, the compiler has to know which places
9180: are unreachable. It knows this automatically after @code{AHEAD},
9181: @code{AGAIN}, @code{EXIT} and @code{LEAVE}; in other cases (e.g., after
9182: most @code{THROW}s), you can use the word @code{UNREACHABLE} to tell the
9183: compiler that the control flow never reaches that place. If
9184: @code{UNREACHABLE} is not used where it could, the only consequence is
9185: that the visibility of some locals is more limited than the rule above
9186: says. If @code{UNREACHABLE} is used where it should not (i.e., if you
9187: lie to the compiler), buggy code will be produced.
9188:
9189:
9190: doc-unreachable
9191:
9192:
9193: Another problem with this rule is that at @code{BEGIN}, the compiler
9194: does not know which locals will be visible on the incoming
9195: back-edge. All problems discussed in the following are due to this
9196: ignorance of the compiler (we discuss the problems using @code{BEGIN}
9197: loops as examples; the discussion also applies to @code{?DO} and other
9198: loops). Perhaps the most insidious example is:
9199: @example
9200: AHEAD
9201: BEGIN
9202: x
9203: [ 1 CS-ROLL ] THEN
9204: @{ x @}
9205: ...
9206: UNTIL
9207: @end example
9208:
9209: This should be legal according to the visibility rule. The use of
9210: @code{x} can only be reached through the definition; but that appears
9211: textually below the use.
9212:
9213: From this example it is clear that the visibility rules cannot be fully
9214: implemented without major headaches. Our implementation treats common
9215: cases as advertised and the exceptions are treated in a safe way: The
9216: compiler makes a reasonable guess about the locals visible after a
9217: @code{BEGIN}; if it is too pessimistic, the
9218: user will get a spurious error about the local not being defined; if the
9219: compiler is too optimistic, it will notice this later and issue a
9220: warning. In the case above the compiler would complain about @code{x}
9221: being undefined at its use. You can see from the obscure examples in
9222: this section that it takes quite unusual control structures to get the
9223: compiler into trouble, and even then it will often do fine.
9224:
9225: If the @code{BEGIN} is reachable from above, the most optimistic guess
9226: is that all locals visible before the @code{BEGIN} will also be
9227: visible after the @code{BEGIN}. This guess is valid for all loops that
9228: are entered only through the @code{BEGIN}, in particular, for normal
9229: @code{BEGIN}...@code{WHILE}...@code{REPEAT} and
9230: @code{BEGIN}...@code{UNTIL} loops and it is implemented in our
9231: compiler. When the branch to the @code{BEGIN} is finally generated by
9232: @code{AGAIN} or @code{UNTIL}, the compiler checks the guess and
9233: warns the user if it was too optimistic:
9234: @example
9235: IF
9236: @{ x @}
9237: BEGIN
9238: \ x ?
9239: [ 1 cs-roll ] THEN
9240: ...
9241: UNTIL
9242: @end example
9243:
9244: Here, @code{x} lives only until the @code{BEGIN}, but the compiler
9245: optimistically assumes that it lives until the @code{THEN}. It notices
9246: this difference when it compiles the @code{UNTIL} and issues a
9247: warning. The user can avoid the warning, and make sure that @code{x}
9248: is not used in the wrong area by using explicit scoping:
9249: @example
9250: IF
9251: SCOPE
9252: @{ x @}
9253: ENDSCOPE
9254: BEGIN
9255: [ 1 cs-roll ] THEN
9256: ...
9257: UNTIL
9258: @end example
9259:
9260: Since the guess is optimistic, there will be no spurious error messages
9261: about undefined locals.
9262:
9263: If the @code{BEGIN} is not reachable from above (e.g., after
9264: @code{AHEAD} or @code{EXIT}), the compiler cannot even make an
9265: optimistic guess, as the locals visible after the @code{BEGIN} may be
9266: defined later. Therefore, the compiler assumes that no locals are
9267: visible after the @code{BEGIN}. However, the user can use
9268: @code{ASSUME-LIVE} to make the compiler assume that the same locals are
9269: visible at the BEGIN as at the point where the top control-flow stack
9270: item was created.
9271:
9272:
9273: doc-assume-live
9274:
9275:
9276: @noindent
9277: E.g.,
9278: @example
9279: @{ x @}
9280: AHEAD
9281: ASSUME-LIVE
9282: BEGIN
9283: x
9284: [ 1 CS-ROLL ] THEN
9285: ...
9286: UNTIL
9287: @end example
9288:
9289: Other cases where the locals are defined before the @code{BEGIN} can be
9290: handled by inserting an appropriate @code{CS-ROLL} before the
9291: @code{ASSUME-LIVE} (and changing the control-flow stack manipulation
9292: behind the @code{ASSUME-LIVE}).
9293:
9294: Cases where locals are defined after the @code{BEGIN} (but should be
9295: visible immediately after the @code{BEGIN}) can only be handled by
9296: rearranging the loop. E.g., the ``most insidious'' example above can be
9297: arranged into:
9298: @example
9299: BEGIN
9300: @{ x @}
9301: ... 0=
9302: WHILE
9303: x
9304: REPEAT
9305: @end example
9306:
9307: @node How long do locals live?, Locals programming style, Where are locals visible by name?, Gforth locals
9308: @subsubsection How long do locals live?
9309: @cindex locals lifetime
9310: @cindex lifetime of locals
9311:
9312: The right answer for the lifetime question would be: A local lives at
9313: least as long as it can be accessed. For a value-flavoured local this
9314: means: until the end of its visibility. However, a variable-flavoured
9315: local could be accessed through its address far beyond its visibility
9316: scope. Ultimately, this would mean that such locals would have to be
9317: garbage collected. Since this entails un-Forth-like implementation
9318: complexities, I adopted the same cowardly solution as some other
9319: languages (e.g., C): The local lives only as long as it is visible;
9320: afterwards its address is invalid (and programs that access it
9321: afterwards are erroneous).
9322:
9323: @node Locals programming style, Locals implementation, How long do locals live?, Gforth locals
9324: @subsubsection Locals programming style
9325: @cindex locals programming style
9326: @cindex programming style, locals
9327:
9328: The freedom to define locals anywhere has the potential to change
9329: programming styles dramatically. In particular, the need to use the
9330: return stack for intermediate storage vanishes. Moreover, all stack
9331: manipulations (except @code{PICK}s and @code{ROLL}s with run-time
9332: determined arguments) can be eliminated: If the stack items are in the
9333: wrong order, just write a locals definition for all of them; then
9334: write the items in the order you want.
9335:
9336: This seems a little far-fetched and eliminating stack manipulations is
9337: unlikely to become a conscious programming objective. Still, the number
9338: of stack manipulations will be reduced dramatically if local variables
9339: are used liberally (e.g., compare @code{max} (@pxref{Gforth locals}) with
9340: a traditional implementation of @code{max}).
9341:
9342: This shows one potential benefit of locals: making Forth programs more
9343: readable. Of course, this benefit will only be realized if the
9344: programmers continue to honour the principle of factoring instead of
9345: using the added latitude to make the words longer.
9346:
9347: @cindex single-assignment style for locals
9348: Using @code{TO} can and should be avoided. Without @code{TO},
9349: every value-flavoured local has only a single assignment and many
9350: advantages of functional languages apply to Forth. I.e., programs are
9351: easier to analyse, to optimize and to read: It is clear from the
9352: definition what the local stands for, it does not turn into something
9353: different later.
9354:
9355: E.g., a definition using @code{TO} might look like this:
9356: @example
9357: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9358: u1 u2 min 0
9359: ?do
9360: addr1 c@@ addr2 c@@ -
9361: ?dup-if
9362: unloop exit
9363: then
9364: addr1 char+ TO addr1
9365: addr2 char+ TO addr2
9366: loop
9367: u1 u2 - ;
9368: @end example
9369: Here, @code{TO} is used to update @code{addr1} and @code{addr2} at
9370: every loop iteration. @code{strcmp} is a typical example of the
9371: readability problems of using @code{TO}. When you start reading
9372: @code{strcmp}, you think that @code{addr1} refers to the start of the
9373: string. Only near the end of the loop you realize that it is something
9374: else.
9375:
9376: This can be avoided by defining two locals at the start of the loop that
9377: are initialized with the right value for the current iteration.
9378: @example
9379: : strcmp @{ addr1 u1 addr2 u2 -- n @}
9380: addr1 addr2
9381: u1 u2 min 0
9382: ?do @{ s1 s2 @}
9383: s1 c@@ s2 c@@ -
9384: ?dup-if
9385: unloop exit
9386: then
9387: s1 char+ s2 char+
9388: loop
9389: 2drop
9390: u1 u2 - ;
9391: @end example
9392: Here it is clear from the start that @code{s1} has a different value
9393: in every loop iteration.
9394:
9395: @node Locals implementation, , Locals programming style, Gforth locals
9396: @subsubsection Locals implementation
9397: @cindex locals implementation
9398: @cindex implementation of locals
9399:
9400: @cindex locals stack
9401: Gforth uses an extra locals stack. The most compelling reason for
9402: this is that the return stack is not float-aligned; using an extra stack
9403: also eliminates the problems and restrictions of using the return stack
9404: as locals stack. Like the other stacks, the locals stack grows toward
9405: lower addresses. A few primitives allow an efficient implementation:
9406:
9407:
9408: doc-@local#
9409: doc-f@local#
9410: doc-laddr#
9411: doc-lp+!#
9412: doc-lp!
9413: doc->l
9414: doc-f>l
9415:
9416:
9417: In addition to these primitives, some specializations of these
9418: primitives for commonly occurring inline arguments are provided for
9419: efficiency reasons, e.g., @code{@@local0} as specialization of
9420: @code{@@local#} for the inline argument 0. The following compiling words
9421: compile the right specialized version, or the general version, as
9422: appropriate:
9423:
9424:
9425: doc-compile-@local
9426: doc-compile-f@local
9427: doc-compile-lp+!
9428:
9429:
9430: Combinations of conditional branches and @code{lp+!#} like
9431: @code{?branch-lp+!#} (the locals pointer is only changed if the branch
9432: is taken) are provided for efficiency and correctness in loops.
9433:
9434: A special area in the dictionary space is reserved for keeping the
9435: local variable names. @code{@{} switches the dictionary pointer to this
9436: area and @code{@}} switches it back and generates the locals
9437: initializing code. @code{W:} etc.@ are normal defining words. This
9438: special area is cleared at the start of every colon definition.
9439:
9440: @cindex word list for defining locals
9441: A special feature of Gforth's dictionary is used to implement the
9442: definition of locals without type specifiers: every word list (aka
9443: vocabulary) has its own methods for searching
9444: etc. (@pxref{Word Lists}). For the present purpose we defined a word list
9445: with a special search method: When it is searched for a word, it
9446: actually creates that word using @code{W:}. @code{@{} changes the search
9447: order to first search the word list containing @code{@}}, @code{W:} etc.,
9448: and then the word list for defining locals without type specifiers.
9449:
9450: The lifetime rules support a stack discipline within a colon
9451: definition: The lifetime of a local is either nested with other locals
9452: lifetimes or it does not overlap them.
9453:
9454: At @code{BEGIN}, @code{IF}, and @code{AHEAD} no code for locals stack
9455: pointer manipulation is generated. Between control structure words
9456: locals definitions can push locals onto the locals stack. @code{AGAIN}
9457: is the simplest of the other three control flow words. It has to
9458: restore the locals stack depth of the corresponding @code{BEGIN}
9459: before branching. The code looks like this:
9460: @format
9461: @code{lp+!#} current-locals-size @minus{} dest-locals-size
9462: @code{branch} <begin>
9463: @end format
9464:
9465: @code{UNTIL} is a little more complicated: If it branches back, it
9466: must adjust the stack just like @code{AGAIN}. But if it falls through,
9467: the locals stack must not be changed. The compiler generates the
9468: following code:
9469: @format
9470: @code{?branch-lp+!#} <begin> current-locals-size @minus{} dest-locals-size
9471: @end format
9472: The locals stack pointer is only adjusted if the branch is taken.
9473:
9474: @code{THEN} can produce somewhat inefficient code:
9475: @format
9476: @code{lp+!#} current-locals-size @minus{} orig-locals-size
9477: <orig target>:
9478: @code{lp+!#} orig-locals-size @minus{} new-locals-size
9479: @end format
9480: The second @code{lp+!#} adjusts the locals stack pointer from the
9481: level at the @i{orig} point to the level after the @code{THEN}. The
9482: first @code{lp+!#} adjusts the locals stack pointer from the current
9483: level to the level at the orig point, so the complete effect is an
9484: adjustment from the current level to the right level after the
9485: @code{THEN}.
9486:
9487: @cindex locals information on the control-flow stack
9488: @cindex control-flow stack items, locals information
9489: In a conventional Forth implementation a dest control-flow stack entry
9490: is just the target address and an orig entry is just the address to be
9491: patched. Our locals implementation adds a word list to every orig or dest
9492: item. It is the list of locals visible (or assumed visible) at the point
9493: described by the entry. Our implementation also adds a tag to identify
9494: the kind of entry, in particular to differentiate between live and dead
9495: (reachable and unreachable) orig entries.
9496:
9497: A few unusual operations have to be performed on locals word lists:
9498:
9499:
9500: doc-common-list
9501: doc-sub-list?
9502: doc-list-size
9503:
9504:
9505: Several features of our locals word list implementation make these
9506: operations easy to implement: The locals word lists are organised as
9507: linked lists; the tails of these lists are shared, if the lists
9508: contain some of the same locals; and the address of a name is greater
9509: than the address of the names behind it in the list.
9510:
9511: Another important implementation detail is the variable
9512: @code{dead-code}. It is used by @code{BEGIN} and @code{THEN} to
9513: determine if they can be reached directly or only through the branch
9514: that they resolve. @code{dead-code} is set by @code{UNREACHABLE},
9515: @code{AHEAD}, @code{EXIT} etc., and cleared at the start of a colon
9516: definition, by @code{BEGIN} and usually by @code{THEN}.
9517:
9518: Counted loops are similar to other loops in most respects, but
9519: @code{LEAVE} requires special attention: It performs basically the same
9520: service as @code{AHEAD}, but it does not create a control-flow stack
9521: entry. Therefore the information has to be stored elsewhere;
9522: traditionally, the information was stored in the target fields of the
9523: branches created by the @code{LEAVE}s, by organizing these fields into a
9524: linked list. Unfortunately, this clever trick does not provide enough
9525: space for storing our extended control flow information. Therefore, we
9526: introduce another stack, the leave stack. It contains the control-flow
9527: stack entries for all unresolved @code{LEAVE}s.
9528:
9529: Local names are kept until the end of the colon definition, even if
9530: they are no longer visible in any control-flow path. In a few cases
9531: this may lead to increased space needs for the locals name area, but
9532: usually less than reclaiming this space would cost in code size.
9533:
9534:
9535: @node ANS Forth locals, , Gforth locals, Locals
9536: @subsection ANS Forth locals
9537: @cindex locals, ANS Forth style
9538:
9539: The ANS Forth locals wordset does not define a syntax for locals, but
9540: words that make it possible to define various syntaxes. One of the
9541: possible syntaxes is a subset of the syntax we used in the Gforth locals
9542: wordset, i.e.:
9543:
9544: @example
9545: @{ local1 local2 ... -- comment @}
9546: @end example
9547: @noindent
9548: or
9549: @example
9550: @{ local1 local2 ... @}
9551: @end example
9552:
9553: The order of the locals corresponds to the order in a stack comment. The
9554: restrictions are:
9555:
9556: @itemize @bullet
9557: @item
9558: Locals can only be cell-sized values (no type specifiers are allowed).
9559: @item
9560: Locals can be defined only outside control structures.
9561: @item
9562: Locals can interfere with explicit usage of the return stack. For the
9563: exact (and long) rules, see the standard. If you don't use return stack
9564: accessing words in a definition using locals, you will be all right. The
9565: purpose of this rule is to make locals implementation on the return
9566: stack easier.
9567: @item
9568: The whole definition must be in one line.
9569: @end itemize
9570:
9571: Locals defined in ANS Forth behave like @code{VALUE}s
9572: (@pxref{Values}). I.e., they are initialized from the stack. Using their
9573: name produces their value. Their value can be changed using @code{TO}.
9574:
9575: Since the syntax above is supported by Gforth directly, you need not do
9576: anything to use it. If you want to port a program using this syntax to
9577: another ANS Forth system, use @file{compat/anslocal.fs} to implement the
9578: syntax on the other system.
9579:
9580: Note that a syntax shown in the standard, section A.13 looks
9581: similar, but is quite different in having the order of locals
9582: reversed. Beware!
9583:
9584: The ANS Forth locals wordset itself consists of one word:
9585:
9586: doc-(local)
9587:
9588: The ANS Forth locals extension wordset defines a syntax using
9589: @code{locals|}, but it is so awful that we strongly recommend not to use
9590: it. We have implemented this syntax to make porting to Gforth easy, but
9591: do not document it here. The problem with this syntax is that the locals
9592: are defined in an order reversed with respect to the standard stack
9593: comment notation, making programs harder to read, and easier to misread
9594: and miswrite. The only merit of this syntax is that it is easy to
9595: implement using the ANS Forth locals wordset.
9596:
9597:
9598: @c ----------------------------------------------------------
9599: @node Structures, Object-oriented Forth, Locals, Words
9600: @section Structures
9601: @cindex structures
9602: @cindex records
9603:
9604: This section presents the structure package that comes with Gforth. A
9605: version of the package implemented in ANS Forth is available in
9606: @file{compat/struct.fs}. This package was inspired by a posting on
9607: comp.lang.forth in 1989 (unfortunately I don't remember, by whom;
9608: possibly John Hayes). A version of this section has been published in
9609: M. Anton Ertl,
9610: @uref{http://www.complang.tuwien.ac.at/forth/objects/structs.html, Yet
9611: Another Forth Structures Package}, Forth Dimensions 19(3), pages
9612: 13--16. Marcel Hendrix provided helpful comments.
9613:
9614: @menu
9615: * Why explicit structure support?::
9616: * Structure Usage::
9617: * Structure Naming Convention::
9618: * Structure Implementation::
9619: * Structure Glossary::
9620: @end menu
9621:
9622: @node Why explicit structure support?, Structure Usage, Structures, Structures
9623: @subsection Why explicit structure support?
9624:
9625: @cindex address arithmetic for structures
9626: @cindex structures using address arithmetic
9627: If we want to use a structure containing several fields, we could simply
9628: reserve memory for it, and access the fields using address arithmetic
9629: (@pxref{Address arithmetic}). As an example, consider a structure with
9630: the following fields
9631:
9632: @table @code
9633: @item a
9634: is a float
9635: @item b
9636: is a cell
9637: @item c
9638: is a float
9639: @end table
9640:
9641: Given the (float-aligned) base address of the structure we get the
9642: address of the field
9643:
9644: @table @code
9645: @item a
9646: without doing anything further.
9647: @item b
9648: with @code{float+}
9649: @item c
9650: with @code{float+ cell+ faligned}
9651: @end table
9652:
9653: It is easy to see that this can become quite tiring.
9654:
9655: Moreover, it is not very readable, because seeing a
9656: @code{cell+} tells us neither which kind of structure is
9657: accessed nor what field is accessed; we have to somehow infer the kind
9658: of structure, and then look up in the documentation, which field of
9659: that structure corresponds to that offset.
9660:
9661: Finally, this kind of address arithmetic also causes maintenance
9662: troubles: If you add or delete a field somewhere in the middle of the
9663: structure, you have to find and change all computations for the fields
9664: afterwards.
9665:
9666: So, instead of using @code{cell+} and friends directly, how
9667: about storing the offsets in constants:
9668:
9669: @example
9670: 0 constant a-offset
9671: 0 float+ constant b-offset
9672: 0 float+ cell+ faligned c-offset
9673: @end example
9674:
9675: Now we can get the address of field @code{x} with @code{x-offset
9676: +}. This is much better in all respects. Of course, you still
9677: have to change all later offset definitions if you add a field. You can
9678: fix this by declaring the offsets in the following way:
9679:
9680: @example
9681: 0 constant a-offset
9682: a-offset float+ constant b-offset
9683: b-offset cell+ faligned constant c-offset
9684: @end example
9685:
9686: Since we always use the offsets with @code{+}, we could use a defining
9687: word @code{cfield} that includes the @code{+} in the action of the
9688: defined word:
9689:
9690: @example
9691: : cfield ( n "name" -- )
9692: create ,
9693: does> ( name execution: addr1 -- addr2 )
9694: @@ + ;
9695:
9696: 0 cfield a
9697: 0 a float+ cfield b
9698: 0 b cell+ faligned cfield c
9699: @end example
9700:
9701: Instead of @code{x-offset +}, we now simply write @code{x}.
9702:
9703: The structure field words now can be used quite nicely. However,
9704: their definition is still a bit cumbersome: We have to repeat the
9705: name, the information about size and alignment is distributed before
9706: and after the field definitions etc. The structure package presented
9707: here addresses these problems.
9708:
9709: @node Structure Usage, Structure Naming Convention, Why explicit structure support?, Structures
9710: @subsection Structure Usage
9711: @cindex structure usage
9712:
9713: @cindex @code{field} usage
9714: @cindex @code{struct} usage
9715: @cindex @code{end-struct} usage
9716: You can define a structure for a (data-less) linked list with:
9717: @example
9718: struct
9719: cell% field list-next
9720: end-struct list%
9721: @end example
9722:
9723: With the address of the list node on the stack, you can compute the
9724: address of the field that contains the address of the next node with
9725: @code{list-next}. E.g., you can determine the length of a list
9726: with:
9727:
9728: @example
9729: : list-length ( list -- n )
9730: \ "list" is a pointer to the first element of a linked list
9731: \ "n" is the length of the list
9732: 0 BEGIN ( list1 n1 )
9733: over
9734: WHILE ( list1 n1 )
9735: 1+ swap list-next @@ swap
9736: REPEAT
9737: nip ;
9738: @end example
9739:
9740: You can reserve memory for a list node in the dictionary with
9741: @code{list% %allot}, which leaves the address of the list node on the
9742: stack. For the equivalent allocation on the heap you can use @code{list%
9743: %alloc} (or, for an @code{allocate}-like stack effect (i.e., with ior),
9744: use @code{list% %allocate}). You can get the the size of a list
9745: node with @code{list% %size} and its alignment with @code{list%
9746: %alignment}.
9747:
9748: Note that in ANS Forth the body of a @code{create}d word is
9749: @code{aligned} but not necessarily @code{faligned};
9750: therefore, if you do a:
9751:
9752: @example
9753: create @emph{name} foo% %allot drop
9754: @end example
9755:
9756: @noindent
9757: then the memory alloted for @code{foo%} is guaranteed to start at the
9758: body of @code{@emph{name}} only if @code{foo%} contains only character,
9759: cell and double fields. Therefore, if your structure contains floats,
9760: better use
9761:
9762: @example
9763: foo% %allot constant @emph{name}
9764: @end example
9765:
9766: @cindex structures containing structures
9767: You can include a structure @code{foo%} as a field of
9768: another structure, like this:
9769: @example
9770: struct
9771: ...
9772: foo% field ...
9773: ...
9774: end-struct ...
9775: @end example
9776:
9777: @cindex structure extension
9778: @cindex extended records
9779: Instead of starting with an empty structure, you can extend an
9780: existing structure. E.g., a plain linked list without data, as defined
9781: above, is hardly useful; You can extend it to a linked list of integers,
9782: like this:@footnote{This feature is also known as @emph{extended
9783: records}. It is the main innovation in the Oberon language; in other
9784: words, adding this feature to Modula-2 led Wirth to create a new
9785: language, write a new compiler etc. Adding this feature to Forth just
9786: required a few lines of code.}
9787:
9788: @example
9789: list%
9790: cell% field intlist-int
9791: end-struct intlist%
9792: @end example
9793:
9794: @code{intlist%} is a structure with two fields:
9795: @code{list-next} and @code{intlist-int}.
9796:
9797: @cindex structures containing arrays
9798: You can specify an array type containing @emph{n} elements of
9799: type @code{foo%} like this:
9800:
9801: @example
9802: foo% @emph{n} *
9803: @end example
9804:
9805: You can use this array type in any place where you can use a normal
9806: type, e.g., when defining a @code{field}, or with
9807: @code{%allot}.
9808:
9809: @cindex first field optimization
9810: The first field is at the base address of a structure and the word for
9811: this field (e.g., @code{list-next}) actually does not change the address
9812: on the stack. You may be tempted to leave it away in the interest of
9813: run-time and space efficiency. This is not necessary, because the
9814: structure package optimizes this case: If you compile a first-field
9815: words, no code is generated. So, in the interest of readability and
9816: maintainability you should include the word for the field when accessing
9817: the field.
9818:
9819:
9820: @node Structure Naming Convention, Structure Implementation, Structure Usage, Structures
9821: @subsection Structure Naming Convention
9822: @cindex structure naming convention
9823:
9824: The field names that come to (my) mind are often quite generic, and,
9825: if used, would cause frequent name clashes. E.g., many structures
9826: probably contain a @code{counter} field. The structure names
9827: that come to (my) mind are often also the logical choice for the names
9828: of words that create such a structure.
9829:
9830: Therefore, I have adopted the following naming conventions:
9831:
9832: @itemize @bullet
9833: @cindex field naming convention
9834: @item
9835: The names of fields are of the form
9836: @code{@emph{struct}-@emph{field}}, where
9837: @code{@emph{struct}} is the basic name of the structure, and
9838: @code{@emph{field}} is the basic name of the field. You can
9839: think of field words as converting the (address of the)
9840: structure into the (address of the) field.
9841:
9842: @cindex structure naming convention
9843: @item
9844: The names of structures are of the form
9845: @code{@emph{struct}%}, where
9846: @code{@emph{struct}} is the basic name of the structure.
9847: @end itemize
9848:
9849: This naming convention does not work that well for fields of extended
9850: structures; e.g., the integer list structure has a field
9851: @code{intlist-int}, but has @code{list-next}, not
9852: @code{intlist-next}.
9853:
9854: @node Structure Implementation, Structure Glossary, Structure Naming Convention, Structures
9855: @subsection Structure Implementation
9856: @cindex structure implementation
9857: @cindex implementation of structures
9858:
9859: The central idea in the implementation is to pass the data about the
9860: structure being built on the stack, not in some global
9861: variable. Everything else falls into place naturally once this design
9862: decision is made.
9863:
9864: The type description on the stack is of the form @emph{align
9865: size}. Keeping the size on the top-of-stack makes dealing with arrays
9866: very simple.
9867:
9868: @code{field} is a defining word that uses @code{Create}
9869: and @code{DOES>}. The body of the field contains the offset
9870: of the field, and the normal @code{DOES>} action is simply:
9871:
9872: @example
9873: @@ +
9874: @end example
9875:
9876: @noindent
9877: i.e., add the offset to the address, giving the stack effect
9878: @i{addr1 -- addr2} for a field.
9879:
9880: @cindex first field optimization, implementation
9881: This simple structure is slightly complicated by the optimization
9882: for fields with offset 0, which requires a different
9883: @code{DOES>}-part (because we cannot rely on there being
9884: something on the stack if such a field is invoked during
9885: compilation). Therefore, we put the different @code{DOES>}-parts
9886: in separate words, and decide which one to invoke based on the
9887: offset. For a zero offset, the field is basically a noop; it is
9888: immediate, and therefore no code is generated when it is compiled.
9889:
9890: @node Structure Glossary, , Structure Implementation, Structures
9891: @subsection Structure Glossary
9892: @cindex structure glossary
9893:
9894:
9895: doc-%align
9896: doc-%alignment
9897: doc-%alloc
9898: doc-%allocate
9899: doc-%allot
9900: doc-cell%
9901: doc-char%
9902: doc-dfloat%
9903: doc-double%
9904: doc-end-struct
9905: doc-field
9906: doc-float%
9907: doc-naligned
9908: doc-sfloat%
9909: doc-%size
9910: doc-struct
9911:
9912:
9913: @c -------------------------------------------------------------
9914: @node Object-oriented Forth, Programming Tools, Structures, Words
9915: @section Object-oriented Forth
9916:
9917: Gforth comes with three packages for object-oriented programming:
9918: @file{objects.fs}, @file{oof.fs}, and @file{mini-oof.fs}; none of them
9919: is preloaded, so you have to @code{include} them before use. The most
9920: important differences between these packages (and others) are discussed
9921: in @ref{Comparison with other object models}. All packages are written
9922: in ANS Forth and can be used with any other ANS Forth.
9923:
9924: @menu
9925: * Why object-oriented programming?::
9926: * Object-Oriented Terminology::
9927: * Objects::
9928: * OOF::
9929: * Mini-OOF::
9930: * Comparison with other object models::
9931: @end menu
9932:
9933: @c ----------------------------------------------------------------
9934: @node Why object-oriented programming?, Object-Oriented Terminology, Object-oriented Forth, Object-oriented Forth
9935: @subsection Why object-oriented programming?
9936: @cindex object-oriented programming motivation
9937: @cindex motivation for object-oriented programming
9938:
9939: Often we have to deal with several data structures (@emph{objects}),
9940: that have to be treated similarly in some respects, but differently in
9941: others. Graphical objects are the textbook example: circles, triangles,
9942: dinosaurs, icons, and others, and we may want to add more during program
9943: development. We want to apply some operations to any graphical object,
9944: e.g., @code{draw} for displaying it on the screen. However, @code{draw}
9945: has to do something different for every kind of object.
9946: @comment TODO add some other operations eg perimeter, area
9947: @comment and tie in to concrete examples later..
9948:
9949: We could implement @code{draw} as a big @code{CASE}
9950: control structure that executes the appropriate code depending on the
9951: kind of object to be drawn. This would be not be very elegant, and,
9952: moreover, we would have to change @code{draw} every time we add
9953: a new kind of graphical object (say, a spaceship).
9954:
9955: What we would rather do is: When defining spaceships, we would tell
9956: the system: ``Here's how you @code{draw} a spaceship; you figure
9957: out the rest''.
9958:
9959: This is the problem that all systems solve that (rightfully) call
9960: themselves object-oriented; the object-oriented packages presented here
9961: solve this problem (and not much else).
9962: @comment TODO ?list properties of oo systems.. oo vs o-based?
9963:
9964: @c ------------------------------------------------------------------------
9965: @node Object-Oriented Terminology, Objects, Why object-oriented programming?, Object-oriented Forth
9966: @subsection Object-Oriented Terminology
9967: @cindex object-oriented terminology
9968: @cindex terminology for object-oriented programming
9969:
9970: This section is mainly for reference, so you don't have to understand
9971: all of it right away. The terminology is mainly Smalltalk-inspired. In
9972: short:
9973:
9974: @table @emph
9975: @cindex class
9976: @item class
9977: a data structure definition with some extras.
9978:
9979: @cindex object
9980: @item object
9981: an instance of the data structure described by the class definition.
9982:
9983: @cindex instance variables
9984: @item instance variables
9985: fields of the data structure.
9986:
9987: @cindex selector
9988: @cindex method selector
9989: @cindex virtual function
9990: @item selector
9991: (or @emph{method selector}) a word (e.g.,
9992: @code{draw}) that performs an operation on a variety of data
9993: structures (classes). A selector describes @emph{what} operation to
9994: perform. In C++ terminology: a (pure) virtual function.
9995:
9996: @cindex method
9997: @item method
9998: the concrete definition that performs the operation
9999: described by the selector for a specific class. A method specifies
10000: @emph{how} the operation is performed for a specific class.
10001:
10002: @cindex selector invocation
10003: @cindex message send
10004: @cindex invoking a selector
10005: @item selector invocation
10006: a call of a selector. One argument of the call (the TOS (top-of-stack))
10007: is used for determining which method is used. In Smalltalk terminology:
10008: a message (consisting of the selector and the other arguments) is sent
10009: to the object.
10010:
10011: @cindex receiving object
10012: @item receiving object
10013: the object used for determining the method executed by a selector
10014: invocation. In the @file{objects.fs} model, it is the object that is on
10015: the TOS when the selector is invoked. (@emph{Receiving} comes from
10016: the Smalltalk @emph{message} terminology.)
10017:
10018: @cindex child class
10019: @cindex parent class
10020: @cindex inheritance
10021: @item child class
10022: a class that has (@emph{inherits}) all properties (instance variables,
10023: selectors, methods) from a @emph{parent class}. In Smalltalk
10024: terminology: The subclass inherits from the superclass. In C++
10025: terminology: The derived class inherits from the base class.
10026:
10027: @end table
10028:
10029: @c If you wonder about the message sending terminology, it comes from
10030: @c a time when each object had it's own task and objects communicated via
10031: @c message passing; eventually the Smalltalk developers realized that
10032: @c they can do most things through simple (indirect) calls. They kept the
10033: @c terminology.
10034:
10035: @c --------------------------------------------------------------
10036: @node Objects, OOF, Object-Oriented Terminology, Object-oriented Forth
10037: @subsection The @file{objects.fs} model
10038: @cindex objects
10039: @cindex object-oriented programming
10040:
10041: @cindex @file{objects.fs}
10042: @cindex @file{oof.fs}
10043:
10044: This section describes the @file{objects.fs} package. This material also
10045: has been published in M. Anton Ertl,
10046: @cite{@uref{http://www.complang.tuwien.ac.at/forth/objects/objects.html,
10047: Yet Another Forth Objects Package}}, Forth Dimensions 19(2), pages
10048: 37--43.
10049: @c McKewan's and Zsoter's packages
10050:
10051: This section assumes that you have read @ref{Structures}.
10052:
10053: The techniques on which this model is based have been used to implement
10054: the parser generator, Gray, and have also been used in Gforth for
10055: implementing the various flavours of word lists (hashed or not,
10056: case-sensitive or not, special-purpose word lists for locals etc.).
10057:
10058:
10059: @menu
10060: * Properties of the Objects model::
10061: * Basic Objects Usage::
10062: * The Objects base class::
10063: * Creating objects::
10064: * Object-Oriented Programming Style::
10065: * Class Binding::
10066: * Method conveniences::
10067: * Classes and Scoping::
10068: * Dividing classes::
10069: * Object Interfaces::
10070: * Objects Implementation::
10071: * Objects Glossary::
10072: @end menu
10073:
10074: Marcel Hendrix provided helpful comments on this section.
10075:
10076: @node Properties of the Objects model, Basic Objects Usage, Objects, Objects
10077: @subsubsection Properties of the @file{objects.fs} model
10078: @cindex @file{objects.fs} properties
10079:
10080: @itemize @bullet
10081: @item
10082: It is straightforward to pass objects on the stack. Passing
10083: selectors on the stack is a little less convenient, but possible.
10084:
10085: @item
10086: Objects are just data structures in memory, and are referenced by their
10087: address. You can create words for objects with normal defining words
10088: like @code{constant}. Likewise, there is no difference between instance
10089: variables that contain objects and those that contain other data.
10090:
10091: @item
10092: Late binding is efficient and easy to use.
10093:
10094: @item
10095: It avoids parsing, and thus avoids problems with state-smartness
10096: and reduced extensibility; for convenience there are a few parsing
10097: words, but they have non-parsing counterparts. There are also a few
10098: defining words that parse. This is hard to avoid, because all standard
10099: defining words parse (except @code{:noname}); however, such
10100: words are not as bad as many other parsing words, because they are not
10101: state-smart.
10102:
10103: @item
10104: It does not try to incorporate everything. It does a few things and does
10105: them well (IMO). In particular, this model was not designed to support
10106: information hiding (although it has features that may help); you can use
10107: a separate package for achieving this.
10108:
10109: @item
10110: It is layered; you don't have to learn and use all features to use this
10111: model. Only a few features are necessary (@pxref{Basic Objects Usage},
10112: @pxref{The Objects base class}, @pxref{Creating objects}.), the others
10113: are optional and independent of each other.
10114:
10115: @item
10116: An implementation in ANS Forth is available.
10117:
10118: @end itemize
10119:
10120:
10121: @node Basic Objects Usage, The Objects base class, Properties of the Objects model, Objects
10122: @subsubsection Basic @file{objects.fs} Usage
10123: @cindex basic objects usage
10124: @cindex objects, basic usage
10125:
10126: You can define a class for graphical objects like this:
10127:
10128: @cindex @code{class} usage
10129: @cindex @code{end-class} usage
10130: @cindex @code{selector} usage
10131: @example
10132: object class \ "object" is the parent class
10133: selector draw ( x y graphical -- )
10134: end-class graphical
10135: @end example
10136:
10137: This code defines a class @code{graphical} with an
10138: operation @code{draw}. We can perform the operation
10139: @code{draw} on any @code{graphical} object, e.g.:
10140:
10141: @example
10142: 100 100 t-rex draw
10143: @end example
10144:
10145: @noindent
10146: where @code{t-rex} is a word (say, a constant) that produces a
10147: graphical object.
10148:
10149: @comment TODO add a 2nd operation eg perimeter.. and use for
10150: @comment a concrete example
10151:
10152: @cindex abstract class
10153: How do we create a graphical object? With the present definitions,
10154: we cannot create a useful graphical object. The class
10155: @code{graphical} describes graphical objects in general, but not
10156: any concrete graphical object type (C++ users would call it an
10157: @emph{abstract class}); e.g., there is no method for the selector
10158: @code{draw} in the class @code{graphical}.
10159:
10160: For concrete graphical objects, we define child classes of the
10161: class @code{graphical}, e.g.:
10162:
10163: @cindex @code{overrides} usage
10164: @cindex @code{field} usage in class definition
10165: @example
10166: graphical class \ "graphical" is the parent class
10167: cell% field circle-radius
10168:
10169: :noname ( x y circle -- )
10170: circle-radius @@ draw-circle ;
10171: overrides draw
10172:
10173: :noname ( n-radius circle -- )
10174: circle-radius ! ;
10175: overrides construct
10176:
10177: end-class circle
10178: @end example
10179:
10180: Here we define a class @code{circle} as a child of @code{graphical},
10181: with field @code{circle-radius} (which behaves just like a field
10182: (@pxref{Structures}); it defines (using @code{overrides}) new methods
10183: for the selectors @code{draw} and @code{construct} (@code{construct} is
10184: defined in @code{object}, the parent class of @code{graphical}).
10185:
10186: Now we can create a circle on the heap (i.e.,
10187: @code{allocate}d memory) with:
10188:
10189: @cindex @code{heap-new} usage
10190: @example
10191: 50 circle heap-new constant my-circle
10192: @end example
10193:
10194: @noindent
10195: @code{heap-new} invokes @code{construct}, thus
10196: initializing the field @code{circle-radius} with 50. We can draw
10197: this new circle at (100,100) with:
10198:
10199: @example
10200: 100 100 my-circle draw
10201: @end example
10202:
10203: @cindex selector invocation, restrictions
10204: @cindex class definition, restrictions
10205: Note: You can only invoke a selector if the object on the TOS
10206: (the receiving object) belongs to the class where the selector was
10207: defined or one of its descendents; e.g., you can invoke
10208: @code{draw} only for objects belonging to @code{graphical}
10209: or its descendents (e.g., @code{circle}). Immediately before
10210: @code{end-class}, the search order has to be the same as
10211: immediately after @code{class}.
10212:
10213: @node The Objects base class, Creating objects, Basic Objects Usage, Objects
10214: @subsubsection The @file{object.fs} base class
10215: @cindex @code{object} class
10216:
10217: When you define a class, you have to specify a parent class. So how do
10218: you start defining classes? There is one class available from the start:
10219: @code{object}. It is ancestor for all classes and so is the
10220: only class that has no parent. It has two selectors: @code{construct}
10221: and @code{print}.
10222:
10223: @node Creating objects, Object-Oriented Programming Style, The Objects base class, Objects
10224: @subsubsection Creating objects
10225: @cindex creating objects
10226: @cindex object creation
10227: @cindex object allocation options
10228:
10229: @cindex @code{heap-new} discussion
10230: @cindex @code{dict-new} discussion
10231: @cindex @code{construct} discussion
10232: You can create and initialize an object of a class on the heap with
10233: @code{heap-new} ( ... class -- object ) and in the dictionary
10234: (allocation with @code{allot}) with @code{dict-new} (
10235: ... class -- object ). Both words invoke @code{construct}, which
10236: consumes the stack items indicated by "..." above.
10237:
10238: @cindex @code{init-object} discussion
10239: @cindex @code{class-inst-size} discussion
10240: If you want to allocate memory for an object yourself, you can get its
10241: alignment and size with @code{class-inst-size 2@@} ( class --
10242: align size ). Once you have memory for an object, you can initialize
10243: it with @code{init-object} ( ... class object -- );
10244: @code{construct} does only a part of the necessary work.
10245:
10246: @node Object-Oriented Programming Style, Class Binding, Creating objects, Objects
10247: @subsubsection Object-Oriented Programming Style
10248: @cindex object-oriented programming style
10249: @cindex programming style, object-oriented
10250:
10251: This section is not exhaustive.
10252:
10253: @cindex stack effects of selectors
10254: @cindex selectors and stack effects
10255: In general, it is a good idea to ensure that all methods for the
10256: same selector have the same stack effect: when you invoke a selector,
10257: you often have no idea which method will be invoked, so, unless all
10258: methods have the same stack effect, you will not know the stack effect
10259: of the selector invocation.
10260:
10261: One exception to this rule is methods for the selector
10262: @code{construct}. We know which method is invoked, because we
10263: specify the class to be constructed at the same place. Actually, I
10264: defined @code{construct} as a selector only to give the users a
10265: convenient way to specify initialization. The way it is used, a
10266: mechanism different from selector invocation would be more natural
10267: (but probably would take more code and more space to explain).
10268:
10269: @node Class Binding, Method conveniences, Object-Oriented Programming Style, Objects
10270: @subsubsection Class Binding
10271: @cindex class binding
10272: @cindex early binding
10273:
10274: @cindex late binding
10275: Normal selector invocations determine the method at run-time depending
10276: on the class of the receiving object. This run-time selection is called
10277: @i{late binding}.
10278:
10279: Sometimes it's preferable to invoke a different method. For example,
10280: you might want to use the simple method for @code{print}ing
10281: @code{object}s instead of the possibly long-winded @code{print} method
10282: of the receiver class. You can achieve this by replacing the invocation
10283: of @code{print} with:
10284:
10285: @cindex @code{[bind]} usage
10286: @example
10287: [bind] object print
10288: @end example
10289:
10290: @noindent
10291: in compiled code or:
10292:
10293: @cindex @code{bind} usage
10294: @example
10295: bind object print
10296: @end example
10297:
10298: @cindex class binding, alternative to
10299: @noindent
10300: in interpreted code. Alternatively, you can define the method with a
10301: name (e.g., @code{print-object}), and then invoke it through the
10302: name. Class binding is just a (often more convenient) way to achieve
10303: the same effect; it avoids name clutter and allows you to invoke
10304: methods directly without naming them first.
10305:
10306: @cindex superclass binding
10307: @cindex parent class binding
10308: A frequent use of class binding is this: When we define a method
10309: for a selector, we often want the method to do what the selector does
10310: in the parent class, and a little more. There is a special word for
10311: this purpose: @code{[parent]}; @code{[parent]
10312: @emph{selector}} is equivalent to @code{[bind] @emph{parent
10313: selector}}, where @code{@emph{parent}} is the parent
10314: class of the current class. E.g., a method definition might look like:
10315:
10316: @cindex @code{[parent]} usage
10317: @example
10318: :noname
10319: dup [parent] foo \ do parent's foo on the receiving object
10320: ... \ do some more
10321: ; overrides foo
10322: @end example
10323:
10324: @cindex class binding as optimization
10325: In @cite{Object-oriented programming in ANS Forth} (Forth Dimensions,
10326: March 1997), Andrew McKewan presents class binding as an optimization
10327: technique. I recommend not using it for this purpose unless you are in
10328: an emergency. Late binding is pretty fast with this model anyway, so the
10329: benefit of using class binding is small; the cost of using class binding
10330: where it is not appropriate is reduced maintainability.
10331:
10332: While we are at programming style questions: You should bind
10333: selectors only to ancestor classes of the receiving object. E.g., say,
10334: you know that the receiving object is of class @code{foo} or its
10335: descendents; then you should bind only to @code{foo} and its
10336: ancestors.
10337:
10338: @node Method conveniences, Classes and Scoping, Class Binding, Objects
10339: @subsubsection Method conveniences
10340: @cindex method conveniences
10341:
10342: In a method you usually access the receiving object pretty often. If
10343: you define the method as a plain colon definition (e.g., with
10344: @code{:noname}), you may have to do a lot of stack
10345: gymnastics. To avoid this, you can define the method with @code{m:
10346: ... ;m}. E.g., you could define the method for
10347: @code{draw}ing a @code{circle} with
10348:
10349: @cindex @code{this} usage
10350: @cindex @code{m:} usage
10351: @cindex @code{;m} usage
10352: @example
10353: m: ( x y circle -- )
10354: ( x y ) this circle-radius @@ draw-circle ;m
10355: @end example
10356:
10357: @cindex @code{exit} in @code{m: ... ;m}
10358: @cindex @code{exitm} discussion
10359: @cindex @code{catch} in @code{m: ... ;m}
10360: When this method is executed, the receiver object is removed from the
10361: stack; you can access it with @code{this} (admittedly, in this
10362: example the use of @code{m: ... ;m} offers no advantage). Note
10363: that I specify the stack effect for the whole method (i.e. including
10364: the receiver object), not just for the code between @code{m:}
10365: and @code{;m}. You cannot use @code{exit} in
10366: @code{m:...;m}; instead, use
10367: @code{exitm}.@footnote{Moreover, for any word that calls
10368: @code{catch} and was defined before loading
10369: @code{objects.fs}, you have to redefine it like I redefined
10370: @code{catch}: @code{: catch this >r catch r> to-this ;}}
10371:
10372: @cindex @code{inst-var} usage
10373: You will frequently use sequences of the form @code{this
10374: @emph{field}} (in the example above: @code{this
10375: circle-radius}). If you use the field only in this way, you can
10376: define it with @code{inst-var} and eliminate the
10377: @code{this} before the field name. E.g., the @code{circle}
10378: class above could also be defined with:
10379:
10380: @example
10381: graphical class
10382: cell% inst-var radius
10383:
10384: m: ( x y circle -- )
10385: radius @@ draw-circle ;m
10386: overrides draw
10387:
10388: m: ( n-radius circle -- )
10389: radius ! ;m
10390: overrides construct
10391:
10392: end-class circle
10393: @end example
10394:
10395: @code{radius} can only be used in @code{circle} and its
10396: descendent classes and inside @code{m:...;m}.
10397:
10398: @cindex @code{inst-value} usage
10399: You can also define fields with @code{inst-value}, which is
10400: to @code{inst-var} what @code{value} is to
10401: @code{variable}. You can change the value of such a field with
10402: @code{[to-inst]}. E.g., we could also define the class
10403: @code{circle} like this:
10404:
10405: @example
10406: graphical class
10407: inst-value radius
10408:
10409: m: ( x y circle -- )
10410: radius draw-circle ;m
10411: overrides draw
10412:
10413: m: ( n-radius circle -- )
10414: [to-inst] radius ;m
10415: overrides construct
10416:
10417: end-class circle
10418: @end example
10419:
10420: @c !! :m is easy to confuse with m:. Another name would be better.
10421:
10422: @c Finally, you can define named methods with @code{:m}. One use of this
10423: @c feature is the definition of words that occur only in one class and are
10424: @c not intended to be overridden, but which still need method context
10425: @c (e.g., for accessing @code{inst-var}s). Another use is for methods that
10426: @c would be bound frequently, if defined anonymously.
10427:
10428:
10429: @node Classes and Scoping, Dividing classes, Method conveniences, Objects
10430: @subsubsection Classes and Scoping
10431: @cindex classes and scoping
10432: @cindex scoping and classes
10433:
10434: Inheritance is frequent, unlike structure extension. This exacerbates
10435: the problem with the field name convention (@pxref{Structure Naming
10436: Convention}): One always has to remember in which class the field was
10437: originally defined; changing a part of the class structure would require
10438: changes for renaming in otherwise unaffected code.
10439:
10440: @cindex @code{inst-var} visibility
10441: @cindex @code{inst-value} visibility
10442: To solve this problem, I added a scoping mechanism (which was not in my
10443: original charter): A field defined with @code{inst-var} (or
10444: @code{inst-value}) is visible only in the class where it is defined and in
10445: the descendent classes of this class. Using such fields only makes
10446: sense in @code{m:}-defined methods in these classes anyway.
10447:
10448: This scoping mechanism allows us to use the unadorned field name,
10449: because name clashes with unrelated words become much less likely.
10450:
10451: @cindex @code{protected} discussion
10452: @cindex @code{private} discussion
10453: Once we have this mechanism, we can also use it for controlling the
10454: visibility of other words: All words defined after
10455: @code{protected} are visible only in the current class and its
10456: descendents. @code{public} restores the compilation
10457: (i.e. @code{current}) word list that was in effect before. If you
10458: have several @code{protected}s without an intervening
10459: @code{public} or @code{set-current}, @code{public}
10460: will restore the compilation word list in effect before the first of
10461: these @code{protected}s.
10462:
10463: @node Dividing classes, Object Interfaces, Classes and Scoping, Objects
10464: @subsubsection Dividing classes
10465: @cindex Dividing classes
10466: @cindex @code{methods}...@code{end-methods}
10467:
10468: You may want to do the definition of methods separate from the
10469: definition of the class, its selectors, fields, and instance variables,
10470: i.e., separate the implementation from the definition. You can do this
10471: in the following way:
10472:
10473: @example
10474: graphical class
10475: inst-value radius
10476: end-class circle
10477:
10478: ... \ do some other stuff
10479:
10480: circle methods \ now we are ready
10481:
10482: m: ( x y circle -- )
10483: radius draw-circle ;m
10484: overrides draw
10485:
10486: m: ( n-radius circle -- )
10487: [to-inst] radius ;m
10488: overrides construct
10489:
10490: end-methods
10491: @end example
10492:
10493: You can use several @code{methods}...@code{end-methods} sections. The
10494: only things you can do to the class in these sections are: defining
10495: methods, and overriding the class's selectors. You must not define new
10496: selectors or fields.
10497:
10498: Note that you often have to override a selector before using it. In
10499: particular, you usually have to override @code{construct} with a new
10500: method before you can invoke @code{heap-new} and friends. E.g., you
10501: must not create a circle before the @code{overrides construct} sequence
10502: in the example above.
10503:
10504: @node Object Interfaces, Objects Implementation, Dividing classes, Objects
10505: @subsubsection Object Interfaces
10506: @cindex object interfaces
10507: @cindex interfaces for objects
10508:
10509: In this model you can only call selectors defined in the class of the
10510: receiving objects or in one of its ancestors. If you call a selector
10511: with a receiving object that is not in one of these classes, the
10512: result is undefined; if you are lucky, the program crashes
10513: immediately.
10514:
10515: @cindex selectors common to hardly-related classes
10516: Now consider the case when you want to have a selector (or several)
10517: available in two classes: You would have to add the selector to a
10518: common ancestor class, in the worst case to @code{object}. You
10519: may not want to do this, e.g., because someone else is responsible for
10520: this ancestor class.
10521:
10522: The solution for this problem is interfaces. An interface is a
10523: collection of selectors. If a class implements an interface, the
10524: selectors become available to the class and its descendents. A class
10525: can implement an unlimited number of interfaces. For the problem
10526: discussed above, we would define an interface for the selector(s), and
10527: both classes would implement the interface.
10528:
10529: As an example, consider an interface @code{storage} for
10530: writing objects to disk and getting them back, and a class
10531: @code{foo} that implements it. The code would look like this:
10532:
10533: @cindex @code{interface} usage
10534: @cindex @code{end-interface} usage
10535: @cindex @code{implementation} usage
10536: @example
10537: interface
10538: selector write ( file object -- )
10539: selector read1 ( file object -- )
10540: end-interface storage
10541:
10542: bar class
10543: storage implementation
10544:
10545: ... overrides write
10546: ... overrides read1
10547: ...
10548: end-class foo
10549: @end example
10550:
10551: @noindent
10552: (I would add a word @code{read} @i{( file -- object )} that uses
10553: @code{read1} internally, but that's beyond the point illustrated
10554: here.)
10555:
10556: Note that you cannot use @code{protected} in an interface; and
10557: of course you cannot define fields.
10558:
10559: In the Neon model, all selectors are available for all classes;
10560: therefore it does not need interfaces. The price you pay in this model
10561: is slower late binding, and therefore, added complexity to avoid late
10562: binding.
10563:
10564: @node Objects Implementation, Objects Glossary, Object Interfaces, Objects
10565: @subsubsection @file{objects.fs} Implementation
10566: @cindex @file{objects.fs} implementation
10567:
10568: @cindex @code{object-map} discussion
10569: An object is a piece of memory, like one of the data structures
10570: described with @code{struct...end-struct}. It has a field
10571: @code{object-map} that points to the method map for the object's
10572: class.
10573:
10574: @cindex method map
10575: @cindex virtual function table
10576: The @emph{method map}@footnote{This is Self terminology; in C++
10577: terminology: virtual function table.} is an array that contains the
10578: execution tokens (@i{xt}s) of the methods for the object's class. Each
10579: selector contains an offset into a method map.
10580:
10581: @cindex @code{selector} implementation, class
10582: @code{selector} is a defining word that uses
10583: @code{CREATE} and @code{DOES>}. The body of the
10584: selector contains the offset; the @code{DOES>} action for a
10585: class selector is, basically:
10586:
10587: @example
10588: ( object addr ) @@ over object-map @@ + @@ execute
10589: @end example
10590:
10591: Since @code{object-map} is the first field of the object, it
10592: does not generate any code. As you can see, calling a selector has a
10593: small, constant cost.
10594:
10595: @cindex @code{current-interface} discussion
10596: @cindex class implementation and representation
10597: A class is basically a @code{struct} combined with a method
10598: map. During the class definition the alignment and size of the class
10599: are passed on the stack, just as with @code{struct}s, so
10600: @code{field} can also be used for defining class
10601: fields. However, passing more items on the stack would be
10602: inconvenient, so @code{class} builds a data structure in memory,
10603: which is accessed through the variable
10604: @code{current-interface}. After its definition is complete, the
10605: class is represented on the stack by a pointer (e.g., as parameter for
10606: a child class definition).
10607:
10608: A new class starts off with the alignment and size of its parent,
10609: and a copy of the parent's method map. Defining new fields extends the
10610: size and alignment; likewise, defining new selectors extends the
10611: method map. @code{overrides} just stores a new @i{xt} in the method
10612: map at the offset given by the selector.
10613:
10614: @cindex class binding, implementation
10615: Class binding just gets the @i{xt} at the offset given by the selector
10616: from the class's method map and @code{compile,}s (in the case of
10617: @code{[bind]}) it.
10618:
10619: @cindex @code{this} implementation
10620: @cindex @code{catch} and @code{this}
10621: @cindex @code{this} and @code{catch}
10622: I implemented @code{this} as a @code{value}. At the
10623: start of an @code{m:...;m} method the old @code{this} is
10624: stored to the return stack and restored at the end; and the object on
10625: the TOS is stored @code{TO this}. This technique has one
10626: disadvantage: If the user does not leave the method via
10627: @code{;m}, but via @code{throw} or @code{exit},
10628: @code{this} is not restored (and @code{exit} may
10629: crash). To deal with the @code{throw} problem, I have redefined
10630: @code{catch} to save and restore @code{this}; the same
10631: should be done with any word that can catch an exception. As for
10632: @code{exit}, I simply forbid it (as a replacement, there is
10633: @code{exitm}).
10634:
10635: @cindex @code{inst-var} implementation
10636: @code{inst-var} is just the same as @code{field}, with
10637: a different @code{DOES>} action:
10638: @example
10639: @@ this +
10640: @end example
10641: Similar for @code{inst-value}.
10642:
10643: @cindex class scoping implementation
10644: Each class also has a word list that contains the words defined with
10645: @code{inst-var} and @code{inst-value}, and its protected
10646: words. It also has a pointer to its parent. @code{class} pushes
10647: the word lists of the class and all its ancestors onto the search order stack,
10648: and @code{end-class} drops them.
10649:
10650: @cindex interface implementation
10651: An interface is like a class without fields, parent and protected
10652: words; i.e., it just has a method map. If a class implements an
10653: interface, its method map contains a pointer to the method map of the
10654: interface. The positive offsets in the map are reserved for class
10655: methods, therefore interface map pointers have negative
10656: offsets. Interfaces have offsets that are unique throughout the
10657: system, unlike class selectors, whose offsets are only unique for the
10658: classes where the selector is available (invokable).
10659:
10660: This structure means that interface selectors have to perform one
10661: indirection more than class selectors to find their method. Their body
10662: contains the interface map pointer offset in the class method map, and
10663: the method offset in the interface method map. The
10664: @code{does>} action for an interface selector is, basically:
10665:
10666: @example
10667: ( object selector-body )
10668: 2dup selector-interface @@ ( object selector-body object interface-offset )
10669: swap object-map @@ + @@ ( object selector-body map )
10670: swap selector-offset @@ + @@ execute
10671: @end example
10672:
10673: where @code{object-map} and @code{selector-offset} are
10674: first fields and generate no code.
10675:
10676: As a concrete example, consider the following code:
10677:
10678: @example
10679: interface
10680: selector if1sel1
10681: selector if1sel2
10682: end-interface if1
10683:
10684: object class
10685: if1 implementation
10686: selector cl1sel1
10687: cell% inst-var cl1iv1
10688:
10689: ' m1 overrides construct
10690: ' m2 overrides if1sel1
10691: ' m3 overrides if1sel2
10692: ' m4 overrides cl1sel2
10693: end-class cl1
10694:
10695: create obj1 object dict-new drop
10696: create obj2 cl1 dict-new drop
10697: @end example
10698:
10699: The data structure created by this code (including the data structure
10700: for @code{object}) is shown in the
10701: @uref{objects-implementation.eps,figure}, assuming a cell size of 4.
10702: @comment TODO add this diagram..
10703:
10704: @node Objects Glossary, , Objects Implementation, Objects
10705: @subsubsection @file{objects.fs} Glossary
10706: @cindex @file{objects.fs} Glossary
10707:
10708:
10709: doc---objects-bind
10710: doc---objects-<bind>
10711: doc---objects-bind'
10712: doc---objects-[bind]
10713: doc---objects-class
10714: doc---objects-class->map
10715: doc---objects-class-inst-size
10716: doc---objects-class-override!
10717: doc---objects-class-previous
10718: doc---objects-class>order
10719: doc---objects-construct
10720: doc---objects-current'
10721: doc---objects-[current]
10722: doc---objects-current-interface
10723: doc---objects-dict-new
10724: doc---objects-end-class
10725: doc---objects-end-class-noname
10726: doc---objects-end-interface
10727: doc---objects-end-interface-noname
10728: doc---objects-end-methods
10729: doc---objects-exitm
10730: doc---objects-heap-new
10731: doc---objects-implementation
10732: doc---objects-init-object
10733: doc---objects-inst-value
10734: doc---objects-inst-var
10735: doc---objects-interface
10736: doc---objects-m:
10737: doc---objects-:m
10738: doc---objects-;m
10739: doc---objects-method
10740: doc---objects-methods
10741: doc---objects-object
10742: doc---objects-overrides
10743: doc---objects-[parent]
10744: doc---objects-print
10745: doc---objects-protected
10746: doc---objects-public
10747: doc---objects-selector
10748: doc---objects-this
10749: doc---objects-<to-inst>
10750: doc---objects-[to-inst]
10751: doc---objects-to-this
10752: doc---objects-xt-new
10753:
10754:
10755: @c -------------------------------------------------------------
10756: @node OOF, Mini-OOF, Objects, Object-oriented Forth
10757: @subsection The @file{oof.fs} model
10758: @cindex oof
10759: @cindex object-oriented programming
10760:
10761: @cindex @file{objects.fs}
10762: @cindex @file{oof.fs}
10763:
10764: This section describes the @file{oof.fs} package.
10765:
10766: The package described in this section has been used in bigFORTH since 1991, and
10767: used for two large applications: a chromatographic system used to
10768: create new medicaments, and a graphic user interface library (MINOS).
10769:
10770: You can find a description (in German) of @file{oof.fs} in @cite{Object
10771: oriented bigFORTH} by Bernd Paysan, published in @cite{Vierte Dimension}
10772: 10(2), 1994.
10773:
10774: @menu
10775: * Properties of the OOF model::
10776: * Basic OOF Usage::
10777: * The OOF base class::
10778: * Class Declaration::
10779: * Class Implementation::
10780: @end menu
10781:
10782: @node Properties of the OOF model, Basic OOF Usage, OOF, OOF
10783: @subsubsection Properties of the @file{oof.fs} model
10784: @cindex @file{oof.fs} properties
10785:
10786: @itemize @bullet
10787: @item
10788: This model combines object oriented programming with information
10789: hiding. It helps you writing large application, where scoping is
10790: necessary, because it provides class-oriented scoping.
10791:
10792: @item
10793: Named objects, object pointers, and object arrays can be created,
10794: selector invocation uses the ``object selector'' syntax. Selector invocation
10795: to objects and/or selectors on the stack is a bit less convenient, but
10796: possible.
10797:
10798: @item
10799: Selector invocation and instance variable usage of the active object is
10800: straightforward, since both make use of the active object.
10801:
10802: @item
10803: Late binding is efficient and easy to use.
10804:
10805: @item
10806: State-smart objects parse selectors. However, extensibility is provided
10807: using a (parsing) selector @code{postpone} and a selector @code{'}.
10808:
10809: @item
10810: An implementation in ANS Forth is available.
10811:
10812: @end itemize
10813:
10814:
10815: @node Basic OOF Usage, The OOF base class, Properties of the OOF model, OOF
10816: @subsubsection Basic @file{oof.fs} Usage
10817: @cindex @file{oof.fs} usage
10818:
10819: This section uses the same example as for @code{objects} (@pxref{Basic Objects Usage}).
10820:
10821: You can define a class for graphical objects like this:
10822:
10823: @cindex @code{class} usage
10824: @cindex @code{class;} usage
10825: @cindex @code{method} usage
10826: @example
10827: object class graphical \ "object" is the parent class
10828: method draw ( x y graphical -- )
10829: class;
10830: @end example
10831:
10832: This code defines a class @code{graphical} with an
10833: operation @code{draw}. We can perform the operation
10834: @code{draw} on any @code{graphical} object, e.g.:
10835:
10836: @example
10837: 100 100 t-rex draw
10838: @end example
10839:
10840: @noindent
10841: where @code{t-rex} is an object or object pointer, created with e.g.
10842: @code{graphical : t-rex}.
10843:
10844: @cindex abstract class
10845: How do we create a graphical object? With the present definitions,
10846: we cannot create a useful graphical object. The class
10847: @code{graphical} describes graphical objects in general, but not
10848: any concrete graphical object type (C++ users would call it an
10849: @emph{abstract class}); e.g., there is no method for the selector
10850: @code{draw} in the class @code{graphical}.
10851:
10852: For concrete graphical objects, we define child classes of the
10853: class @code{graphical}, e.g.:
10854:
10855: @example
10856: graphical class circle \ "graphical" is the parent class
10857: cell var circle-radius
10858: how:
10859: : draw ( x y -- )
10860: circle-radius @@ draw-circle ;
10861:
10862: : init ( n-radius -- (
10863: circle-radius ! ;
10864: class;
10865: @end example
10866:
10867: Here we define a class @code{circle} as a child of @code{graphical},
10868: with a field @code{circle-radius}; it defines new methods for the
10869: selectors @code{draw} and @code{init} (@code{init} is defined in
10870: @code{object}, the parent class of @code{graphical}).
10871:
10872: Now we can create a circle in the dictionary with:
10873:
10874: @example
10875: 50 circle : my-circle
10876: @end example
10877:
10878: @noindent
10879: @code{:} invokes @code{init}, thus initializing the field
10880: @code{circle-radius} with 50. We can draw this new circle at (100,100)
10881: with:
10882:
10883: @example
10884: 100 100 my-circle draw
10885: @end example
10886:
10887: @cindex selector invocation, restrictions
10888: @cindex class definition, restrictions
10889: Note: You can only invoke a selector if the receiving object belongs to
10890: the class where the selector was defined or one of its descendents;
10891: e.g., you can invoke @code{draw} only for objects belonging to
10892: @code{graphical} or its descendents (e.g., @code{circle}). The scoping
10893: mechanism will check if you try to invoke a selector that is not
10894: defined in this class hierarchy, so you'll get an error at compilation
10895: time.
10896:
10897:
10898: @node The OOF base class, Class Declaration, Basic OOF Usage, OOF
10899: @subsubsection The @file{oof.fs} base class
10900: @cindex @file{oof.fs} base class
10901:
10902: When you define a class, you have to specify a parent class. So how do
10903: you start defining classes? There is one class available from the start:
10904: @code{object}. You have to use it as ancestor for all classes. It is the
10905: only class that has no parent. Classes are also objects, except that
10906: they don't have instance variables; class manipulation such as
10907: inheritance or changing definitions of a class is handled through
10908: selectors of the class @code{object}.
10909:
10910: @code{object} provides a number of selectors:
10911:
10912: @itemize @bullet
10913: @item
10914: @code{class} for subclassing, @code{definitions} to add definitions
10915: later on, and @code{class?} to get type informations (is the class a
10916: subclass of the class passed on the stack?).
10917:
10918: doc---object-class
10919: doc---object-definitions
10920: doc---object-class?
10921:
10922:
10923: @item
10924: @code{init} and @code{dispose} as constructor and destructor of the
10925: object. @code{init} is invocated after the object's memory is allocated,
10926: while @code{dispose} also handles deallocation. Thus if you redefine
10927: @code{dispose}, you have to call the parent's dispose with @code{super
10928: dispose}, too.
10929:
10930: doc---object-init
10931: doc---object-dispose
10932:
10933:
10934: @item
10935: @code{new}, @code{new[]}, @code{:}, @code{ptr}, @code{asptr}, and
10936: @code{[]} to create named and unnamed objects and object arrays or
10937: object pointers.
10938:
10939: doc---object-new
10940: doc---object-new[]
10941: doc---object-:
10942: doc---object-ptr
10943: doc---object-asptr
10944: doc---object-[]
10945:
10946:
10947: @item
10948: @code{::} and @code{super} for explicit scoping. You should use explicit
10949: scoping only for super classes or classes with the same set of instance
10950: variables. Explicitly-scoped selectors use early binding.
10951:
10952: doc---object-::
10953: doc---object-super
10954:
10955:
10956: @item
10957: @code{self} to get the address of the object
10958:
10959: doc---object-self
10960:
10961:
10962: @item
10963: @code{bind}, @code{bound}, @code{link}, and @code{is} to assign object
10964: pointers and instance defers.
10965:
10966: doc---object-bind
10967: doc---object-bound
10968: doc---object-link
10969: doc---object-is
10970:
10971:
10972: @item
10973: @code{'} to obtain selector tokens, @code{send} to invocate selectors
10974: form the stack, and @code{postpone} to generate selector invocation code.
10975:
10976: doc---object-'
10977: doc---object-postpone
10978:
10979:
10980: @item
10981: @code{with} and @code{endwith} to select the active object from the
10982: stack, and enable its scope. Using @code{with} and @code{endwith}
10983: also allows you to create code using selector @code{postpone} without being
10984: trapped by the state-smart objects.
10985:
10986: doc---object-with
10987: doc---object-endwith
10988:
10989:
10990: @end itemize
10991:
10992: @node Class Declaration, Class Implementation, The OOF base class, OOF
10993: @subsubsection Class Declaration
10994: @cindex class declaration
10995:
10996: @itemize @bullet
10997: @item
10998: Instance variables
10999:
11000: doc---oof-var
11001:
11002:
11003: @item
11004: Object pointers
11005:
11006: doc---oof-ptr
11007: doc---oof-asptr
11008:
11009:
11010: @item
11011: Instance defers
11012:
11013: doc---oof-defer
11014:
11015:
11016: @item
11017: Method selectors
11018:
11019: doc---oof-early
11020: doc---oof-method
11021:
11022:
11023: @item
11024: Class-wide variables
11025:
11026: doc---oof-static
11027:
11028:
11029: @item
11030: End declaration
11031:
11032: doc---oof-how:
11033: doc---oof-class;
11034:
11035:
11036: @end itemize
11037:
11038: @c -------------------------------------------------------------
11039: @node Class Implementation, , Class Declaration, OOF
11040: @subsubsection Class Implementation
11041: @cindex class implementation
11042:
11043: @c -------------------------------------------------------------
11044: @node Mini-OOF, Comparison with other object models, OOF, Object-oriented Forth
11045: @subsection The @file{mini-oof.fs} model
11046: @cindex mini-oof
11047:
11048: Gforth's third object oriented Forth package is a 12-liner. It uses a
11049: mixture of the @file{objects.fs} and the @file{oof.fs} syntax,
11050: and reduces to the bare minimum of features. This is based on a posting
11051: of Bernd Paysan in comp.lang.forth.
11052:
11053: @menu
11054: * Basic Mini-OOF Usage::
11055: * Mini-OOF Example::
11056: * Mini-OOF Implementation::
11057: @end menu
11058:
11059: @c -------------------------------------------------------------
11060: @node Basic Mini-OOF Usage, Mini-OOF Example, Mini-OOF, Mini-OOF
11061: @subsubsection Basic @file{mini-oof.fs} Usage
11062: @cindex mini-oof usage
11063:
11064: There is a base class (@code{class}, which allocates one cell for the
11065: object pointer) plus seven other words: to define a method, a variable,
11066: a class; to end a class, to resolve binding, to allocate an object and
11067: to compile a class method.
11068: @comment TODO better description of the last one
11069:
11070:
11071: doc-object
11072: doc-method
11073: doc-var
11074: doc-class
11075: doc-end-class
11076: doc-defines
11077: doc-new
11078: doc-::
11079:
11080:
11081:
11082: @c -------------------------------------------------------------
11083: @node Mini-OOF Example, Mini-OOF Implementation, Basic Mini-OOF Usage, Mini-OOF
11084: @subsubsection Mini-OOF Example
11085: @cindex mini-oof example
11086:
11087: A short example shows how to use this package. This example, in slightly
11088: extended form, is supplied as @file{moof-exm.fs}
11089: @comment TODO could flesh this out with some comments from the Forthwrite article
11090:
11091: @example
11092: object class
11093: method init
11094: method draw
11095: end-class graphical
11096: @end example
11097:
11098: This code defines a class @code{graphical} with an
11099: operation @code{draw}. We can perform the operation
11100: @code{draw} on any @code{graphical} object, e.g.:
11101:
11102: @example
11103: 100 100 t-rex draw
11104: @end example
11105:
11106: where @code{t-rex} is an object or object pointer, created with e.g.
11107: @code{graphical new Constant t-rex}.
11108:
11109: For concrete graphical objects, we define child classes of the
11110: class @code{graphical}, e.g.:
11111:
11112: @example
11113: graphical class
11114: cell var circle-radius
11115: end-class circle \ "graphical" is the parent class
11116:
11117: :noname ( x y -- )
11118: circle-radius @@ draw-circle ; circle defines draw
11119: :noname ( r -- )
11120: circle-radius ! ; circle defines init
11121: @end example
11122:
11123: There is no implicit init method, so we have to define one. The creation
11124: code of the object now has to call init explicitely.
11125:
11126: @example
11127: circle new Constant my-circle
11128: 50 my-circle init
11129: @end example
11130:
11131: It is also possible to add a function to create named objects with
11132: automatic call of @code{init}, given that all objects have @code{init}
11133: on the same place:
11134:
11135: @example
11136: : new: ( .. o "name" -- )
11137: new dup Constant init ;
11138: 80 circle new: large-circle
11139: @end example
11140:
11141: We can draw this new circle at (100,100) with:
11142:
11143: @example
11144: 100 100 my-circle draw
11145: @end example
11146:
11147: @node Mini-OOF Implementation, , Mini-OOF Example, Mini-OOF
11148: @subsubsection @file{mini-oof.fs} Implementation
11149:
11150: Object-oriented systems with late binding typically use a
11151: ``vtable''-approach: the first variable in each object is a pointer to a
11152: table, which contains the methods as function pointers. The vtable
11153: may also contain other information.
11154:
11155: So first, let's declare selectors:
11156:
11157: @example
11158: : method ( m v "name" -- m' v ) Create over , swap cell+ swap
11159: DOES> ( ... o -- ... ) @@ over @@ + @@ execute ;
11160: @end example
11161:
11162: During selector declaration, the number of selectors and instance
11163: variables is on the stack (in address units). @code{method} creates one
11164: selector and increments the selector number. To execute a selector, it
11165: takes the object, fetches the vtable pointer, adds the offset, and
11166: executes the method @i{xt} stored there. Each selector takes the object
11167: it is invoked with as top of stack parameter; it passes the parameters
11168: (including the object) unchanged to the appropriate method which should
11169: consume that object.
11170:
11171: Now, we also have to declare instance variables
11172:
11173: @example
11174: : var ( m v size "name" -- m v' ) Create over , +
11175: DOES> ( o -- addr ) @@ + ;
11176: @end example
11177:
11178: As before, a word is created with the current offset. Instance
11179: variables can have different sizes (cells, floats, doubles, chars), so
11180: all we do is take the size and add it to the offset. If your machine
11181: has alignment restrictions, put the proper @code{aligned} or
11182: @code{faligned} before the variable, to adjust the variable
11183: offset. That's why it is on the top of stack.
11184:
11185: We need a starting point (the base object) and some syntactic sugar:
11186:
11187: @example
11188: Create object 1 cells , 2 cells ,
11189: : class ( class -- class selectors vars ) dup 2@@ ;
11190: @end example
11191:
11192: For inheritance, the vtable of the parent object has to be
11193: copied when a new, derived class is declared. This gives all the
11194: methods of the parent class, which can be overridden, though.
11195:
11196: @example
11197: : end-class ( class selectors vars "name" -- )
11198: Create here >r , dup , 2 cells ?DO ['] noop , 1 cells +LOOP
11199: cell+ dup cell+ r> rot @@ 2 cells /string move ;
11200: @end example
11201:
11202: The first line creates the vtable, initialized with
11203: @code{noop}s. The second line is the inheritance mechanism, it
11204: copies the xts from the parent vtable.
11205:
11206: We still have no way to define new methods, let's do that now:
11207:
11208: @example
11209: : defines ( xt class "name" -- ) ' >body @@ + ! ;
11210: @end example
11211:
11212: To allocate a new object, we need a word, too:
11213:
11214: @example
11215: : new ( class -- o ) here over @@ allot swap over ! ;
11216: @end example
11217:
11218: Sometimes derived classes want to access the method of the
11219: parent object. There are two ways to achieve this with Mini-OOF:
11220: first, you could use named words, and second, you could look up the
11221: vtable of the parent object.
11222:
11223: @example
11224: : :: ( class "name" -- ) ' >body @@ + @@ compile, ;
11225: @end example
11226:
11227:
11228: Nothing can be more confusing than a good example, so here is
11229: one. First let's declare a text object (called
11230: @code{button}), that stores text and position:
11231:
11232: @example
11233: object class
11234: cell var text
11235: cell var len
11236: cell var x
11237: cell var y
11238: method init
11239: method draw
11240: end-class button
11241: @end example
11242:
11243: @noindent
11244: Now, implement the two methods, @code{draw} and @code{init}:
11245:
11246: @example
11247: :noname ( o -- )
11248: >r r@@ x @@ r@@ y @@ at-xy r@@ text @@ r> len @@ type ;
11249: button defines draw
11250: :noname ( addr u o -- )
11251: >r 0 r@@ x ! 0 r@@ y ! r@@ len ! r> text ! ;
11252: button defines init
11253: @end example
11254:
11255: @noindent
11256: To demonstrate inheritance, we define a class @code{bold-button}, with no
11257: new data and no new selectors:
11258:
11259: @example
11260: button class
11261: end-class bold-button
11262:
11263: : bold 27 emit ." [1m" ;
11264: : normal 27 emit ." [0m" ;
11265: @end example
11266:
11267: @noindent
11268: The class @code{bold-button} has a different draw method to
11269: @code{button}, but the new method is defined in terms of the draw method
11270: for @code{button}:
11271:
11272: @example
11273: :noname bold [ button :: draw ] normal ; bold-button defines draw
11274: @end example
11275:
11276: @noindent
11277: Finally, create two objects and apply selectors:
11278:
11279: @example
11280: button new Constant foo
11281: s" thin foo" foo init
11282: page
11283: foo draw
11284: bold-button new Constant bar
11285: s" fat bar" bar init
11286: 1 bar y !
11287: bar draw
11288: @end example
11289:
11290:
11291: @node Comparison with other object models, , Mini-OOF, Object-oriented Forth
11292: @subsection Comparison with other object models
11293: @cindex comparison of object models
11294: @cindex object models, comparison
11295:
11296: Many object-oriented Forth extensions have been proposed (@cite{A survey
11297: of object-oriented Forths} (SIGPLAN Notices, April 1996) by Bradford
11298: J. Rodriguez and W. F. S. Poehlman lists 17). This section discusses the
11299: relation of the object models described here to two well-known and two
11300: closely-related (by the use of method maps) models. Andras Zsoter
11301: helped us with this section.
11302:
11303: @cindex Neon model
11304: The most popular model currently seems to be the Neon model (see
11305: @cite{Object-oriented programming in ANS Forth} (Forth Dimensions, March
11306: 1997) by Andrew McKewan) but this model has a number of limitations
11307: @footnote{A longer version of this critique can be
11308: found in @cite{On Standardizing Object-Oriented Forth Extensions} (Forth
11309: Dimensions, May 1997) by Anton Ertl.}:
11310:
11311: @itemize @bullet
11312: @item
11313: It uses a @code{@emph{selector object}} syntax, which makes it unnatural
11314: to pass objects on the stack.
11315:
11316: @item
11317: It requires that the selector parses the input stream (at
11318: compile time); this leads to reduced extensibility and to bugs that are
11319: hard to find.
11320:
11321: @item
11322: It allows using every selector on every object; this eliminates the
11323: need for interfaces, but makes it harder to create efficient
11324: implementations.
11325: @end itemize
11326:
11327: @cindex Pountain's object-oriented model
11328: Another well-known publication is @cite{Object-Oriented Forth} (Academic
11329: Press, London, 1987) by Dick Pountain. However, it is not really about
11330: object-oriented programming, because it hardly deals with late
11331: binding. Instead, it focuses on features like information hiding and
11332: overloading that are characteristic of modular languages like Ada (83).
11333:
11334: @cindex Zsoter's object-oriented model
11335: In @uref{http://www.forth.org/oopf.html, Does late binding have to be
11336: slow?} (Forth Dimensions 18(1) 1996, pages 31-35) Andras Zsoter
11337: describes a model that makes heavy use of an active object (like
11338: @code{this} in @file{objects.fs}): The active object is not only used
11339: for accessing all fields, but also specifies the receiving object of
11340: every selector invocation; you have to change the active object
11341: explicitly with @code{@{ ... @}}, whereas in @file{objects.fs} it
11342: changes more or less implicitly at @code{m: ... ;m}. Such a change at
11343: the method entry point is unnecessary with Zsoter's model, because the
11344: receiving object is the active object already. On the other hand, the
11345: explicit change is absolutely necessary in that model, because otherwise
11346: no one could ever change the active object. An ANS Forth implementation
11347: of this model is available through
11348: @uref{http://www.forth.org/oopf.html}.
11349:
11350: @cindex @file{oof.fs}, differences to other models
11351: The @file{oof.fs} model combines information hiding and overloading
11352: resolution (by keeping names in various word lists) with object-oriented
11353: programming. It sets the active object implicitly on method entry, but
11354: also allows explicit changing (with @code{>o...o>} or with
11355: @code{with...endwith}). It uses parsing and state-smart objects and
11356: classes for resolving overloading and for early binding: the object or
11357: class parses the selector and determines the method from this. If the
11358: selector is not parsed by an object or class, it performs a call to the
11359: selector for the active object (late binding), like Zsoter's model.
11360: Fields are always accessed through the active object. The big
11361: disadvantage of this model is the parsing and the state-smartness, which
11362: reduces extensibility and increases the opportunities for subtle bugs;
11363: essentially, you are only safe if you never tick or @code{postpone} an
11364: object or class (Bernd disagrees, but I (Anton) am not convinced).
11365:
11366: @cindex @file{mini-oof.fs}, differences to other models
11367: The @file{mini-oof.fs} model is quite similar to a very stripped-down
11368: version of the @file{objects.fs} model, but syntactically it is a
11369: mixture of the @file{objects.fs} and @file{oof.fs} models.
11370:
11371:
11372: @c -------------------------------------------------------------
11373: @node Programming Tools, Assembler and Code Words, Object-oriented Forth, Words
11374: @section Programming Tools
11375: @cindex programming tools
11376:
11377: @c !! move this and assembler down below OO stuff.
11378:
11379: @menu
11380: * Examining::
11381: * Forgetting words::
11382: * Debugging:: Simple and quick.
11383: * Assertions:: Making your programs self-checking.
11384: * Singlestep Debugger:: Executing your program word by word.
11385: @end menu
11386:
11387: @node Examining, Forgetting words, Programming Tools, Programming Tools
11388: @subsection Examining data and code
11389: @cindex examining data and code
11390: @cindex data examination
11391: @cindex code examination
11392:
11393: The following words inspect the stack non-destructively:
11394:
11395: doc-.s
11396: doc-f.s
11397:
11398: There is a word @code{.r} but it does @i{not} display the return stack!
11399: It is used for formatted numeric output (@pxref{Simple numeric output}).
11400:
11401: doc-depth
11402: doc-fdepth
11403: doc-clearstack
11404:
11405: The following words inspect memory.
11406:
11407: doc-?
11408: doc-dump
11409:
11410: And finally, @code{see} allows to inspect code:
11411:
11412: doc-see
11413: doc-xt-see
11414:
11415: @node Forgetting words, Debugging, Examining, Programming Tools
11416: @subsection Forgetting words
11417: @cindex words, forgetting
11418: @cindex forgeting words
11419:
11420: @c anton: other, maybe better places for this subsection: Defining Words;
11421: @c Dictionary allocation. At least a reference should be there.
11422:
11423: Forth allows you to forget words (and everything that was alloted in the
11424: dictonary after them) in a LIFO manner.
11425:
11426: doc-marker
11427:
11428: The most common use of this feature is during progam development: when
11429: you change a source file, forget all the words it defined and load it
11430: again (since you also forget everything defined after the source file
11431: was loaded, you have to reload that, too). Note that effects like
11432: storing to variables and destroyed system words are not undone when you
11433: forget words. With a system like Gforth, that is fast enough at
11434: starting up and compiling, I find it more convenient to exit and restart
11435: Gforth, as this gives me a clean slate.
11436:
11437: Here's an example of using @code{marker} at the start of a source file
11438: that you are debugging; it ensures that you only ever have one copy of
11439: the file's definitions compiled at any time:
11440:
11441: @example
11442: [IFDEF] my-code
11443: my-code
11444: [ENDIF]
11445:
11446: marker my-code
11447: init-included-files
11448:
11449: \ .. definitions start here
11450: \ .
11451: \ .
11452: \ end
11453: @end example
11454:
11455:
11456: @node Debugging, Assertions, Forgetting words, Programming Tools
11457: @subsection Debugging
11458: @cindex debugging
11459:
11460: Languages with a slow edit/compile/link/test development loop tend to
11461: require sophisticated tracing/stepping debuggers to facilate debugging.
11462:
11463: A much better (faster) way in fast-compiling languages is to add
11464: printing code at well-selected places, let the program run, look at
11465: the output, see where things went wrong, add more printing code, etc.,
11466: until the bug is found.
11467:
11468: The simple debugging aids provided in @file{debugs.fs}
11469: are meant to support this style of debugging.
11470:
11471: The word @code{~~} prints debugging information (by default the source
11472: location and the stack contents). It is easy to insert. If you use Emacs
11473: it is also easy to remove (@kbd{C-x ~} in the Emacs Forth mode to
11474: query-replace them with nothing). The deferred words
11475: @code{printdebugdata} and @code{printdebugline} control the output of
11476: @code{~~}. The default source location output format works well with
11477: Emacs' compilation mode, so you can step through the program at the
11478: source level using @kbd{C-x `} (the advantage over a stepping debugger
11479: is that you can step in any direction and you know where the crash has
11480: happened or where the strange data has occurred).
11481:
11482: doc-~~
11483: doc-printdebugdata
11484: doc-printdebugline
11485:
11486: @node Assertions, Singlestep Debugger, Debugging, Programming Tools
11487: @subsection Assertions
11488: @cindex assertions
11489:
11490: It is a good idea to make your programs self-checking, especially if you
11491: make an assumption that may become invalid during maintenance (for
11492: example, that a certain field of a data structure is never zero). Gforth
11493: supports @dfn{assertions} for this purpose. They are used like this:
11494:
11495: @example
11496: assert( @i{flag} )
11497: @end example
11498:
11499: The code between @code{assert(} and @code{)} should compute a flag, that
11500: should be true if everything is alright and false otherwise. It should
11501: not change anything else on the stack. The overall stack effect of the
11502: assertion is @code{( -- )}. E.g.
11503:
11504: @example
11505: assert( 1 1 + 2 = ) \ what we learn in school
11506: assert( dup 0<> ) \ assert that the top of stack is not zero
11507: assert( false ) \ this code should not be reached
11508: @end example
11509:
11510: The need for assertions is different at different times. During
11511: debugging, we want more checking, in production we sometimes care more
11512: for speed. Therefore, assertions can be turned off, i.e., the assertion
11513: becomes a comment. Depending on the importance of an assertion and the
11514: time it takes to check it, you may want to turn off some assertions and
11515: keep others turned on. Gforth provides several levels of assertions for
11516: this purpose:
11517:
11518:
11519: doc-assert0(
11520: doc-assert1(
11521: doc-assert2(
11522: doc-assert3(
11523: doc-assert(
11524: doc-)
11525:
11526:
11527: The variable @code{assert-level} specifies the highest assertions that
11528: are turned on. I.e., at the default @code{assert-level} of one,
11529: @code{assert0(} and @code{assert1(} assertions perform checking, while
11530: @code{assert2(} and @code{assert3(} assertions are treated as comments.
11531:
11532: The value of @code{assert-level} is evaluated at compile-time, not at
11533: run-time. Therefore you cannot turn assertions on or off at run-time;
11534: you have to set the @code{assert-level} appropriately before compiling a
11535: piece of code. You can compile different pieces of code at different
11536: @code{assert-level}s (e.g., a trusted library at level 1 and
11537: newly-written code at level 3).
11538:
11539:
11540: doc-assert-level
11541:
11542:
11543: If an assertion fails, a message compatible with Emacs' compilation mode
11544: is produced and the execution is aborted (currently with @code{ABORT"}.
11545: If there is interest, we will introduce a special throw code. But if you
11546: intend to @code{catch} a specific condition, using @code{throw} is
11547: probably more appropriate than an assertion).
11548:
11549: Definitions in ANS Forth for these assertion words are provided
11550: in @file{compat/assert.fs}.
11551:
11552:
11553: @node Singlestep Debugger, , Assertions, Programming Tools
11554: @subsection Singlestep Debugger
11555: @cindex singlestep Debugger
11556: @cindex debugging Singlestep
11557:
11558: When you create a new word there's often the need to check whether it
11559: behaves correctly or not. You can do this by typing @code{dbg
11560: badword}. A debug session might look like this:
11561:
11562: @example
11563: : badword 0 DO i . LOOP ; ok
11564: 2 dbg badword
11565: : badword
11566: Scanning code...
11567:
11568: Nesting debugger ready!
11569:
11570: 400D4738 8049BC4 0 -> [ 2 ] 00002 00000
11571: 400D4740 8049F68 DO -> [ 0 ]
11572: 400D4744 804A0C8 i -> [ 1 ] 00000
11573: 400D4748 400C5E60 . -> 0 [ 0 ]
11574: 400D474C 8049D0C LOOP -> [ 0 ]
11575: 400D4744 804A0C8 i -> [ 1 ] 00001
11576: 400D4748 400C5E60 . -> 1 [ 0 ]
11577: 400D474C 8049D0C LOOP -> [ 0 ]
11578: 400D4758 804B384 ; -> ok
11579: @end example
11580:
11581: Each line displayed is one step. You always have to hit return to
11582: execute the next word that is displayed. If you don't want to execute
11583: the next word in a whole, you have to type @kbd{n} for @code{nest}. Here is
11584: an overview what keys are available:
11585:
11586: @table @i
11587:
11588: @item @key{RET}
11589: Next; Execute the next word.
11590:
11591: @item n
11592: Nest; Single step through next word.
11593:
11594: @item u
11595: Unnest; Stop debugging and execute rest of word. If we got to this word
11596: with nest, continue debugging with the calling word.
11597:
11598: @item d
11599: Done; Stop debugging and execute rest.
11600:
11601: @item s
11602: Stop; Abort immediately.
11603:
11604: @end table
11605:
11606: Debugging large application with this mechanism is very difficult, because
11607: you have to nest very deeply into the program before the interesting part
11608: begins. This takes a lot of time.
11609:
11610: To do it more directly put a @code{BREAK:} command into your source code.
11611: When program execution reaches @code{BREAK:} the single step debugger is
11612: invoked and you have all the features described above.
11613:
11614: If you have more than one part to debug it is useful to know where the
11615: program has stopped at the moment. You can do this by the
11616: @code{BREAK" string"} command. This behaves like @code{BREAK:} except that
11617: string is typed out when the ``breakpoint'' is reached.
11618:
11619:
11620: doc-dbg
11621: doc-break:
11622: doc-break"
11623:
11624:
11625:
11626: @c -------------------------------------------------------------
11627: @node Assembler and Code Words, Threading Words, Programming Tools, Words
11628: @section Assembler and Code Words
11629: @cindex assembler
11630: @cindex code words
11631:
11632: @menu
11633: * Code and ;code::
11634: * Common Assembler:: Assembler Syntax
11635: * Common Disassembler::
11636: * 386 Assembler:: Deviations and special cases
11637: * Alpha Assembler:: Deviations and special cases
11638: * MIPS assembler:: Deviations and special cases
11639: * Other assemblers:: How to write them
11640: @end menu
11641:
11642: @node Code and ;code, Common Assembler, Assembler and Code Words, Assembler and Code Words
11643: @subsection @code{Code} and @code{;code}
11644:
11645: Gforth provides some words for defining primitives (words written in
11646: machine code), and for defining the machine-code equivalent of
11647: @code{DOES>}-based defining words. However, the machine-independent
11648: nature of Gforth poses a few problems: First of all, Gforth runs on
11649: several architectures, so it can provide no standard assembler. What's
11650: worse is that the register allocation not only depends on the processor,
11651: but also on the @code{gcc} version and options used.
11652:
11653: The words that Gforth offers encapsulate some system dependences (e.g.,
11654: the header structure), so a system-independent assembler may be used in
11655: Gforth. If you do not have an assembler, you can compile machine code
11656: directly with @code{,} and @code{c,}@footnote{This isn't portable,
11657: because these words emit stuff in @i{data} space; it works because
11658: Gforth has unified code/data spaces. Assembler isn't likely to be
11659: portable anyway.}.
11660:
11661:
11662: doc-assembler
11663: doc-init-asm
11664: doc-code
11665: doc-end-code
11666: doc-;code
11667: doc-flush-icache
11668:
11669:
11670: If @code{flush-icache} does not work correctly, @code{code} words
11671: etc. will not work (reliably), either.
11672:
11673: The typical usage of these @code{code} words can be shown most easily by
11674: analogy to the equivalent high-level defining words:
11675:
11676: @example
11677: : foo code foo
11678: <high-level Forth words> <assembler>
11679: ; end-code
11680:
11681: : bar : bar
11682: <high-level Forth words> <high-level Forth words>
11683: CREATE CREATE
11684: <high-level Forth words> <high-level Forth words>
11685: DOES> ;code
11686: <high-level Forth words> <assembler>
11687: ; end-code
11688: @end example
11689:
11690: @c anton: the following stuff is also in "Common Assembler", in less detail.
11691:
11692: @cindex registers of the inner interpreter
11693: In the assembly code you will want to refer to the inner interpreter's
11694: registers (e.g., the data stack pointer) and you may want to use other
11695: registers for temporary storage. Unfortunately, the register allocation
11696: is installation-dependent.
11697:
11698: In particular, @code{ip} (Forth instruction pointer) and @code{rp}
11699: (return stack pointer) are in different places in @code{gforth} and
11700: @code{gforth-fast}. This means that you cannot write a @code{NEXT}
11701: routine that works on both versions; so for doing @code{NEXT}, I
11702: recomment jumping to @code{' noop >code-address}, which contains nothing
11703: but a @code{NEXT}.
11704:
11705: For general accesses to the inner interpreter's registers, the easiest
11706: solution is to use explicit register declarations (@pxref{Explicit Reg
11707: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) for
11708: all of the inner interpreter's registers: You have to compile Gforth
11709: with @code{-DFORCE_REG} (configure option @code{--enable-force-reg}) and
11710: the appropriate declarations must be present in the @code{machine.h}
11711: file (see @code{mips.h} for an example; you can find a full list of all
11712: declarable register symbols with @code{grep register engine.c}). If you
11713: give explicit registers to all variables that are declared at the
11714: beginning of @code{engine()}, you should be able to use the other
11715: caller-saved registers for temporary storage. Alternatively, you can use
11716: the @code{gcc} option @code{-ffixed-REG} (@pxref{Code Gen Options, ,
11717: Options for Code Generation Conventions, gcc.info, GNU C Manual}) to
11718: reserve a register (however, this restriction on register allocation may
11719: slow Gforth significantly).
11720:
11721: If this solution is not viable (e.g., because @code{gcc} does not allow
11722: you to explicitly declare all the registers you need), you have to find
11723: out by looking at the code where the inner interpreter's registers
11724: reside and which registers can be used for temporary storage. You can
11725: get an assembly listing of the engine's code with @code{make engine.s}.
11726:
11727: In any case, it is good practice to abstract your assembly code from the
11728: actual register allocation. E.g., if the data stack pointer resides in
11729: register @code{$17}, create an alias for this register called @code{sp},
11730: and use that in your assembly code.
11731:
11732: @cindex code words, portable
11733: Another option for implementing normal and defining words efficiently
11734: is to add the desired functionality to the source of Gforth. For normal
11735: words you just have to edit @file{primitives} (@pxref{Automatic
11736: Generation}). Defining words (equivalent to @code{;CODE} words, for fast
11737: defined words) may require changes in @file{engine.c}, @file{kernel.fs},
11738: @file{prims2x.fs}, and possibly @file{cross.fs}.
11739:
11740: @node Common Assembler, Common Disassembler, Code and ;code, Assembler and Code Words
11741: @subsection Common Assembler
11742:
11743: The assemblers in Gforth generally use a postfix syntax, i.e., the
11744: instruction name follows the operands.
11745:
11746: The operands are passed in the usual order (the same that is used in the
11747: manual of the architecture). Since they all are Forth words, they have
11748: to be separated by spaces; you can also use Forth words to compute the
11749: operands.
11750:
11751: The instruction names usually end with a @code{,}. This makes it easier
11752: to visually separate instructions if you put several of them on one
11753: line; it also avoids shadowing other Forth words (e.g., @code{and}).
11754:
11755: Registers are usually specified by number; e.g., (decimal) @code{11}
11756: specifies registers R11 and F11 on the Alpha architecture (which one,
11757: depends on the instruction). The usual names are also available, e.g.,
11758: @code{s2} for R11 on Alpha.
11759:
11760: Control flow is specified similar to normal Forth code (@pxref{Arbitrary
11761: control structures}), with @code{if,}, @code{ahead,}, @code{then,},
11762: @code{begin,}, @code{until,}, @code{again,}, @code{cs-roll},
11763: @code{cs-pick}, @code{else,}, @code{while,}, and @code{repeat,}. The
11764: conditions are specified in a way specific to each assembler.
11765:
11766: Note that the register assignments of the Gforth engine can change
11767: between Gforth versions, or even between different compilations of the
11768: same Gforth version (e.g., if you use a different GCC version). So if
11769: you want to refer to Gforth's registers (e.g., the stack pointer or
11770: TOS), I recommend defining your own words for refering to these
11771: registers, and using them later on; then you can easily adapt to a
11772: changed register assignment. The stability of the register assignment
11773: is usually better if you build Gforth with @code{--enable-force-reg}.
11774:
11775: In particular, the return stack pointer and the instruction pointer are
11776: in memory in @code{gforth}, and usually in registers in
11777: @code{gforth-fast}. The most common use of these registers is to
11778: dispatch to the next word (the @code{next} routine). A portable way to
11779: do this is to jump to @code{' noop >code-address} (of course, this is
11780: less efficient than integrating the @code{next} code and scheduling it
11781: well).
11782:
11783: @node Common Disassembler, 386 Assembler, Common Assembler, Assembler and Code Words
11784: @subsection Common Disassembler
11785:
11786: You can disassemble a @code{code} word with @code{see}
11787: (@pxref{Debugging}). You can disassemble a section of memory with
11788:
11789: doc-disasm
11790:
11791: The disassembler generally produces output that can be fed into the
11792: assembler (i.e., same syntax, etc.). It also includes additional
11793: information in comments. In particular, the address of the instruction
11794: is given in a comment before the instruction.
11795:
11796: @code{See} may display more or less than the actual code of the word,
11797: because the recognition of the end of the code is unreliable. You can
11798: use @code{disasm} if it did not display enough. It may display more, if
11799: the code word is not immediately followed by a named word. If you have
11800: something else there, you can follow the word with @code{align last @ ,}
11801: to ensure that the end is recognized.
11802:
11803: @node 386 Assembler, Alpha Assembler, Common Disassembler, Assembler and Code Words
11804: @subsection 386 Assembler
11805:
11806: The 386 assembler included in Gforth was written by Bernd Paysan, it's
11807: available under GPL, and originally part of bigFORTH.
11808:
11809: The 386 disassembler included in Gforth was written by Andrew McKewan
11810: and is in the public domain.
11811:
11812: The disassembler displays code in prefix Intel syntax.
11813:
11814: The assembler uses a postfix syntax with reversed parameters.
11815:
11816: The assembler includes all instruction of the Athlon, i.e. 486 core
11817: instructions, Pentium and PPro extensions, floating point, MMX, 3Dnow!,
11818: but not ISSE. It's an integrated 16- and 32-bit assembler. Default is 32
11819: bit, you can switch to 16 bit with .86 and back to 32 bit with .386.
11820:
11821: There are several prefixes to switch between different operation sizes,
11822: @code{.b} for byte accesses, @code{.w} for word accesses, @code{.d} for
11823: double-word accesses. Addressing modes can be switched with @code{.wa}
11824: for 16 bit addresses, and @code{.da} for 32 bit addresses. You don't
11825: need a prefix for byte register names (@code{AL} et al).
11826:
11827: For floating point operations, the prefixes are @code{.fs} (IEEE
11828: single), @code{.fl} (IEEE double), @code{.fx} (extended), @code{.fw}
11829: (word), @code{.fd} (double-word), and @code{.fq} (quad-word).
11830:
11831: The MMX opcodes don't have size prefixes, they are spelled out like in
11832: the Intel assembler. Instead of move from and to memory, there are
11833: PLDQ/PLDD and PSTQ/PSTD.
11834:
11835: The registers lack the 'e' prefix; even in 32 bit mode, eax is called
11836: ax. Immediate values are indicated by postfixing them with @code{#},
11837: e.g., @code{3 #}. Here are some examples of addressing modes:
11838:
11839: @example
11840: 3 # \ immediate
11841: 1000 #) \ absolute
11842: ax \ register
11843: 100 di d) \ 100[edi]
11844: 4 bx cx di) \ 4[ebx][ecx]
11845: di ax *4 i) \ [edi][eax*4]
11846: 20 ax *4 i#) \ 20[eax*4]
11847: @end example
11848:
11849: Some example of instructions are:
11850:
11851: @example
11852: ax bx mov \ move ebx,eax
11853: 3 # ax mov \ mov eax,3
11854: 100 di ) ax mov \ mov eax,100[edi]
11855: 4 bx cx di) ax mov \ mov eax,4[ebx][ecx]
11856: .w ax bx mov \ mov bx,ax
11857: @end example
11858:
11859: The following forms are supported for binary instructions:
11860:
11861: @example
11862: <reg> <reg> <inst>
11863: <n> # <reg> <inst>
11864: <mem> <reg> <inst>
11865: <reg> <mem> <inst>
11866: @end example
11867:
11868: Immediate to memory is not supported. The shift/rotate syntax is:
11869:
11870: @example
11871: <reg/mem> 1 # shl \ shortens to shift without immediate
11872: <reg/mem> 4 # shl
11873: <reg/mem> cl shl
11874: @end example
11875:
11876: Precede string instructions (@code{movs} etc.) with @code{.b} to get
11877: the byte version.
11878:
11879: The control structure words @code{IF} @code{UNTIL} etc. must be preceded
11880: by one of these conditions: @code{vs vc u< u>= 0= 0<> u<= u> 0< 0>= ps
11881: pc < >= <= >}. (Note that most of these words shadow some Forth words
11882: when @code{assembler} is in front of @code{forth} in the search path,
11883: e.g., in @code{code} words). Currently the control structure words use
11884: one stack item, so you have to use @code{roll} instead of @code{cs-roll}
11885: to shuffle them (you can also use @code{swap} etc.).
11886:
11887: Here is an example of a @code{code} word (assumes that the stack pointer
11888: is in esi and the TOS is in ebx):
11889:
11890: @example
11891: code my+ ( n1 n2 -- n )
11892: 4 si D) bx add
11893: 4 # si add
11894: Next
11895: end-code
11896: @end example
11897:
11898: @node Alpha Assembler, MIPS assembler, 386 Assembler, Assembler and Code Words
11899: @subsection Alpha Assembler
11900:
11901: The Alpha assembler and disassembler were originally written by Bernd
11902: Thallner.
11903:
11904: The register names @code{a0}--@code{a5} are not available to avoid
11905: shadowing hex numbers.
11906:
11907: Immediate forms of arithmetic instructions are distinguished by a
11908: @code{#} just before the @code{,}, e.g., @code{and#,} (note: @code{lda,}
11909: does not count as arithmetic instruction).
11910:
11911: You have to specify all operands to an instruction, even those that
11912: other assemblers consider optional, e.g., the destination register for
11913: @code{br,}, or the destination register and hint for @code{jmp,}.
11914:
11915: You can specify conditions for @code{if,} by removing the first @code{b}
11916: and the trailing @code{,} from a branch with a corresponding name; e.g.,
11917:
11918: @example
11919: 11 fgt if, \ if F11>0e
11920: ...
11921: endif,
11922: @end example
11923:
11924: @code{fbgt,} gives @code{fgt}.
11925:
11926: @node MIPS assembler, Other assemblers, Alpha Assembler, Assembler and Code Words
11927: @subsection MIPS assembler
11928:
11929: The MIPS assembler was originally written by Christian Pirker.
11930:
11931: Currently the assembler and disassembler only cover the MIPS-I
11932: architecture (R3000), and don't support FP instructions.
11933:
11934: The register names @code{$a0}--@code{$a3} are not available to avoid
11935: shadowing hex numbers.
11936:
11937: Because there is no way to distinguish registers from immediate values,
11938: you have to explicitly use the immediate forms of instructions, i.e.,
11939: @code{addiu,}, not just @code{addu,} (@command{as} does this
11940: implicitly).
11941:
11942: If the architecture manual specifies several formats for the instruction
11943: (e.g., for @code{jalr,}), you usually have to use the one with more
11944: arguments (i.e., two for @code{jalr,}). When in doubt, see
11945: @code{arch/mips/testasm.fs} for an example of correct use.
11946:
11947: Branches and jumps in the MIPS architecture have a delay slot. You have
11948: to fill it yourself (the simplest way is to use @code{nop,}), the
11949: assembler does not do it for you (unlike @command{as}). Even
11950: @code{if,}, @code{ahead,}, @code{until,}, @code{again,}, @code{while,},
11951: @code{else,} and @code{repeat,} need a delay slot. Since @code{begin,}
11952: and @code{then,} just specify branch targets, they are not affected.
11953:
11954: Note that you must not put branches, jumps, or @code{li,} into the delay
11955: slot: @code{li,} may expand to several instructions, and control flow
11956: instructions may not be put into the branch delay slot in any case.
11957:
11958: For branches the argument specifying the target is a relative address;
11959: You have to add the address of the delay slot to get the absolute
11960: address.
11961:
11962: The MIPS architecture also has load delay slots and restrictions on
11963: using @code{mfhi,} and @code{mflo,}; you have to order the instructions
11964: yourself to satisfy these restrictions, the assembler does not do it for
11965: you.
11966:
11967: You can specify the conditions for @code{if,} etc. by taking a
11968: conditional branch and leaving away the @code{b} at the start and the
11969: @code{,} at the end. E.g.,
11970:
11971: @example
11972: 4 5 eq if,
11973: ... \ do something if $4 equals $5
11974: then,
11975: @end example
11976:
11977: @node Other assemblers, , MIPS assembler, Assembler and Code Words
11978: @subsection Other assemblers
11979:
11980: If you want to contribute another assembler/disassembler, please contact
11981: us (@email{bug-gforth@@gnu.org}) to check if we have such an assembler
11982: already. If you are writing them from scratch, please use a similar
11983: syntax style as the one we use (i.e., postfix, commas at the end of the
11984: instruction names, @pxref{Common Assembler}); make the output of the
11985: disassembler be valid input for the assembler, and keep the style
11986: similar to the style we used.
11987:
11988: Hints on implementation: The most important part is to have a good test
11989: suite that contains all instructions. Once you have that, the rest is
11990: easy. For actual coding you can take a look at
11991: @file{arch/mips/disasm.fs} to get some ideas on how to use data for both
11992: the assembler and disassembler, avoiding redundancy and some potential
11993: bugs. You can also look at that file (and @pxref{Advanced does> usage
11994: example}) to get ideas how to factor a disassembler.
11995:
11996: Start with the disassembler, because it's easier to reuse data from the
11997: disassembler for the assembler than the other way round.
11998:
11999: For the assembler, take a look at @file{arch/alpha/asm.fs}, which shows
12000: how simple it can be.
12001:
12002: @c -------------------------------------------------------------
12003: @node Threading Words, Passing Commands to the OS, Assembler and Code Words, Words
12004: @section Threading Words
12005: @cindex threading words
12006:
12007: @cindex code address
12008: These words provide access to code addresses and other threading stuff
12009: in Gforth (and, possibly, other interpretive Forths). It more or less
12010: abstracts away the differences between direct and indirect threading
12011: (and, for direct threading, the machine dependences). However, at
12012: present this wordset is still incomplete. It is also pretty low-level;
12013: some day it will hopefully be made unnecessary by an internals wordset
12014: that abstracts implementation details away completely.
12015:
12016: The terminology used here stems from indirect threaded Forth systems; in
12017: such a system, the XT of a word is represented by the CFA (code field
12018: address) of a word; the CFA points to a cell that contains the code
12019: address. The code address is the address of some machine code that
12020: performs the run-time action of invoking the word (e.g., the
12021: @code{dovar:} routine pushes the address of the body of the word (a
12022: variable) on the stack
12023: ).
12024:
12025: @cindex code address
12026: @cindex code field address
12027: In an indirect threaded Forth, you can get the code address of @i{name}
12028: with @code{' @i{name} @@}; in Gforth you can get it with @code{' @i{name}
12029: >code-address}, independent of the threading method.
12030:
12031: doc-threading-method
12032: doc->code-address
12033: doc-code-address!
12034:
12035: @cindex @code{does>}-handler
12036: @cindex @code{does>}-code
12037: For a word defined with @code{DOES>}, the code address usually points to
12038: a jump instruction (the @dfn{does-handler}) that jumps to the dodoes
12039: routine (in Gforth on some platforms, it can also point to the dodoes
12040: routine itself). What you are typically interested in, though, is
12041: whether a word is a @code{DOES>}-defined word, and what Forth code it
12042: executes; @code{>does-code} tells you that.
12043:
12044: doc->does-code
12045:
12046: To create a @code{DOES>}-defined word with the following basic words,
12047: you have to set up a @code{DOES>}-handler with @code{does-handler!};
12048: @code{/does-handler} aus behind you have to place your executable Forth
12049: code. Finally you have to create a word and modify its behaviour with
12050: @code{does-handler!}.
12051:
12052: doc-does-code!
12053: doc-does-handler!
12054: doc-/does-handler
12055:
12056: The code addresses produced by various defining words are produced by
12057: the following words:
12058:
12059: doc-docol:
12060: doc-docon:
12061: doc-dovar:
12062: doc-douser:
12063: doc-dodefer:
12064: doc-dofield:
12065:
12066: @c -------------------------------------------------------------
12067: @node Passing Commands to the OS, Keeping track of Time, Threading Words, Words
12068: @section Passing Commands to the Operating System
12069: @cindex operating system - passing commands
12070: @cindex shell commands
12071:
12072: Gforth allows you to pass an arbitrary string to the host operating
12073: system shell (if such a thing exists) for execution.
12074:
12075:
12076: doc-sh
12077: doc-system
12078: doc-$?
12079: doc-getenv
12080:
12081:
12082: @c -------------------------------------------------------------
12083: @node Keeping track of Time, Miscellaneous Words, Passing Commands to the OS, Words
12084: @section Keeping track of Time
12085: @cindex time-related words
12086:
12087: doc-ms
12088: doc-time&date
12089: doc-utime
12090: doc-cputime
12091:
12092:
12093: @c -------------------------------------------------------------
12094: @node Miscellaneous Words, , Keeping track of Time, Words
12095: @section Miscellaneous Words
12096: @cindex miscellaneous words
12097:
12098: @comment TODO find homes for these
12099:
12100: These section lists the ANS Forth words that are not documented
12101: elsewhere in this manual. Ultimately, they all need proper homes.
12102:
12103: doc-quit
12104:
12105: The following ANS Forth words are not currently supported by Gforth
12106: (@pxref{ANS conformance}):
12107:
12108: @code{EDITOR}
12109: @code{EMIT?}
12110: @code{FORGET}
12111:
12112: @c ******************************************************************
12113: @node Error messages, Tools, Words, Top
12114: @chapter Error messages
12115: @cindex error messages
12116: @cindex backtrace
12117:
12118: A typical Gforth error message looks like this:
12119:
12120: @example
12121: in file included from \evaluated string/:-1
12122: in file included from ./yyy.fs:1
12123: ./xxx.fs:4: Invalid memory address
12124: bar
12125: ^^^
12126: Backtrace:
12127: $400E664C @@
12128: $400E6664 foo
12129: @end example
12130:
12131: The message identifying the error is @code{Invalid memory address}. The
12132: error happened when text-interpreting line 4 of the file
12133: @file{./xxx.fs}. This line is given (it contains @code{bar}), and the
12134: word on the line where the error happened, is pointed out (with
12135: @code{^^^}).
12136:
12137: The file containing the error was included in line 1 of @file{./yyy.fs},
12138: and @file{yyy.fs} was included from a non-file (in this case, by giving
12139: @file{yyy.fs} as command-line parameter to Gforth).
12140:
12141: At the end of the error message you find a return stack dump that can be
12142: interpreted as a backtrace (possibly empty). On top you find the top of
12143: the return stack when the @code{throw} happened, and at the bottom you
12144: find the return stack entry just above the return stack of the topmost
12145: text interpreter.
12146:
12147: To the right of most return stack entries you see a guess for the word
12148: that pushed that return stack entry as its return address. This gives a
12149: backtrace. In our case we see that @code{bar} called @code{foo}, and
12150: @code{foo} called @code{@@} (and @code{@@} had an @emph{Invalid memory
12151: address} exception).
12152:
12153: Note that the backtrace is not perfect: We don't know which return stack
12154: entries are return addresses (so we may get false positives); and in
12155: some cases (e.g., for @code{abort"}) we cannot determine from the return
12156: address the word that pushed the return address, so for some return
12157: addresses you see no names in the return stack dump.
12158:
12159: @cindex @code{catch} and backtraces
12160: The return stack dump represents the return stack at the time when a
12161: specific @code{throw} was executed. In programs that make use of
12162: @code{catch}, it is not necessarily clear which @code{throw} should be
12163: used for the return stack dump (e.g., consider one @code{throw} that
12164: indicates an error, which is caught, and during recovery another error
12165: happens; which @code{throw} should be used for the stack dump?). Gforth
12166: presents the return stack dump for the first @code{throw} after the last
12167: executed (not returned-to) @code{catch}; this works well in the usual
12168: case.
12169:
12170: @cindex @code{gforth-fast} and backtraces
12171: @cindex @code{gforth-fast}, difference from @code{gforth}
12172: @cindex backtraces with @code{gforth-fast}
12173: @cindex return stack dump with @code{gforth-fast}
12174: @code{Gforth} is able to do a return stack dump for throws generated
12175: from primitives (e.g., invalid memory address, stack empty etc.);
12176: @code{gforth-fast} is only able to do a return stack dump from a
12177: directly called @code{throw} (including @code{abort} etc.). This is the
12178: only difference (apart from a speed factor of between 1.15 (K6-2) and
12179: 2 (21264)) between @code{gforth} and @code{gforth-fast}. Given an
12180: exception caused by a primitive in @code{gforth-fast}, you will
12181: typically see no return stack dump at all; however, if the exception is
12182: caught by @code{catch} (e.g., for restoring some state), and then
12183: @code{throw}n again, the return stack dump will be for the first such
12184: @code{throw}.
12185:
12186: @c ******************************************************************
12187: @node Tools, ANS conformance, Error messages, Top
12188: @chapter Tools
12189:
12190: @menu
12191: * ANS Report:: Report the words used, sorted by wordset.
12192: @end menu
12193:
12194: See also @ref{Emacs and Gforth}.
12195:
12196: @node ANS Report, , Tools, Tools
12197: @section @file{ans-report.fs}: Report the words used, sorted by wordset
12198: @cindex @file{ans-report.fs}
12199: @cindex report the words used in your program
12200: @cindex words used in your program
12201:
12202: If you want to label a Forth program as ANS Forth Program, you must
12203: document which wordsets the program uses; for extension wordsets, it is
12204: helpful to list the words the program requires from these wordsets
12205: (because Forth systems are allowed to provide only some words of them).
12206:
12207: The @file{ans-report.fs} tool makes it easy for you to determine which
12208: words from which wordset and which non-ANS words your application
12209: uses. You simply have to include @file{ans-report.fs} before loading the
12210: program you want to check. After loading your program, you can get the
12211: report with @code{print-ans-report}. A typical use is to run this as
12212: batch job like this:
12213: @example
12214: gforth ans-report.fs myprog.fs -e "print-ans-report bye"
12215: @end example
12216:
12217: The output looks like this (for @file{compat/control.fs}):
12218: @example
12219: The program uses the following words
12220: from CORE :
12221: : POSTPONE THEN ; immediate ?dup IF 0=
12222: from BLOCK-EXT :
12223: \
12224: from FILE :
12225: (
12226: @end example
12227:
12228: @subsection Caveats
12229:
12230: Note that @file{ans-report.fs} just checks which words are used, not whether
12231: they are used in an ANS Forth conforming way!
12232:
12233: Some words are defined in several wordsets in the
12234: standard. @file{ans-report.fs} reports them for only one of the
12235: wordsets, and not necessarily the one you expect. It depends on usage
12236: which wordset is the right one to specify. E.g., if you only use the
12237: compilation semantics of @code{S"}, it is a Core word; if you also use
12238: its interpretation semantics, it is a File word.
12239:
12240: @c ******************************************************************
12241: @node ANS conformance, Standard vs Extensions, Tools, Top
12242: @chapter ANS conformance
12243: @cindex ANS conformance of Gforth
12244:
12245: To the best of our knowledge, Gforth is an
12246:
12247: ANS Forth System
12248: @itemize @bullet
12249: @item providing the Core Extensions word set
12250: @item providing the Block word set
12251: @item providing the Block Extensions word set
12252: @item providing the Double-Number word set
12253: @item providing the Double-Number Extensions word set
12254: @item providing the Exception word set
12255: @item providing the Exception Extensions word set
12256: @item providing the Facility word set
12257: @item providing @code{EKEY}, @code{EKEY>CHAR}, @code{EKEY?}, @code{MS} and @code{TIME&DATE} from the Facility Extensions word set
12258: @item providing the File Access word set
12259: @item providing the File Access Extensions word set
12260: @item providing the Floating-Point word set
12261: @item providing the Floating-Point Extensions word set
12262: @item providing the Locals word set
12263: @item providing the Locals Extensions word set
12264: @item providing the Memory-Allocation word set
12265: @item providing the Memory-Allocation Extensions word set (that one's easy)
12266: @item providing the Programming-Tools word set
12267: @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
12268: @item providing the Search-Order word set
12269: @item providing the Search-Order Extensions word set
12270: @item providing the String word set
12271: @item providing the String Extensions word set (another easy one)
12272: @end itemize
12273:
12274: @cindex system documentation
12275: In addition, ANS Forth systems are required to document certain
12276: implementation choices. This chapter tries to meet these
12277: requirements. In many cases it gives a way to ask the system for the
12278: information instead of providing the information directly, in
12279: particular, if the information depends on the processor, the operating
12280: system or the installation options chosen, or if they are likely to
12281: change during the maintenance of Gforth.
12282:
12283: @comment The framework for the rest has been taken from pfe.
12284:
12285: @menu
12286: * The Core Words::
12287: * The optional Block word set::
12288: * The optional Double Number word set::
12289: * The optional Exception word set::
12290: * The optional Facility word set::
12291: * The optional File-Access word set::
12292: * The optional Floating-Point word set::
12293: * The optional Locals word set::
12294: * The optional Memory-Allocation word set::
12295: * The optional Programming-Tools word set::
12296: * The optional Search-Order word set::
12297: @end menu
12298:
12299:
12300: @c =====================================================================
12301: @node The Core Words, The optional Block word set, ANS conformance, ANS conformance
12302: @comment node-name, next, previous, up
12303: @section The Core Words
12304: @c =====================================================================
12305: @cindex core words, system documentation
12306: @cindex system documentation, core words
12307:
12308: @menu
12309: * core-idef:: Implementation Defined Options
12310: * core-ambcond:: Ambiguous Conditions
12311: * core-other:: Other System Documentation
12312: @end menu
12313:
12314: @c ---------------------------------------------------------------------
12315: @node core-idef, core-ambcond, The Core Words, The Core Words
12316: @subsection Implementation Defined Options
12317: @c ---------------------------------------------------------------------
12318: @cindex core words, implementation-defined options
12319: @cindex implementation-defined options, core words
12320:
12321:
12322: @table @i
12323: @item (Cell) aligned addresses:
12324: @cindex cell-aligned addresses
12325: @cindex aligned addresses
12326: processor-dependent. Gforth's alignment words perform natural alignment
12327: (e.g., an address aligned for a datum of size 8 is divisible by
12328: 8). Unaligned accesses usually result in a @code{-23 THROW}.
12329:
12330: @item @code{EMIT} and non-graphic characters:
12331: @cindex @code{EMIT} and non-graphic characters
12332: @cindex non-graphic characters and @code{EMIT}
12333: The character is output using the C library function (actually, macro)
12334: @code{putc}.
12335:
12336: @item character editing of @code{ACCEPT} and @code{EXPECT}:
12337: @cindex character editing of @code{ACCEPT} and @code{EXPECT}
12338: @cindex editing in @code{ACCEPT} and @code{EXPECT}
12339: @cindex @code{ACCEPT}, editing
12340: @cindex @code{EXPECT}, editing
12341: This is modeled on the GNU readline library (@pxref{Readline
12342: Interaction, , Command Line Editing, readline, The GNU Readline
12343: Library}) with Emacs-like key bindings. @kbd{Tab} deviates a little by
12344: producing a full word completion every time you type it (instead of
12345: producing the common prefix of all completions). @xref{Command-line editing}.
12346:
12347: @item character set:
12348: @cindex character set
12349: The character set of your computer and display device. Gforth is
12350: 8-bit-clean (but some other component in your system may make trouble).
12351:
12352: @item Character-aligned address requirements:
12353: @cindex character-aligned address requirements
12354: installation-dependent. Currently a character is represented by a C
12355: @code{unsigned char}; in the future we might switch to @code{wchar_t}
12356: (Comments on that requested).
12357:
12358: @item character-set extensions and matching of names:
12359: @cindex character-set extensions and matching of names
12360: @cindex case-sensitivity for name lookup
12361: @cindex name lookup, case-sensitivity
12362: @cindex locale and case-sensitivity
12363: Any character except the ASCII NUL character can be used in a
12364: name. Matching is case-insensitive (except in @code{TABLE}s). The
12365: matching is performed using the C library function @code{strncasecmp}, whose
12366: function is probably influenced by the locale. E.g., the @code{C} locale
12367: does not know about accents and umlauts, so they are matched
12368: case-sensitively in that locale. For portability reasons it is best to
12369: write programs such that they work in the @code{C} locale. Then one can
12370: use libraries written by a Polish programmer (who might use words
12371: containing ISO Latin-2 encoded characters) and by a French programmer
12372: (ISO Latin-1) in the same program (of course, @code{WORDS} will produce
12373: funny results for some of the words (which ones, depends on the font you
12374: are using)). Also, the locale you prefer may not be available in other
12375: operating systems. Hopefully, Unicode will solve these problems one day.
12376:
12377: @item conditions under which control characters match a space delimiter:
12378: @cindex space delimiters
12379: @cindex control characters as delimiters
12380: If @code{WORD} is called with the space character as a delimiter, all
12381: white-space characters (as identified by the C macro @code{isspace()})
12382: are delimiters. @code{PARSE}, on the other hand, treats space like other
12383: delimiters. @code{SWORD} treats space like @code{WORD}, but behaves
12384: like @code{PARSE} otherwise. @code{Name}, which is used by the outer
12385: interpreter (aka text interpreter) by default, treats all white-space
12386: characters as delimiters.
12387:
12388: @item format of the control-flow stack:
12389: @cindex control-flow stack, format
12390: The data stack is used as control-flow stack. The size of a control-flow
12391: stack item in cells is given by the constant @code{cs-item-size}. At the
12392: time of this writing, an item consists of a (pointer to a) locals list
12393: (third), an address in the code (second), and a tag for identifying the
12394: item (TOS). The following tags are used: @code{defstart},
12395: @code{live-orig}, @code{dead-orig}, @code{dest}, @code{do-dest},
12396: @code{scopestart}.
12397:
12398: @item conversion of digits > 35
12399: @cindex digits > 35
12400: The characters @code{[\]^_'} are the digits with the decimal value
12401: 36@minus{}41. There is no way to input many of the larger digits.
12402:
12403: @item display after input terminates in @code{ACCEPT} and @code{EXPECT}:
12404: @cindex @code{EXPECT}, display after end of input
12405: @cindex @code{ACCEPT}, display after end of input
12406: The cursor is moved to the end of the entered string. If the input is
12407: terminated using the @kbd{Return} key, a space is typed.
12408:
12409: @item exception abort sequence of @code{ABORT"}:
12410: @cindex exception abort sequence of @code{ABORT"}
12411: @cindex @code{ABORT"}, exception abort sequence
12412: The error string is stored into the variable @code{"error} and a
12413: @code{-2 throw} is performed.
12414:
12415: @item input line terminator:
12416: @cindex input line terminator
12417: @cindex line terminator on input
12418: @cindex newline character on input
12419: For interactive input, @kbd{C-m} (CR) and @kbd{C-j} (LF) terminate
12420: lines. One of these characters is typically produced when you type the
12421: @kbd{Enter} or @kbd{Return} key.
12422:
12423: @item maximum size of a counted string:
12424: @cindex maximum size of a counted string
12425: @cindex counted string, maximum size
12426: @code{s" /counted-string" environment? drop .}. Currently 255 characters
12427: on all platforms, but this may change.
12428:
12429: @item maximum size of a parsed string:
12430: @cindex maximum size of a parsed string
12431: @cindex parsed string, maximum size
12432: Given by the constant @code{/line}. Currently 255 characters.
12433:
12434: @item maximum size of a definition name, in characters:
12435: @cindex maximum size of a definition name, in characters
12436: @cindex name, maximum length
12437: 31
12438:
12439: @item maximum string length for @code{ENVIRONMENT?}, in characters:
12440: @cindex maximum string length for @code{ENVIRONMENT?}, in characters
12441: @cindex @code{ENVIRONMENT?} string length, maximum
12442: 31
12443:
12444: @item method of selecting the user input device:
12445: @cindex user input device, method of selecting
12446: The user input device is the standard input. There is currently no way to
12447: change it from within Gforth. However, the input can typically be
12448: redirected in the command line that starts Gforth.
12449:
12450: @item method of selecting the user output device:
12451: @cindex user output device, method of selecting
12452: @code{EMIT} and @code{TYPE} output to the file-id stored in the value
12453: @code{outfile-id} (@code{stdout} by default). Gforth uses unbuffered
12454: output when the user output device is a terminal, otherwise the output
12455: is buffered.
12456:
12457: @item methods of dictionary compilation:
12458: What are we expected to document here?
12459:
12460: @item number of bits in one address unit:
12461: @cindex number of bits in one address unit
12462: @cindex address unit, size in bits
12463: @code{s" address-units-bits" environment? drop .}. 8 in all current
12464: platforms.
12465:
12466: @item number representation and arithmetic:
12467: @cindex number representation and arithmetic
12468: Processor-dependent. Binary two's complement on all current platforms.
12469:
12470: @item ranges for integer types:
12471: @cindex ranges for integer types
12472: @cindex integer types, ranges
12473: Installation-dependent. Make environmental queries for @code{MAX-N},
12474: @code{MAX-U}, @code{MAX-D} and @code{MAX-UD}. The lower bounds for
12475: unsigned (and positive) types is 0. The lower bound for signed types on
12476: two's complement and one's complement machines machines can be computed
12477: by adding 1 to the upper bound.
12478:
12479: @item read-only data space regions:
12480: @cindex read-only data space regions
12481: @cindex data-space, read-only regions
12482: The whole Forth data space is writable.
12483:
12484: @item size of buffer at @code{WORD}:
12485: @cindex size of buffer at @code{WORD}
12486: @cindex @code{WORD} buffer size
12487: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12488: shared with the pictured numeric output string. If overwriting
12489: @code{PAD} is acceptable, it is as large as the remaining dictionary
12490: space, although only as much can be sensibly used as fits in a counted
12491: string.
12492:
12493: @item size of one cell in address units:
12494: @cindex cell size
12495: @code{1 cells .}.
12496:
12497: @item size of one character in address units:
12498: @cindex char size
12499: @code{1 chars .}. 1 on all current platforms.
12500:
12501: @item size of the keyboard terminal buffer:
12502: @cindex size of the keyboard terminal buffer
12503: @cindex terminal buffer, size
12504: Varies. You can determine the size at a specific time using @code{lp@@
12505: tib - .}. It is shared with the locals stack and TIBs of files that
12506: include the current file. You can change the amount of space for TIBs
12507: and locals stack at Gforth startup with the command line option
12508: @code{-l}.
12509:
12510: @item size of the pictured numeric output buffer:
12511: @cindex size of the pictured numeric output buffer
12512: @cindex pictured numeric output buffer, size
12513: @code{PAD HERE - .}. 104 characters on 32-bit machines. The buffer is
12514: shared with @code{WORD}.
12515:
12516: @item size of the scratch area returned by @code{PAD}:
12517: @cindex size of the scratch area returned by @code{PAD}
12518: @cindex @code{PAD} size
12519: The remainder of dictionary space. @code{unused pad here - - .}.
12520:
12521: @item system case-sensitivity characteristics:
12522: @cindex case-sensitivity characteristics
12523: Dictionary searches are case-insensitive (except in
12524: @code{TABLE}s). However, as explained above under @i{character-set
12525: extensions}, the matching for non-ASCII characters is determined by the
12526: locale you are using. In the default @code{C} locale all non-ASCII
12527: characters are matched case-sensitively.
12528:
12529: @item system prompt:
12530: @cindex system prompt
12531: @cindex prompt
12532: @code{ ok} in interpret state, @code{ compiled} in compile state.
12533:
12534: @item division rounding:
12535: @cindex division rounding
12536: installation dependent. @code{s" floored" environment? drop .}. We leave
12537: the choice to @code{gcc} (what to use for @code{/}) and to you (whether
12538: to use @code{fm/mod}, @code{sm/rem} or simply @code{/}).
12539:
12540: @item values of @code{STATE} when true:
12541: @cindex @code{STATE} values
12542: -1.
12543:
12544: @item values returned after arithmetic overflow:
12545: On two's complement machines, arithmetic is performed modulo
12546: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12547: arithmetic (with appropriate mapping for signed types). Division by zero
12548: typically results in a @code{-55 throw} (Floating-point unidentified
12549: fault) or @code{-10 throw} (divide by zero).
12550:
12551: @item whether the current definition can be found after @t{DOES>}:
12552: @cindex @t{DOES>}, visibility of current definition
12553: No.
12554:
12555: @end table
12556:
12557: @c ---------------------------------------------------------------------
12558: @node core-ambcond, core-other, core-idef, The Core Words
12559: @subsection Ambiguous conditions
12560: @c ---------------------------------------------------------------------
12561: @cindex core words, ambiguous conditions
12562: @cindex ambiguous conditions, core words
12563:
12564: @table @i
12565:
12566: @item a name is neither a word nor a number:
12567: @cindex name not found
12568: @cindex undefined word
12569: @code{-13 throw} (Undefined word).
12570:
12571: @item a definition name exceeds the maximum length allowed:
12572: @cindex word name too long
12573: @code{-19 throw} (Word name too long)
12574:
12575: @item addressing a region not inside the various data spaces of the forth system:
12576: @cindex Invalid memory address
12577: The stacks, code space and header space are accessible. Machine code space is
12578: typically readable. Accessing other addresses gives results dependent on
12579: the operating system. On decent systems: @code{-9 throw} (Invalid memory
12580: address).
12581:
12582: @item argument type incompatible with parameter:
12583: @cindex argument type mismatch
12584: This is usually not caught. Some words perform checks, e.g., the control
12585: flow words, and issue a @code{ABORT"} or @code{-12 THROW} (Argument type
12586: mismatch).
12587:
12588: @item attempting to obtain the execution token of a word with undefined execution semantics:
12589: @cindex Interpreting a compile-only word, for @code{'} etc.
12590: @cindex execution token of words with undefined execution semantics
12591: @code{-14 throw} (Interpreting a compile-only word). In some cases, you
12592: get an execution token for @code{compile-only-error} (which performs a
12593: @code{-14 throw} when executed).
12594:
12595: @item dividing by zero:
12596: @cindex dividing by zero
12597: @cindex floating point unidentified fault, integer division
12598: On some platforms, this produces a @code{-10 throw} (Division by
12599: zero); on other systems, this typically results in a @code{-55 throw}
12600: (Floating-point unidentified fault).
12601:
12602: @item insufficient data stack or return stack space:
12603: @cindex insufficient data stack or return stack space
12604: @cindex stack overflow
12605: @cindex address alignment exception, stack overflow
12606: @cindex Invalid memory address, stack overflow
12607: Depending on the operating system, the installation, and the invocation
12608: of Gforth, this is either checked by the memory management hardware, or
12609: it is not checked. If it is checked, you typically get a @code{-3 throw}
12610: (Stack overflow), @code{-5 throw} (Return stack overflow), or @code{-9
12611: throw} (Invalid memory address) (depending on the platform and how you
12612: achieved the overflow) as soon as the overflow happens. If it is not
12613: checked, overflows typically result in mysterious illegal memory
12614: accesses, producing @code{-9 throw} (Invalid memory address) or
12615: @code{-23 throw} (Address alignment exception); they might also destroy
12616: the internal data structure of @code{ALLOCATE} and friends, resulting in
12617: various errors in these words.
12618:
12619: @item insufficient space for loop control parameters:
12620: @cindex insufficient space for loop control parameters
12621: Like other return stack overflows.
12622:
12623: @item insufficient space in the dictionary:
12624: @cindex insufficient space in the dictionary
12625: @cindex dictionary overflow
12626: If you try to allot (either directly with @code{allot}, or indirectly
12627: with @code{,}, @code{create} etc.) more memory than available in the
12628: dictionary, you get a @code{-8 throw} (Dictionary overflow). If you try
12629: to access memory beyond the end of the dictionary, the results are
12630: similar to stack overflows.
12631:
12632: @item interpreting a word with undefined interpretation semantics:
12633: @cindex interpreting a word with undefined interpretation semantics
12634: @cindex Interpreting a compile-only word
12635: For some words, we have defined interpretation semantics. For the
12636: others: @code{-14 throw} (Interpreting a compile-only word).
12637:
12638: @item modifying the contents of the input buffer or a string literal:
12639: @cindex modifying the contents of the input buffer or a string literal
12640: These are located in writable memory and can be modified.
12641:
12642: @item overflow of the pictured numeric output string:
12643: @cindex overflow of the pictured numeric output string
12644: @cindex pictured numeric output string, overflow
12645: @code{-17 throw} (Pictured numeric ouput string overflow).
12646:
12647: @item parsed string overflow:
12648: @cindex parsed string overflow
12649: @code{PARSE} cannot overflow. @code{WORD} does not check for overflow.
12650:
12651: @item producing a result out of range:
12652: @cindex result out of range
12653: On two's complement machines, arithmetic is performed modulo
12654: 2**bits-per-cell for single arithmetic and 4**bits-per-cell for double
12655: arithmetic (with appropriate mapping for signed types). Division by zero
12656: typically results in a @code{-10 throw} (divide by zero) or @code{-55
12657: throw} (floating point unidentified fault). @code{convert} and
12658: @code{>number} currently overflow silently.
12659:
12660: @item reading from an empty data or return stack:
12661: @cindex stack empty
12662: @cindex stack underflow
12663: @cindex return stack underflow
12664: The data stack is checked by the outer (aka text) interpreter after
12665: every word executed. If it has underflowed, a @code{-4 throw} (Stack
12666: underflow) is performed. Apart from that, stacks may be checked or not,
12667: depending on operating system, installation, and invocation. If they are
12668: caught by a check, they typically result in @code{-4 throw} (Stack
12669: underflow), @code{-6 throw} (Return stack underflow) or @code{-9 throw}
12670: (Invalid memory address), depending on the platform and which stack
12671: underflows and by how much. Note that even if the system uses checking
12672: (through the MMU), your program may have to underflow by a significant
12673: number of stack items to trigger the reaction (the reason for this is
12674: that the MMU, and therefore the checking, works with a page-size
12675: granularity). If there is no checking, the symptoms resulting from an
12676: underflow are similar to those from an overflow. Unbalanced return
12677: stack errors can result in a variety of symptoms, including @code{-9 throw}
12678: (Invalid memory address) and Illegal Instruction (typically @code{-260
12679: throw}).
12680:
12681: @item unexpected end of the input buffer, resulting in an attempt to use a zero-length string as a name:
12682: @cindex unexpected end of the input buffer
12683: @cindex zero-length string as a name
12684: @cindex Attempt to use zero-length string as a name
12685: @code{Create} and its descendants perform a @code{-16 throw} (Attempt to
12686: use zero-length string as a name). Words like @code{'} probably will not
12687: find what they search. Note that it is possible to create zero-length
12688: names with @code{nextname} (should it not?).
12689:
12690: @item @code{>IN} greater than input buffer:
12691: @cindex @code{>IN} greater than input buffer
12692: The next invocation of a parsing word returns a string with length 0.
12693:
12694: @item @code{RECURSE} appears after @code{DOES>}:
12695: @cindex @code{RECURSE} appears after @code{DOES>}
12696: Compiles a recursive call to the defining word, not to the defined word.
12697:
12698: @item argument input source different than current input source for @code{RESTORE-INPUT}:
12699: @cindex argument input source different than current input source for @code{RESTORE-INPUT}
12700: @cindex argument type mismatch, @code{RESTORE-INPUT}
12701: @cindex @code{RESTORE-INPUT}, Argument type mismatch
12702: @code{-12 THROW}. Note that, once an input file is closed (e.g., because
12703: the end of the file was reached), its source-id may be
12704: reused. Therefore, restoring an input source specification referencing a
12705: closed file may lead to unpredictable results instead of a @code{-12
12706: THROW}.
12707:
12708: In the future, Gforth may be able to restore input source specifications
12709: from other than the current input source.
12710:
12711: @item data space containing definitions gets de-allocated:
12712: @cindex data space containing definitions gets de-allocated
12713: Deallocation with @code{allot} is not checked. This typically results in
12714: memory access faults or execution of illegal instructions.
12715:
12716: @item data space read/write with incorrect alignment:
12717: @cindex data space read/write with incorrect alignment
12718: @cindex alignment faults
12719: @cindex address alignment exception
12720: Processor-dependent. Typically results in a @code{-23 throw} (Address
12721: alignment exception). Under Linux-Intel on a 486 or later processor with
12722: alignment turned on, incorrect alignment results in a @code{-9 throw}
12723: (Invalid memory address). There are reportedly some processors with
12724: alignment restrictions that do not report violations.
12725:
12726: @item data space pointer not properly aligned, @code{,}, @code{C,}:
12727: @cindex data space pointer not properly aligned, @code{,}, @code{C,}
12728: Like other alignment errors.
12729:
12730: @item less than u+2 stack items (@code{PICK} and @code{ROLL}):
12731: Like other stack underflows.
12732:
12733: @item loop control parameters not available:
12734: @cindex loop control parameters not available
12735: Not checked. The counted loop words simply assume that the top of return
12736: stack items are loop control parameters and behave accordingly.
12737:
12738: @item most recent definition does not have a name (@code{IMMEDIATE}):
12739: @cindex most recent definition does not have a name (@code{IMMEDIATE})
12740: @cindex last word was headerless
12741: @code{abort" last word was headerless"}.
12742:
12743: @item name not defined by @code{VALUE} used by @code{TO}:
12744: @cindex name not defined by @code{VALUE} used by @code{TO}
12745: @cindex @code{TO} on non-@code{VALUE}s
12746: @cindex Invalid name argument, @code{TO}
12747: @code{-32 throw} (Invalid name argument) (unless name is a local or was
12748: defined by @code{CONSTANT}; in the latter case it just changes the constant).
12749:
12750: @item name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}):
12751: @cindex name not found (@code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]})
12752: @cindex undefined word, @code{'}, @code{POSTPONE}, @code{[']}, @code{[COMPILE]}
12753: @code{-13 throw} (Undefined word)
12754:
12755: @item parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN}):
12756: @cindex parameters are not of the same type (@code{DO}, @code{?DO}, @code{WITHIN})
12757: Gforth behaves as if they were of the same type. I.e., you can predict
12758: the behaviour by interpreting all parameters as, e.g., signed.
12759:
12760: @item @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}:
12761: @cindex @code{POSTPONE} or @code{[COMPILE]} applied to @code{TO}
12762: Assume @code{: X POSTPONE TO ; IMMEDIATE}. @code{X} performs the
12763: compilation semantics of @code{TO}.
12764:
12765: @item String longer than a counted string returned by @code{WORD}:
12766: @cindex string longer than a counted string returned by @code{WORD}
12767: @cindex @code{WORD}, string overflow
12768: Not checked. The string will be ok, but the count will, of course,
12769: contain only the least significant bits of the length.
12770:
12771: @item u greater than or equal to the number of bits in a cell (@code{LSHIFT}, @code{RSHIFT}):
12772: @cindex @code{LSHIFT}, large shift counts
12773: @cindex @code{RSHIFT}, large shift counts
12774: Processor-dependent. Typical behaviours are returning 0 and using only
12775: the low bits of the shift count.
12776:
12777: @item word not defined via @code{CREATE}:
12778: @cindex @code{>BODY} of non-@code{CREATE}d words
12779: @code{>BODY} produces the PFA of the word no matter how it was defined.
12780:
12781: @cindex @code{DOES>} of non-@code{CREATE}d words
12782: @code{DOES>} changes the execution semantics of the last defined word no
12783: matter how it was defined. E.g., @code{CONSTANT DOES>} is equivalent to
12784: @code{CREATE , DOES>}.
12785:
12786: @item words improperly used outside @code{<#} and @code{#>}:
12787: Not checked. As usual, you can expect memory faults.
12788:
12789: @end table
12790:
12791:
12792: @c ---------------------------------------------------------------------
12793: @node core-other, , core-ambcond, The Core Words
12794: @subsection Other system documentation
12795: @c ---------------------------------------------------------------------
12796: @cindex other system documentation, core words
12797: @cindex core words, other system documentation
12798:
12799: @table @i
12800: @item nonstandard words using @code{PAD}:
12801: @cindex @code{PAD} use by nonstandard words
12802: None.
12803:
12804: @item operator's terminal facilities available:
12805: @cindex operator's terminal facilities available
12806: After processing the OS's command line, Gforth goes into interactive mode,
12807: and you can give commands to Gforth interactively. The actual facilities
12808: available depend on how you invoke Gforth.
12809:
12810: @item program data space available:
12811: @cindex program data space available
12812: @cindex data space available
12813: @code{UNUSED .} gives the remaining dictionary space. The total
12814: dictionary space can be specified with the @code{-m} switch
12815: (@pxref{Invoking Gforth}) when Gforth starts up.
12816:
12817: @item return stack space available:
12818: @cindex return stack space available
12819: You can compute the total return stack space in cells with
12820: @code{s" RETURN-STACK-CELLS" environment? drop .}. You can specify it at
12821: startup time with the @code{-r} switch (@pxref{Invoking Gforth}).
12822:
12823: @item stack space available:
12824: @cindex stack space available
12825: You can compute the total data stack space in cells with
12826: @code{s" STACK-CELLS" environment? drop .}. You can specify it at
12827: startup time with the @code{-d} switch (@pxref{Invoking Gforth}).
12828:
12829: @item system dictionary space required, in address units:
12830: @cindex system dictionary space required, in address units
12831: Type @code{here forthstart - .} after startup. At the time of this
12832: writing, this gives 80080 (bytes) on a 32-bit system.
12833: @end table
12834:
12835:
12836: @c =====================================================================
12837: @node The optional Block word set, The optional Double Number word set, The Core Words, ANS conformance
12838: @section The optional Block word set
12839: @c =====================================================================
12840: @cindex system documentation, block words
12841: @cindex block words, system documentation
12842:
12843: @menu
12844: * block-idef:: Implementation Defined Options
12845: * block-ambcond:: Ambiguous Conditions
12846: * block-other:: Other System Documentation
12847: @end menu
12848:
12849:
12850: @c ---------------------------------------------------------------------
12851: @node block-idef, block-ambcond, The optional Block word set, The optional Block word set
12852: @subsection Implementation Defined Options
12853: @c ---------------------------------------------------------------------
12854: @cindex implementation-defined options, block words
12855: @cindex block words, implementation-defined options
12856:
12857: @table @i
12858: @item the format for display by @code{LIST}:
12859: @cindex @code{LIST} display format
12860: First the screen number is displayed, then 16 lines of 64 characters,
12861: each line preceded by the line number.
12862:
12863: @item the length of a line affected by @code{\}:
12864: @cindex length of a line affected by @code{\}
12865: @cindex @code{\}, line length in blocks
12866: 64 characters.
12867: @end table
12868:
12869:
12870: @c ---------------------------------------------------------------------
12871: @node block-ambcond, block-other, block-idef, The optional Block word set
12872: @subsection Ambiguous conditions
12873: @c ---------------------------------------------------------------------
12874: @cindex block words, ambiguous conditions
12875: @cindex ambiguous conditions, block words
12876:
12877: @table @i
12878: @item correct block read was not possible:
12879: @cindex block read not possible
12880: Typically results in a @code{throw} of some OS-derived value (between
12881: -512 and -2048). If the blocks file was just not long enough, blanks are
12882: supplied for the missing portion.
12883:
12884: @item I/O exception in block transfer:
12885: @cindex I/O exception in block transfer
12886: @cindex block transfer, I/O exception
12887: Typically results in a @code{throw} of some OS-derived value (between
12888: -512 and -2048).
12889:
12890: @item invalid block number:
12891: @cindex invalid block number
12892: @cindex block number invalid
12893: @code{-35 throw} (Invalid block number)
12894:
12895: @item a program directly alters the contents of @code{BLK}:
12896: @cindex @code{BLK}, altering @code{BLK}
12897: The input stream is switched to that other block, at the same
12898: position. If the storing to @code{BLK} happens when interpreting
12899: non-block input, the system will get quite confused when the block ends.
12900:
12901: @item no current block buffer for @code{UPDATE}:
12902: @cindex @code{UPDATE}, no current block buffer
12903: @code{UPDATE} has no effect.
12904:
12905: @end table
12906:
12907: @c ---------------------------------------------------------------------
12908: @node block-other, , block-ambcond, The optional Block word set
12909: @subsection Other system documentation
12910: @c ---------------------------------------------------------------------
12911: @cindex other system documentation, block words
12912: @cindex block words, other system documentation
12913:
12914: @table @i
12915: @item any restrictions a multiprogramming system places on the use of buffer addresses:
12916: No restrictions (yet).
12917:
12918: @item the number of blocks available for source and data:
12919: depends on your disk space.
12920:
12921: @end table
12922:
12923:
12924: @c =====================================================================
12925: @node The optional Double Number word set, The optional Exception word set, The optional Block word set, ANS conformance
12926: @section The optional Double Number word set
12927: @c =====================================================================
12928: @cindex system documentation, double words
12929: @cindex double words, system documentation
12930:
12931: @menu
12932: * double-ambcond:: Ambiguous Conditions
12933: @end menu
12934:
12935:
12936: @c ---------------------------------------------------------------------
12937: @node double-ambcond, , The optional Double Number word set, The optional Double Number word set
12938: @subsection Ambiguous conditions
12939: @c ---------------------------------------------------------------------
12940: @cindex double words, ambiguous conditions
12941: @cindex ambiguous conditions, double words
12942:
12943: @table @i
12944: @item @i{d} outside of range of @i{n} in @code{D>S}:
12945: @cindex @code{D>S}, @i{d} out of range of @i{n}
12946: The least significant cell of @i{d} is produced.
12947:
12948: @end table
12949:
12950:
12951: @c =====================================================================
12952: @node The optional Exception word set, The optional Facility word set, The optional Double Number word set, ANS conformance
12953: @section The optional Exception word set
12954: @c =====================================================================
12955: @cindex system documentation, exception words
12956: @cindex exception words, system documentation
12957:
12958: @menu
12959: * exception-idef:: Implementation Defined Options
12960: @end menu
12961:
12962:
12963: @c ---------------------------------------------------------------------
12964: @node exception-idef, , The optional Exception word set, The optional Exception word set
12965: @subsection Implementation Defined Options
12966: @c ---------------------------------------------------------------------
12967: @cindex implementation-defined options, exception words
12968: @cindex exception words, implementation-defined options
12969:
12970: @table @i
12971: @item @code{THROW}-codes used in the system:
12972: @cindex @code{THROW}-codes used in the system
12973: The codes -256@minus{}-511 are used for reporting signals. The mapping
12974: from OS signal numbers to throw codes is -256@minus{}@i{signal}. The
12975: codes -512@minus{}-2047 are used for OS errors (for file and memory
12976: allocation operations). The mapping from OS error numbers to throw codes
12977: is -512@minus{}@code{errno}. One side effect of this mapping is that
12978: undefined OS errors produce a message with a strange number; e.g.,
12979: @code{-1000 THROW} results in @code{Unknown error 488} on my system.
12980: @end table
12981:
12982: @c =====================================================================
12983: @node The optional Facility word set, The optional File-Access word set, The optional Exception word set, ANS conformance
12984: @section The optional Facility word set
12985: @c =====================================================================
12986: @cindex system documentation, facility words
12987: @cindex facility words, system documentation
12988:
12989: @menu
12990: * facility-idef:: Implementation Defined Options
12991: * facility-ambcond:: Ambiguous Conditions
12992: @end menu
12993:
12994:
12995: @c ---------------------------------------------------------------------
12996: @node facility-idef, facility-ambcond, The optional Facility word set, The optional Facility word set
12997: @subsection Implementation Defined Options
12998: @c ---------------------------------------------------------------------
12999: @cindex implementation-defined options, facility words
13000: @cindex facility words, implementation-defined options
13001:
13002: @table @i
13003: @item encoding of keyboard events (@code{EKEY}):
13004: @cindex keyboard events, encoding in @code{EKEY}
13005: @cindex @code{EKEY}, encoding of keyboard events
13006: Keys corresponding to ASCII characters are encoded as ASCII characters.
13007: Other keys are encoded with the constants @code{k-left}, @code{k-right},
13008: @code{k-up}, @code{k-down}, @code{k-home}, @code{k-end}, @code{k1},
13009: @code{k2}, @code{k3}, @code{k4}, @code{k5}, @code{k6}, @code{k7},
13010: @code{k8}, @code{k9}, @code{k10}, @code{k11}, @code{k12}.
13011:
13012:
13013: @item duration of a system clock tick:
13014: @cindex duration of a system clock tick
13015: @cindex clock tick duration
13016: System dependent. With respect to @code{MS}, the time is specified in
13017: microseconds. How well the OS and the hardware implement this, is
13018: another question.
13019:
13020: @item repeatability to be expected from the execution of @code{MS}:
13021: @cindex repeatability to be expected from the execution of @code{MS}
13022: @cindex @code{MS}, repeatability to be expected
13023: System dependent. On Unix, a lot depends on load. If the system is
13024: lightly loaded, and the delay is short enough that Gforth does not get
13025: swapped out, the performance should be acceptable. Under MS-DOS and
13026: other single-tasking systems, it should be good.
13027:
13028: @end table
13029:
13030:
13031: @c ---------------------------------------------------------------------
13032: @node facility-ambcond, , facility-idef, The optional Facility word set
13033: @subsection Ambiguous conditions
13034: @c ---------------------------------------------------------------------
13035: @cindex facility words, ambiguous conditions
13036: @cindex ambiguous conditions, facility words
13037:
13038: @table @i
13039: @item @code{AT-XY} can't be performed on user output device:
13040: @cindex @code{AT-XY} can't be performed on user output device
13041: Largely terminal dependent. No range checks are done on the arguments.
13042: No errors are reported. You may see some garbage appearing, you may see
13043: simply nothing happen.
13044:
13045: @end table
13046:
13047:
13048: @c =====================================================================
13049: @node The optional File-Access word set, The optional Floating-Point word set, The optional Facility word set, ANS conformance
13050: @section The optional File-Access word set
13051: @c =====================================================================
13052: @cindex system documentation, file words
13053: @cindex file words, system documentation
13054:
13055: @menu
13056: * file-idef:: Implementation Defined Options
13057: * file-ambcond:: Ambiguous Conditions
13058: @end menu
13059:
13060: @c ---------------------------------------------------------------------
13061: @node file-idef, file-ambcond, The optional File-Access word set, The optional File-Access word set
13062: @subsection Implementation Defined Options
13063: @c ---------------------------------------------------------------------
13064: @cindex implementation-defined options, file words
13065: @cindex file words, implementation-defined options
13066:
13067: @table @i
13068: @item file access methods used:
13069: @cindex file access methods used
13070: @code{R/O}, @code{R/W} and @code{BIN} work as you would
13071: expect. @code{W/O} translates into the C file opening mode @code{w} (or
13072: @code{wb}): The file is cleared, if it exists, and created, if it does
13073: not (with both @code{open-file} and @code{create-file}). Under Unix
13074: @code{create-file} creates a file with 666 permissions modified by your
13075: umask.
13076:
13077: @item file exceptions:
13078: @cindex file exceptions
13079: The file words do not raise exceptions (except, perhaps, memory access
13080: faults when you pass illegal addresses or file-ids).
13081:
13082: @item file line terminator:
13083: @cindex file line terminator
13084: System-dependent. Gforth uses C's newline character as line
13085: terminator. What the actual character code(s) of this are is
13086: system-dependent.
13087:
13088: @item file name format:
13089: @cindex file name format
13090: System dependent. Gforth just uses the file name format of your OS.
13091:
13092: @item information returned by @code{FILE-STATUS}:
13093: @cindex @code{FILE-STATUS}, returned information
13094: @code{FILE-STATUS} returns the most powerful file access mode allowed
13095: for the file: Either @code{R/O}, @code{W/O} or @code{R/W}. If the file
13096: cannot be accessed, @code{R/O BIN} is returned. @code{BIN} is applicable
13097: along with the returned mode.
13098:
13099: @item input file state after an exception when including source:
13100: @cindex exception when including source
13101: All files that are left via the exception are closed.
13102:
13103: @item @i{ior} values and meaning:
13104: @cindex @i{ior} values and meaning
13105: @cindex @i{wior} values and meaning
13106: The @i{ior}s returned by the file and memory allocation words are
13107: intended as throw codes. They typically are in the range
13108: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
13109: @i{ior}s is -512@minus{}@i{errno}.
13110:
13111: @item maximum depth of file input nesting:
13112: @cindex maximum depth of file input nesting
13113: @cindex file input nesting, maximum depth
13114: limited by the amount of return stack, locals/TIB stack, and the number
13115: of open files available. This should not give you troubles.
13116:
13117: @item maximum size of input line:
13118: @cindex maximum size of input line
13119: @cindex input line size, maximum
13120: @code{/line}. Currently 255.
13121:
13122: @item methods of mapping block ranges to files:
13123: @cindex mapping block ranges to files
13124: @cindex files containing blocks
13125: @cindex blocks in files
13126: By default, blocks are accessed in the file @file{blocks.fb} in the
13127: current working directory. The file can be switched with @code{USE}.
13128:
13129: @item number of string buffers provided by @code{S"}:
13130: @cindex @code{S"}, number of string buffers
13131: 1
13132:
13133: @item size of string buffer used by @code{S"}:
13134: @cindex @code{S"}, size of string buffer
13135: @code{/line}. currently 255.
13136:
13137: @end table
13138:
13139: @c ---------------------------------------------------------------------
13140: @node file-ambcond, , file-idef, The optional File-Access word set
13141: @subsection Ambiguous conditions
13142: @c ---------------------------------------------------------------------
13143: @cindex file words, ambiguous conditions
13144: @cindex ambiguous conditions, file words
13145:
13146: @table @i
13147: @item attempting to position a file outside its boundaries:
13148: @cindex @code{REPOSITION-FILE}, outside the file's boundaries
13149: @code{REPOSITION-FILE} is performed as usual: Afterwards,
13150: @code{FILE-POSITION} returns the value given to @code{REPOSITION-FILE}.
13151:
13152: @item attempting to read from file positions not yet written:
13153: @cindex reading from file positions not yet written
13154: End-of-file, i.e., zero characters are read and no error is reported.
13155:
13156: @item @i{file-id} is invalid (@code{INCLUDE-FILE}):
13157: @cindex @code{INCLUDE-FILE}, @i{file-id} is invalid
13158: An appropriate exception may be thrown, but a memory fault or other
13159: problem is more probable.
13160:
13161: @item I/O exception reading or closing @i{file-id} (@code{INCLUDE-FILE}, @code{INCLUDED}):
13162: @cindex @code{INCLUDE-FILE}, I/O exception reading or closing @i{file-id}
13163: @cindex @code{INCLUDED}, I/O exception reading or closing @i{file-id}
13164: The @i{ior} produced by the operation, that discovered the problem, is
13165: thrown.
13166:
13167: @item named file cannot be opened (@code{INCLUDED}):
13168: @cindex @code{INCLUDED}, named file cannot be opened
13169: The @i{ior} produced by @code{open-file} is thrown.
13170:
13171: @item requesting an unmapped block number:
13172: @cindex unmapped block numbers
13173: There are no unmapped legal block numbers. On some operating systems,
13174: writing a block with a large number may overflow the file system and
13175: have an error message as consequence.
13176:
13177: @item using @code{source-id} when @code{blk} is non-zero:
13178: @cindex @code{SOURCE-ID}, behaviour when @code{BLK} is non-zero
13179: @code{source-id} performs its function. Typically it will give the id of
13180: the source which loaded the block. (Better ideas?)
13181:
13182: @end table
13183:
13184:
13185: @c =====================================================================
13186: @node The optional Floating-Point word set, The optional Locals word set, The optional File-Access word set, ANS conformance
13187: @section The optional Floating-Point word set
13188: @c =====================================================================
13189: @cindex system documentation, floating-point words
13190: @cindex floating-point words, system documentation
13191:
13192: @menu
13193: * floating-idef:: Implementation Defined Options
13194: * floating-ambcond:: Ambiguous Conditions
13195: @end menu
13196:
13197:
13198: @c ---------------------------------------------------------------------
13199: @node floating-idef, floating-ambcond, The optional Floating-Point word set, The optional Floating-Point word set
13200: @subsection Implementation Defined Options
13201: @c ---------------------------------------------------------------------
13202: @cindex implementation-defined options, floating-point words
13203: @cindex floating-point words, implementation-defined options
13204:
13205: @table @i
13206: @item format and range of floating point numbers:
13207: @cindex format and range of floating point numbers
13208: @cindex floating point numbers, format and range
13209: System-dependent; the @code{double} type of C.
13210:
13211: @item results of @code{REPRESENT} when @i{float} is out of range:
13212: @cindex @code{REPRESENT}, results when @i{float} is out of range
13213: System dependent; @code{REPRESENT} is implemented using the C library
13214: function @code{ecvt()} and inherits its behaviour in this respect.
13215:
13216: @item rounding or truncation of floating-point numbers:
13217: @cindex rounding of floating-point numbers
13218: @cindex truncation of floating-point numbers
13219: @cindex floating-point numbers, rounding or truncation
13220: System dependent; the rounding behaviour is inherited from the hosting C
13221: compiler. IEEE-FP-based (i.e., most) systems by default round to
13222: nearest, and break ties by rounding to even (i.e., such that the last
13223: bit of the mantissa is 0).
13224:
13225: @item size of floating-point stack:
13226: @cindex floating-point stack size
13227: @code{s" FLOATING-STACK" environment? drop .} gives the total size of
13228: the floating-point stack (in floats). You can specify this on startup
13229: with the command-line option @code{-f} (@pxref{Invoking Gforth}).
13230:
13231: @item width of floating-point stack:
13232: @cindex floating-point stack width
13233: @code{1 floats}.
13234:
13235: @end table
13236:
13237:
13238: @c ---------------------------------------------------------------------
13239: @node floating-ambcond, , floating-idef, The optional Floating-Point word set
13240: @subsection Ambiguous conditions
13241: @c ---------------------------------------------------------------------
13242: @cindex floating-point words, ambiguous conditions
13243: @cindex ambiguous conditions, floating-point words
13244:
13245: @table @i
13246: @item @code{df@@} or @code{df!} used with an address that is not double-float aligned:
13247: @cindex @code{df@@} or @code{df!} used with an address that is not double-float aligned
13248: System-dependent. Typically results in a @code{-23 THROW} like other
13249: alignment violations.
13250:
13251: @item @code{f@@} or @code{f!} used with an address that is not float aligned:
13252: @cindex @code{f@@} used with an address that is not float aligned
13253: @cindex @code{f!} used with an address that is not float aligned
13254: System-dependent. Typically results in a @code{-23 THROW} like other
13255: alignment violations.
13256:
13257: @item floating-point result out of range:
13258: @cindex floating-point result out of range
13259: System-dependent. Can result in a @code{-43 throw} (floating point
13260: overflow), @code{-54 throw} (floating point underflow), @code{-41 throw}
13261: (floating point inexact result), @code{-55 THROW} (Floating-point
13262: unidentified fault), or can produce a special value representing, e.g.,
13263: Infinity.
13264:
13265: @item @code{sf@@} or @code{sf!} used with an address that is not single-float aligned:
13266: @cindex @code{sf@@} or @code{sf!} used with an address that is not single-float aligned
13267: System-dependent. Typically results in an alignment fault like other
13268: alignment violations.
13269:
13270: @item @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.}):
13271: @cindex @code{base} is not decimal (@code{REPRESENT}, @code{F.}, @code{FE.}, @code{FS.})
13272: The floating-point number is converted into decimal nonetheless.
13273:
13274: @item Both arguments are equal to zero (@code{FATAN2}):
13275: @cindex @code{FATAN2}, both arguments are equal to zero
13276: System-dependent. @code{FATAN2} is implemented using the C library
13277: function @code{atan2()}.
13278:
13279: @item Using @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero:
13280: @cindex @code{FTAN} on an argument @i{r1} where cos(@i{r1}) is zero
13281: System-dependent. Anyway, typically the cos of @i{r1} will not be zero
13282: because of small errors and the tan will be a very large (or very small)
13283: but finite number.
13284:
13285: @item @i{d} cannot be presented precisely as a float in @code{D>F}:
13286: @cindex @code{D>F}, @i{d} cannot be presented precisely as a float
13287: The result is rounded to the nearest float.
13288:
13289: @item dividing by zero:
13290: @cindex dividing by zero, floating-point
13291: @cindex floating-point dividing by zero
13292: @cindex floating-point unidentified fault, FP divide-by-zero
13293: Platform-dependent; can produce an Infinity, NaN, @code{-42 throw}
13294: (floating point divide by zero) or @code{-55 throw} (Floating-point
13295: unidentified fault).
13296:
13297: @item exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@}):
13298: @cindex exponent too big for conversion (@code{DF!}, @code{DF@@}, @code{SF!}, @code{SF@@})
13299: System dependent. On IEEE-FP based systems the number is converted into
13300: an infinity.
13301:
13302: @item @i{float}<1 (@code{FACOSH}):
13303: @cindex @code{FACOSH}, @i{float}<1
13304: @cindex floating-point unidentified fault, @code{FACOSH}
13305: Platform-dependent; on IEEE-FP systems typically produces a NaN.
13306:
13307: @item @i{float}=<-1 (@code{FLNP1}):
13308: @cindex @code{FLNP1}, @i{float}=<-1
13309: @cindex floating-point unidentified fault, @code{FLNP1}
13310: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13311: negative infinity for @i{float}=-1).
13312:
13313: @item @i{float}=<0 (@code{FLN}, @code{FLOG}):
13314: @cindex @code{FLN}, @i{float}=<0
13315: @cindex @code{FLOG}, @i{float}=<0
13316: @cindex floating-point unidentified fault, @code{FLN} or @code{FLOG}
13317: Platform-dependent; on IEEE-FP systems typically produces a NaN (or a
13318: negative infinity for @i{float}=0).
13319:
13320: @item @i{float}<0 (@code{FASINH}, @code{FSQRT}):
13321: @cindex @code{FASINH}, @i{float}<0
13322: @cindex @code{FSQRT}, @i{float}<0
13323: @cindex floating-point unidentified fault, @code{FASINH} or @code{FSQRT}
13324: Platform-dependent; for @code{fsqrt} this typically gives a NaN, for
13325: @code{fasinh} some platforms produce a NaN, others a number (bug in the
13326: C library?).
13327:
13328: @item |@i{float}|>1 (@code{FACOS}, @code{FASIN}, @code{FATANH}):
13329: @cindex @code{FACOS}, |@i{float}|>1
13330: @cindex @code{FASIN}, |@i{float}|>1
13331: @cindex @code{FATANH}, |@i{float}|>1
13332: @cindex floating-point unidentified fault, @code{FACOS}, @code{FASIN} or @code{FATANH}
13333: Platform-dependent; IEEE-FP systems typically produce a NaN.
13334:
13335: @item integer part of float cannot be represented by @i{d} in @code{F>D}:
13336: @cindex @code{F>D}, integer part of float cannot be represented by @i{d}
13337: @cindex floating-point unidentified fault, @code{F>D}
13338: Platform-dependent; typically, some double number is produced and no
13339: error is reported.
13340:
13341: @item string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.}):
13342: @cindex string larger than pictured numeric output area (@code{f.}, @code{fe.}, @code{fs.})
13343: @code{Precision} characters of the numeric output area are used. If
13344: @code{precision} is too high, these words will smash the data or code
13345: close to @code{here}.
13346: @end table
13347:
13348: @c =====================================================================
13349: @node The optional Locals word set, The optional Memory-Allocation word set, The optional Floating-Point word set, ANS conformance
13350: @section The optional Locals word set
13351: @c =====================================================================
13352: @cindex system documentation, locals words
13353: @cindex locals words, system documentation
13354:
13355: @menu
13356: * locals-idef:: Implementation Defined Options
13357: * locals-ambcond:: Ambiguous Conditions
13358: @end menu
13359:
13360:
13361: @c ---------------------------------------------------------------------
13362: @node locals-idef, locals-ambcond, The optional Locals word set, The optional Locals word set
13363: @subsection Implementation Defined Options
13364: @c ---------------------------------------------------------------------
13365: @cindex implementation-defined options, locals words
13366: @cindex locals words, implementation-defined options
13367:
13368: @table @i
13369: @item maximum number of locals in a definition:
13370: @cindex maximum number of locals in a definition
13371: @cindex locals, maximum number in a definition
13372: @code{s" #locals" environment? drop .}. Currently 15. This is a lower
13373: bound, e.g., on a 32-bit machine there can be 41 locals of up to 8
13374: characters. The number of locals in a definition is bounded by the size
13375: of locals-buffer, which contains the names of the locals.
13376:
13377: @end table
13378:
13379:
13380: @c ---------------------------------------------------------------------
13381: @node locals-ambcond, , locals-idef, The optional Locals word set
13382: @subsection Ambiguous conditions
13383: @c ---------------------------------------------------------------------
13384: @cindex locals words, ambiguous conditions
13385: @cindex ambiguous conditions, locals words
13386:
13387: @table @i
13388: @item executing a named local in interpretation state:
13389: @cindex local in interpretation state
13390: @cindex Interpreting a compile-only word, for a local
13391: Locals have no interpretation semantics. If you try to perform the
13392: interpretation semantics, you will get a @code{-14 throw} somewhere
13393: (Interpreting a compile-only word). If you perform the compilation
13394: semantics, the locals access will be compiled (irrespective of state).
13395:
13396: @item @i{name} not defined by @code{VALUE} or @code{(LOCAL)} (@code{TO}):
13397: @cindex name not defined by @code{VALUE} or @code{(LOCAL)} used by @code{TO}
13398: @cindex @code{TO} on non-@code{VALUE}s and non-locals
13399: @cindex Invalid name argument, @code{TO}
13400: @code{-32 throw} (Invalid name argument)
13401:
13402: @end table
13403:
13404:
13405: @c =====================================================================
13406: @node The optional Memory-Allocation word set, The optional Programming-Tools word set, The optional Locals word set, ANS conformance
13407: @section The optional Memory-Allocation word set
13408: @c =====================================================================
13409: @cindex system documentation, memory-allocation words
13410: @cindex memory-allocation words, system documentation
13411:
13412: @menu
13413: * memory-idef:: Implementation Defined Options
13414: @end menu
13415:
13416:
13417: @c ---------------------------------------------------------------------
13418: @node memory-idef, , The optional Memory-Allocation word set, The optional Memory-Allocation word set
13419: @subsection Implementation Defined Options
13420: @c ---------------------------------------------------------------------
13421: @cindex implementation-defined options, memory-allocation words
13422: @cindex memory-allocation words, implementation-defined options
13423:
13424: @table @i
13425: @item values and meaning of @i{ior}:
13426: @cindex @i{ior} values and meaning
13427: The @i{ior}s returned by the file and memory allocation words are
13428: intended as throw codes. They typically are in the range
13429: -512@minus{}-2047 of OS errors. The mapping from OS error numbers to
13430: @i{ior}s is -512@minus{}@i{errno}.
13431:
13432: @end table
13433:
13434: @c =====================================================================
13435: @node The optional Programming-Tools word set, The optional Search-Order word set, The optional Memory-Allocation word set, ANS conformance
13436: @section The optional Programming-Tools word set
13437: @c =====================================================================
13438: @cindex system documentation, programming-tools words
13439: @cindex programming-tools words, system documentation
13440:
13441: @menu
13442: * programming-idef:: Implementation Defined Options
13443: * programming-ambcond:: Ambiguous Conditions
13444: @end menu
13445:
13446:
13447: @c ---------------------------------------------------------------------
13448: @node programming-idef, programming-ambcond, The optional Programming-Tools word set, The optional Programming-Tools word set
13449: @subsection Implementation Defined Options
13450: @c ---------------------------------------------------------------------
13451: @cindex implementation-defined options, programming-tools words
13452: @cindex programming-tools words, implementation-defined options
13453:
13454: @table @i
13455: @item ending sequence for input following @code{;CODE} and @code{CODE}:
13456: @cindex @code{;CODE} ending sequence
13457: @cindex @code{CODE} ending sequence
13458: @code{END-CODE}
13459:
13460: @item manner of processing input following @code{;CODE} and @code{CODE}:
13461: @cindex @code{;CODE}, processing input
13462: @cindex @code{CODE}, processing input
13463: The @code{ASSEMBLER} vocabulary is pushed on the search order stack, and
13464: the input is processed by the text interpreter, (starting) in interpret
13465: state.
13466:
13467: @item search order capability for @code{EDITOR} and @code{ASSEMBLER}:
13468: @cindex @code{ASSEMBLER}, search order capability
13469: The ANS Forth search order word set.
13470:
13471: @item source and format of display by @code{SEE}:
13472: @cindex @code{SEE}, source and format of output
13473: The source for @code{see} is the executable code used by the inner
13474: interpreter. The current @code{see} tries to output Forth source code
13475: (and on some platforms, assembly code for primitives) as well as
13476: possible.
13477:
13478: @end table
13479:
13480: @c ---------------------------------------------------------------------
13481: @node programming-ambcond, , programming-idef, The optional Programming-Tools word set
13482: @subsection Ambiguous conditions
13483: @c ---------------------------------------------------------------------
13484: @cindex programming-tools words, ambiguous conditions
13485: @cindex ambiguous conditions, programming-tools words
13486:
13487: @table @i
13488:
13489: @item deleting the compilation word list (@code{FORGET}):
13490: @cindex @code{FORGET}, deleting the compilation word list
13491: Not implemented (yet).
13492:
13493: @item fewer than @i{u}+1 items on the control-flow stack (@code{CS-PICK}, @code{CS-ROLL}):
13494: @cindex @code{CS-PICK}, fewer than @i{u}+1 items on the control flow-stack
13495: @cindex @code{CS-ROLL}, fewer than @i{u}+1 items on the control flow-stack
13496: @cindex control-flow stack underflow
13497: This typically results in an @code{abort"} with a descriptive error
13498: message (may change into a @code{-22 throw} (Control structure mismatch)
13499: in the future). You may also get a memory access error. If you are
13500: unlucky, this ambiguous condition is not caught.
13501:
13502: @item @i{name} can't be found (@code{FORGET}):
13503: @cindex @code{FORGET}, @i{name} can't be found
13504: Not implemented (yet).
13505:
13506: @item @i{name} not defined via @code{CREATE}:
13507: @cindex @code{;CODE}, @i{name} not defined via @code{CREATE}
13508: @code{;CODE} behaves like @code{DOES>} in this respect, i.e., it changes
13509: the execution semantics of the last defined word no matter how it was
13510: defined.
13511:
13512: @item @code{POSTPONE} applied to @code{[IF]}:
13513: @cindex @code{POSTPONE} applied to @code{[IF]}
13514: @cindex @code{[IF]} and @code{POSTPONE}
13515: After defining @code{: X POSTPONE [IF] ; IMMEDIATE}. @code{X} is
13516: equivalent to @code{[IF]}.
13517:
13518: @item reaching the end of the input source before matching @code{[ELSE]} or @code{[THEN]}:
13519: @cindex @code{[IF]}, end of the input source before matching @code{[ELSE]} or @code{[THEN]}
13520: Continue in the same state of conditional compilation in the next outer
13521: input source. Currently there is no warning to the user about this.
13522:
13523: @item removing a needed definition (@code{FORGET}):
13524: @cindex @code{FORGET}, removing a needed definition
13525: Not implemented (yet).
13526:
13527: @end table
13528:
13529:
13530: @c =====================================================================
13531: @node The optional Search-Order word set, , The optional Programming-Tools word set, ANS conformance
13532: @section The optional Search-Order word set
13533: @c =====================================================================
13534: @cindex system documentation, search-order words
13535: @cindex search-order words, system documentation
13536:
13537: @menu
13538: * search-idef:: Implementation Defined Options
13539: * search-ambcond:: Ambiguous Conditions
13540: @end menu
13541:
13542:
13543: @c ---------------------------------------------------------------------
13544: @node search-idef, search-ambcond, The optional Search-Order word set, The optional Search-Order word set
13545: @subsection Implementation Defined Options
13546: @c ---------------------------------------------------------------------
13547: @cindex implementation-defined options, search-order words
13548: @cindex search-order words, implementation-defined options
13549:
13550: @table @i
13551: @item maximum number of word lists in search order:
13552: @cindex maximum number of word lists in search order
13553: @cindex search order, maximum depth
13554: @code{s" wordlists" environment? drop .}. Currently 16.
13555:
13556: @item minimum search order:
13557: @cindex minimum search order
13558: @cindex search order, minimum
13559: @code{root root}.
13560:
13561: @end table
13562:
13563: @c ---------------------------------------------------------------------
13564: @node search-ambcond, , search-idef, The optional Search-Order word set
13565: @subsection Ambiguous conditions
13566: @c ---------------------------------------------------------------------
13567: @cindex search-order words, ambiguous conditions
13568: @cindex ambiguous conditions, search-order words
13569:
13570: @table @i
13571: @item changing the compilation word list (during compilation):
13572: @cindex changing the compilation word list (during compilation)
13573: @cindex compilation word list, change before definition ends
13574: The word is entered into the word list that was the compilation word list
13575: at the start of the definition. Any changes to the name field (e.g.,
13576: @code{immediate}) or the code field (e.g., when executing @code{DOES>})
13577: are applied to the latest defined word (as reported by @code{last} or
13578: @code{lastxt}), if possible, irrespective of the compilation word list.
13579:
13580: @item search order empty (@code{previous}):
13581: @cindex @code{previous}, search order empty
13582: @cindex vocstack empty, @code{previous}
13583: @code{abort" Vocstack empty"}.
13584:
13585: @item too many word lists in search order (@code{also}):
13586: @cindex @code{also}, too many word lists in search order
13587: @cindex vocstack full, @code{also}
13588: @code{abort" Vocstack full"}.
13589:
13590: @end table
13591:
13592: @c ***************************************************************
13593: @node Standard vs Extensions, Model, ANS conformance, Top
13594: @chapter Should I use Gforth extensions?
13595: @cindex Gforth extensions
13596:
13597: As you read through the rest of this manual, you will see documentation
13598: for @i{Standard} words, and documentation for some appealing Gforth
13599: @i{extensions}. You might ask yourself the question: @i{``Should I
13600: restrict myself to the standard, or should I use the extensions?''}
13601:
13602: The answer depends on the goals you have for the program you are working
13603: on:
13604:
13605: @itemize @bullet
13606:
13607: @item Is it just for yourself or do you want to share it with others?
13608:
13609: @item
13610: If you want to share it, do the others all use Gforth?
13611:
13612: @item
13613: If it is just for yourself, do you want to restrict yourself to Gforth?
13614:
13615: @end itemize
13616:
13617: If restricting the program to Gforth is ok, then there is no reason not
13618: to use extensions. It is still a good idea to keep to the standard
13619: where it is easy, in case you want to reuse these parts in another
13620: program that you want to be portable.
13621:
13622: If you want to be able to port the program to other Forth systems, there
13623: are the following points to consider:
13624:
13625: @itemize @bullet
13626:
13627: @item
13628: Most Forth systems that are being maintained support the ANS Forth
13629: standard. So if your program complies with the standard, it will be
13630: portable among many systems.
13631:
13632: @item
13633: A number of the Gforth extensions can be implemented in ANS Forth using
13634: public-domain files provided in the @file{compat/} directory. These are
13635: mentioned in the text in passing. There is no reason not to use these
13636: extensions, your program will still be ANS Forth compliant; just include
13637: the appropriate compat files with your program.
13638:
13639: @item
13640: The tool @file{ans-report.fs} (@pxref{ANS Report}) makes it easy to
13641: analyse your program and determine what non-Standard words it relies
13642: upon. However, it does not check whether you use standard words in a
13643: non-standard way.
13644:
13645: @item
13646: Some techniques are not standardized by ANS Forth, and are hard or
13647: impossible to implement in a standard way, but can be implemented in
13648: most Forth systems easily, and usually in similar ways (e.g., accessing
13649: word headers). Forth has a rich historical precedent for programmers
13650: taking advantage of implementation-dependent features of their tools
13651: (for example, relying on a knowledge of the dictionary
13652: structure). Sometimes these techniques are necessary to extract every
13653: last bit of performance from the hardware, sometimes they are just a
13654: programming shorthand.
13655:
13656: @item
13657: Does using a Gforth extension save more work than the porting this part
13658: to other Forth systems (if any) will cost?
13659:
13660: @item
13661: Is the additional functionality worth the reduction in portability and
13662: the additional porting problems?
13663:
13664: @end itemize
13665:
13666: In order to perform these consideratios, you need to know what's
13667: standard and what's not. This manual generally states if something is
13668: non-standard, but the authoritative source is the
13669: @uref{http://www.taygeta.com/forth/dpans.html,standard document}.
13670: Appendix A of the Standard (@var{Rationale}) provides a valuable insight
13671: into the thought processes of the technical committee.
13672:
13673: Note also that portability between Forth systems is not the only
13674: portability issue; there is also the issue of portability between
13675: different platforms (processor/OS combinations).
13676:
13677: @c ***************************************************************
13678: @node Model, Integrating Gforth, Standard vs Extensions, Top
13679: @chapter Model
13680:
13681: This chapter has yet to be written. It will contain information, on
13682: which internal structures you can rely.
13683:
13684: @c ***************************************************************
13685: @node Integrating Gforth, Emacs and Gforth, Model, Top
13686: @chapter Integrating Gforth into C programs
13687:
13688: This is not yet implemented.
13689:
13690: Several people like to use Forth as scripting language for applications
13691: that are otherwise written in C, C++, or some other language.
13692:
13693: The Forth system ATLAST provides facilities for embedding it into
13694: applications; unfortunately it has several disadvantages: most
13695: importantly, it is not based on ANS Forth, and it is apparently dead
13696: (i.e., not developed further and not supported). The facilities
13697: provided by Gforth in this area are inspired by ATLAST's facilities, so
13698: making the switch should not be hard.
13699:
13700: We also tried to design the interface such that it can easily be
13701: implemented by other Forth systems, so that we may one day arrive at a
13702: standardized interface. Such a standard interface would allow you to
13703: replace the Forth system without having to rewrite C code.
13704:
13705: You embed the Gforth interpreter by linking with the library
13706: @code{libgforth.a} (give the compiler the option @code{-lgforth}). All
13707: global symbols in this library that belong to the interface, have the
13708: prefix @code{forth_}. (Global symbols that are used internally have the
13709: prefix @code{gforth_}).
13710:
13711: You can include the declarations of Forth types and the functions and
13712: variables of the interface with @code{#include <forth.h>}.
13713:
13714: Types.
13715:
13716: Variables.
13717:
13718: Data and FP Stack pointer. Area sizes.
13719:
13720: functions.
13721:
13722: forth_init(imagefile)
13723: forth_evaluate(string) exceptions?
13724: forth_goto(address) (or forth_execute(xt)?)
13725: forth_continue() (a corountining mechanism)
13726:
13727: Adding primitives.
13728:
13729: No checking.
13730:
13731: Signals?
13732:
13733: Accessing the Stacks
13734:
13735: @c ******************************************************************
13736: @node Emacs and Gforth, Image Files, Integrating Gforth, Top
13737: @chapter Emacs and Gforth
13738: @cindex Emacs and Gforth
13739:
13740: @cindex @file{gforth.el}
13741: @cindex @file{forth.el}
13742: @cindex Rydqvist, Goran
13743: @cindex comment editing commands
13744: @cindex @code{\}, editing with Emacs
13745: @cindex debug tracer editing commands
13746: @cindex @code{~~}, removal with Emacs
13747: @cindex Forth mode in Emacs
13748: Gforth comes with @file{gforth.el}, an improved version of
13749: @file{forth.el} by Goran Rydqvist (included in the TILE package). The
13750: improvements are:
13751:
13752: @itemize @bullet
13753: @item
13754: A better (but still not perfect) handling of indentation.
13755: @item
13756: Comment paragraph filling (@kbd{M-q})
13757: @item
13758: Commenting (@kbd{C-x \}) and uncommenting (@kbd{C-u C-x \}) of regions
13759: @item
13760: Removal of debugging tracers (@kbd{C-x ~}, @pxref{Debugging}).
13761: @item
13762: Support of the @code{info-lookup} feature for looking up the
13763: documentation of a word.
13764: @end itemize
13765:
13766: I left the stuff I do not use alone, even though some of it only makes
13767: sense for TILE. To get a description of these features, enter Forth mode
13768: and type @kbd{C-h m}.
13769:
13770: @cindex source location of error or debugging output in Emacs
13771: @cindex error output, finding the source location in Emacs
13772: @cindex debugging output, finding the source location in Emacs
13773: In addition, Gforth supports Emacs quite well: The source code locations
13774: given in error messages, debugging output (from @code{~~}) and failed
13775: assertion messages are in the right format for Emacs' compilation mode
13776: (@pxref{Compilation, , Running Compilations under Emacs, emacs, Emacs
13777: Manual}) so the source location corresponding to an error or other
13778: message is only a few keystrokes away (@kbd{C-x `} for the next error,
13779: @kbd{C-c C-c} for the error under the cursor).
13780:
13781: @cindex @file{TAGS} file
13782: @cindex @file{etags.fs}
13783: @cindex viewing the source of a word in Emacs
13784: @cindex @code{require}, placement in files
13785: @cindex @code{include}, placement in files
13786: Also, if you @code{require} @file{etags.fs}, a new @file{TAGS} file will
13787: be produced (@pxref{Tags, , Tags Tables, emacs, Emacs Manual}) that
13788: contains the definitions of all words defined afterwards. You can then
13789: find the source for a word using @kbd{M-.}. Note that emacs can use
13790: several tags files at the same time (e.g., one for the Gforth sources
13791: and one for your program, @pxref{Select Tags Table,,Selecting a Tags
13792: Table,emacs, Emacs Manual}). The TAGS file for the preloaded words is
13793: @file{$(datadir)/gforth/$(VERSION)/TAGS} (e.g.,
13794: @file{/usr/local/share/gforth/0.2.0/TAGS}). To get the best behaviour
13795: with @file{etags.fs}, you should avoid putting definitions both before
13796: and after @code{require} etc., otherwise you will see the same file
13797: visited several times by commands like @code{tags-search}.
13798:
13799: @cindex viewing the documentation of a word in Emacs
13800: @cindex context-sensitive help
13801: Moreover, for words documented in this manual, you can look up the
13802: glossary entry quickly by using @kbd{C-h TAB}
13803: (@code{info-lookup-symbol}, @pxref{Documentation, ,Documentation
13804: Commands, emacs, Emacs Manual}). This feature requires Emacs 20.3 or
13805: later and does not work for words containing @code{:}.
13806:
13807:
13808: @cindex @file{.emacs}
13809: To get all these benefits, add the following lines to your @file{.emacs}
13810: file:
13811:
13812: @example
13813: (autoload 'forth-mode "gforth.el")
13814: (setq auto-mode-alist (cons '("\\.fs\\'" . forth-mode) auto-mode-alist))
13815: @end example
13816:
13817: @c ******************************************************************
13818: @node Image Files, Engine, Emacs and Gforth, Top
13819: @chapter Image Files
13820: @cindex image file
13821: @cindex @file{.fi} files
13822: @cindex precompiled Forth code
13823: @cindex dictionary in persistent form
13824: @cindex persistent form of dictionary
13825:
13826: An image file is a file containing an image of the Forth dictionary,
13827: i.e., compiled Forth code and data residing in the dictionary. By
13828: convention, we use the extension @code{.fi} for image files.
13829:
13830: @menu
13831: * Image Licensing Issues:: Distribution terms for images.
13832: * Image File Background:: Why have image files?
13833: * Non-Relocatable Image Files:: don't always work.
13834: * Data-Relocatable Image Files:: are better.
13835: * Fully Relocatable Image Files:: better yet.
13836: * Stack and Dictionary Sizes:: Setting the default sizes for an image.
13837: * Running Image Files:: @code{gforth -i @i{file}} or @i{file}.
13838: * Modifying the Startup Sequence:: and turnkey applications.
13839: @end menu
13840:
13841: @node Image Licensing Issues, Image File Background, Image Files, Image Files
13842: @section Image Licensing Issues
13843: @cindex license for images
13844: @cindex image license
13845:
13846: An image created with @code{gforthmi} (@pxref{gforthmi}) or
13847: @code{savesystem} (@pxref{Non-Relocatable Image Files}) includes the
13848: original image; i.e., according to copyright law it is a derived work of
13849: the original image.
13850:
13851: Since Gforth is distributed under the GNU GPL, the newly created image
13852: falls under the GNU GPL, too. In particular, this means that if you
13853: distribute the image, you have to make all of the sources for the image
13854: available, including those you wrote. For details see @ref{License, ,
13855: GNU General Public License (Section 3)}.
13856:
13857: If you create an image with @code{cross} (@pxref{cross.fs}), the image
13858: contains only code compiled from the sources you gave it; if none of
13859: these sources is under the GPL, the terms discussed above do not apply
13860: to the image. However, if your image needs an engine (a gforth binary)
13861: that is under the GPL, you should make sure that you distribute both in
13862: a way that is at most a @emph{mere aggregation}, if you don't want the
13863: terms of the GPL to apply to the image.
13864:
13865: @node Image File Background, Non-Relocatable Image Files, Image Licensing Issues, Image Files
13866: @section Image File Background
13867: @cindex image file background
13868:
13869: Gforth consists not only of primitives (in the engine), but also of
13870: definitions written in Forth. Since the Forth compiler itself belongs to
13871: those definitions, it is not possible to start the system with the
13872: engine and the Forth source alone. Therefore we provide the Forth
13873: code as an image file in nearly executable form. When Gforth starts up,
13874: a C routine loads the image file into memory, optionally relocates the
13875: addresses, then sets up the memory (stacks etc.) according to
13876: information in the image file, and (finally) starts executing Forth
13877: code.
13878:
13879: The image file variants represent different compromises between the
13880: goals of making it easy to generate image files and making them
13881: portable.
13882:
13883: @cindex relocation at run-time
13884: Win32Forth 3.4 and Mitch Bradley's @code{cforth} use relocation at
13885: run-time. This avoids many of the complications discussed below (image
13886: files are data relocatable without further ado), but costs performance
13887: (one addition per memory access).
13888:
13889: @cindex relocation at load-time
13890: By contrast, the Gforth loader performs relocation at image load time. The
13891: loader also has to replace tokens that represent primitive calls with the
13892: appropriate code-field addresses (or code addresses in the case of
13893: direct threading).
13894:
13895: There are three kinds of image files, with different degrees of
13896: relocatability: non-relocatable, data-relocatable, and fully relocatable
13897: image files.
13898:
13899: @cindex image file loader
13900: @cindex relocating loader
13901: @cindex loader for image files
13902: These image file variants have several restrictions in common; they are
13903: caused by the design of the image file loader:
13904:
13905: @itemize @bullet
13906: @item
13907: There is only one segment; in particular, this means, that an image file
13908: cannot represent @code{ALLOCATE}d memory chunks (and pointers to
13909: them). The contents of the stacks are not represented, either.
13910:
13911: @item
13912: The only kinds of relocation supported are: adding the same offset to
13913: all cells that represent data addresses; and replacing special tokens
13914: with code addresses or with pieces of machine code.
13915:
13916: If any complex computations involving addresses are performed, the
13917: results cannot be represented in the image file. Several applications that
13918: use such computations come to mind:
13919: @itemize @minus
13920: @item
13921: Hashing addresses (or data structures which contain addresses) for table
13922: lookup. If you use Gforth's @code{table}s or @code{wordlist}s for this
13923: purpose, you will have no problem, because the hash tables are
13924: recomputed automatically when the system is started. If you use your own
13925: hash tables, you will have to do something similar.
13926:
13927: @item
13928: There's a cute implementation of doubly-linked lists that uses
13929: @code{XOR}ed addresses. You could represent such lists as singly-linked
13930: in the image file, and restore the doubly-linked representation on
13931: startup.@footnote{In my opinion, though, you should think thrice before
13932: using a doubly-linked list (whatever implementation).}
13933:
13934: @item
13935: The code addresses of run-time routines like @code{docol:} cannot be
13936: represented in the image file (because their tokens would be replaced by
13937: machine code in direct threaded implementations). As a workaround,
13938: compute these addresses at run-time with @code{>code-address} from the
13939: executions tokens of appropriate words (see the definitions of
13940: @code{docol:} and friends in @file{kernel/getdoers.fs}).
13941:
13942: @item
13943: On many architectures addresses are represented in machine code in some
13944: shifted or mangled form. You cannot put @code{CODE} words that contain
13945: absolute addresses in this form in a relocatable image file. Workarounds
13946: are representing the address in some relative form (e.g., relative to
13947: the CFA, which is present in some register), or loading the address from
13948: a place where it is stored in a non-mangled form.
13949: @end itemize
13950: @end itemize
13951:
13952: @node Non-Relocatable Image Files, Data-Relocatable Image Files, Image File Background, Image Files
13953: @section Non-Relocatable Image Files
13954: @cindex non-relocatable image files
13955: @cindex image file, non-relocatable
13956:
13957: These files are simple memory dumps of the dictionary. They are specific
13958: to the executable (i.e., @file{gforth} file) they were created
13959: with. What's worse, they are specific to the place on which the
13960: dictionary resided when the image was created. Now, there is no
13961: guarantee that the dictionary will reside at the same place the next
13962: time you start Gforth, so there's no guarantee that a non-relocatable
13963: image will work the next time (Gforth will complain instead of crashing,
13964: though).
13965:
13966: You can create a non-relocatable image file with
13967:
13968:
13969: doc-savesystem
13970:
13971:
13972: @node Data-Relocatable Image Files, Fully Relocatable Image Files, Non-Relocatable Image Files, Image Files
13973: @section Data-Relocatable Image Files
13974: @cindex data-relocatable image files
13975: @cindex image file, data-relocatable
13976:
13977: These files contain relocatable data addresses, but fixed code addresses
13978: (instead of tokens). They are specific to the executable (i.e.,
13979: @file{gforth} file) they were created with. For direct threading on some
13980: architectures (e.g., the i386), data-relocatable images do not work. You
13981: get a data-relocatable image, if you use @file{gforthmi} with a
13982: Gforth binary that is not doubly indirect threaded (@pxref{Fully
13983: Relocatable Image Files}).
13984:
13985: @node Fully Relocatable Image Files, Stack and Dictionary Sizes, Data-Relocatable Image Files, Image Files
13986: @section Fully Relocatable Image Files
13987: @cindex fully relocatable image files
13988: @cindex image file, fully relocatable
13989:
13990: @cindex @file{kern*.fi}, relocatability
13991: @cindex @file{gforth.fi}, relocatability
13992: These image files have relocatable data addresses, and tokens for code
13993: addresses. They can be used with different binaries (e.g., with and
13994: without debugging) on the same machine, and even across machines with
13995: the same data formats (byte order, cell size, floating point
13996: format). However, they are usually specific to the version of Gforth
13997: they were created with. The files @file{gforth.fi} and @file{kernl*.fi}
13998: are fully relocatable.
13999:
14000: There are two ways to create a fully relocatable image file:
14001:
14002: @menu
14003: * gforthmi:: The normal way
14004: * cross.fs:: The hard way
14005: @end menu
14006:
14007: @node gforthmi, cross.fs, Fully Relocatable Image Files, Fully Relocatable Image Files
14008: @subsection @file{gforthmi}
14009: @cindex @file{comp-i.fs}
14010: @cindex @file{gforthmi}
14011:
14012: You will usually use @file{gforthmi}. If you want to create an
14013: image @i{file} that contains everything you would load by invoking
14014: Gforth with @code{gforth @i{options}}, you simply say:
14015: @example
14016: gforthmi @i{file} @i{options}
14017: @end example
14018:
14019: E.g., if you want to create an image @file{asm.fi} that has the file
14020: @file{asm.fs} loaded in addition to the usual stuff, you could do it
14021: like this:
14022:
14023: @example
14024: gforthmi asm.fi asm.fs
14025: @end example
14026:
14027: @file{gforthmi} is implemented as a sh script and works like this: It
14028: produces two non-relocatable images for different addresses and then
14029: compares them. Its output reflects this: first you see the output (if
14030: any) of the two Gforth invocations that produce the non-relocatable image
14031: files, then you see the output of the comparing program: It displays the
14032: offset used for data addresses and the offset used for code addresses;
14033: moreover, for each cell that cannot be represented correctly in the
14034: image files, it displays a line like this:
14035:
14036: @example
14037: 78DC BFFFFA50 BFFFFA40
14038: @end example
14039:
14040: This means that at offset $78dc from @code{forthstart}, one input image
14041: contains $bffffa50, and the other contains $bffffa40. Since these cells
14042: cannot be represented correctly in the output image, you should examine
14043: these places in the dictionary and verify that these cells are dead
14044: (i.e., not read before they are written).
14045:
14046: @cindex --application, @code{gforthmi} option
14047: If you insert the option @code{--application} in front of the image file
14048: name, you will get an image that uses the @code{--appl-image} option
14049: instead of the @code{--image-file} option (@pxref{Invoking
14050: Gforth}). When you execute such an image on Unix (by typing the image
14051: name as command), the Gforth engine will pass all options to the image
14052: instead of trying to interpret them as engine options.
14053:
14054: If you type @file{gforthmi} with no arguments, it prints some usage
14055: instructions.
14056:
14057: @cindex @code{savesystem} during @file{gforthmi}
14058: @cindex @code{bye} during @file{gforthmi}
14059: @cindex doubly indirect threaded code
14060: @cindex environment variables
14061: @cindex @code{GFORTHD} -- environment variable
14062: @cindex @code{GFORTH} -- environment variable
14063: @cindex @code{gforth-ditc}
14064: There are a few wrinkles: After processing the passed @i{options}, the
14065: words @code{savesystem} and @code{bye} must be visible. A special doubly
14066: indirect threaded version of the @file{gforth} executable is used for
14067: creating the non-relocatable images; you can pass the exact filename of
14068: this executable through the environment variable @code{GFORTHD}
14069: (default: @file{gforth-ditc}); if you pass a version that is not doubly
14070: indirect threaded, you will not get a fully relocatable image, but a
14071: data-relocatable image (because there is no code address offset). The
14072: normal @file{gforth} executable is used for creating the relocatable
14073: image; you can pass the exact filename of this executable through the
14074: environment variable @code{GFORTH}.
14075:
14076: @node cross.fs, , gforthmi, Fully Relocatable Image Files
14077: @subsection @file{cross.fs}
14078: @cindex @file{cross.fs}
14079: @cindex cross-compiler
14080: @cindex metacompiler
14081: @cindex target compiler
14082:
14083: You can also use @code{cross}, a batch compiler that accepts a Forth-like
14084: programming language (@pxref{Cross Compiler}).
14085:
14086: @code{cross} allows you to create image files for machines with
14087: different data sizes and data formats than the one used for generating
14088: the image file. You can also use it to create an application image that
14089: does not contain a Forth compiler. These features are bought with
14090: restrictions and inconveniences in programming. E.g., addresses have to
14091: be stored in memory with special words (@code{A!}, @code{A,}, etc.) in
14092: order to make the code relocatable.
14093:
14094:
14095: @node Stack and Dictionary Sizes, Running Image Files, Fully Relocatable Image Files, Image Files
14096: @section Stack and Dictionary Sizes
14097: @cindex image file, stack and dictionary sizes
14098: @cindex dictionary size default
14099: @cindex stack size default
14100:
14101: If you invoke Gforth with a command line flag for the size
14102: (@pxref{Invoking Gforth}), the size you specify is stored in the
14103: dictionary. If you save the dictionary with @code{savesystem} or create
14104: an image with @file{gforthmi}, this size will become the default
14105: for the resulting image file. E.g., the following will create a
14106: fully relocatable version of @file{gforth.fi} with a 1MB dictionary:
14107:
14108: @example
14109: gforthmi gforth.fi -m 1M
14110: @end example
14111:
14112: In other words, if you want to set the default size for the dictionary
14113: and the stacks of an image, just invoke @file{gforthmi} with the
14114: appropriate options when creating the image.
14115:
14116: @cindex stack size, cache-friendly
14117: Note: For cache-friendly behaviour (i.e., good performance), you should
14118: make the sizes of the stacks modulo, say, 2K, somewhat different. E.g.,
14119: the default stack sizes are: data: 16k (mod 2k=0); fp: 15.5k (mod
14120: 2k=1.5k); return: 15k(mod 2k=1k); locals: 14.5k (mod 2k=0.5k).
14121:
14122: @node Running Image Files, Modifying the Startup Sequence, Stack and Dictionary Sizes, Image Files
14123: @section Running Image Files
14124: @cindex running image files
14125: @cindex invoking image files
14126: @cindex image file invocation
14127:
14128: @cindex -i, invoke image file
14129: @cindex --image file, invoke image file
14130: You can invoke Gforth with an image file @i{image} instead of the
14131: default @file{gforth.fi} with the @code{-i} flag (@pxref{Invoking Gforth}):
14132: @example
14133: gforth -i @i{image}
14134: @end example
14135:
14136: @cindex executable image file
14137: @cindex image file, executable
14138: If your operating system supports starting scripts with a line of the
14139: form @code{#! ...}, you just have to type the image file name to start
14140: Gforth with this image file (note that the file extension @code{.fi} is
14141: just a convention). I.e., to run Gforth with the image file @i{image},
14142: you can just type @i{image} instead of @code{gforth -i @i{image}}.
14143: This works because every @code{.fi} file starts with a line of this
14144: format:
14145:
14146: @example
14147: #! /usr/local/bin/gforth-0.4.0 -i
14148: @end example
14149:
14150: The file and pathname for the Gforth engine specified on this line is
14151: the specific Gforth executable that it was built against; i.e. the value
14152: of the environment variable @code{GFORTH} at the time that
14153: @file{gforthmi} was executed.
14154:
14155: You can make use of the same shell capability to make a Forth source
14156: file into an executable. For example, if you place this text in a file:
14157:
14158: @example
14159: #! /usr/local/bin/gforth
14160:
14161: ." Hello, world" CR
14162: bye
14163: @end example
14164:
14165: @noindent
14166: and then make the file executable (chmod +x in Unix), you can run it
14167: directly from the command line. The sequence @code{#!} is used in two
14168: ways; firstly, it is recognised as a ``magic sequence'' by the operating
14169: system@footnote{The Unix kernel actually recognises two types of files:
14170: executable files and files of data, where the data is processed by an
14171: interpreter that is specified on the ``interpreter line'' -- the first
14172: line of the file, starting with the sequence #!. There may be a small
14173: limit (e.g., 32) on the number of characters that may be specified on
14174: the interpreter line.} secondly it is treated as a comment character by
14175: Gforth. Because of the second usage, a space is required between
14176: @code{#!} and the path to the executable (moreover, some Unixes
14177: require the sequence @code{#! /}).
14178:
14179: The disadvantage of this latter technique, compared with using
14180: @file{gforthmi}, is that it is slightly slower; the Forth source code is
14181: compiled on-the-fly, each time the program is invoked.
14182:
14183: doc-#!
14184:
14185:
14186: @node Modifying the Startup Sequence, , Running Image Files, Image Files
14187: @section Modifying the Startup Sequence
14188: @cindex startup sequence for image file
14189: @cindex image file initialization sequence
14190: @cindex initialization sequence of image file
14191:
14192: You can add your own initialization to the startup sequence through the
14193: deferred word @code{'cold}. @code{'cold} is invoked just before the
14194: image-specific command line processing (i.e., loading files and
14195: evaluating (@code{-e}) strings) starts.
14196:
14197: A sequence for adding your initialization usually looks like this:
14198:
14199: @example
14200: :noname
14201: Defers 'cold \ do other initialization stuff (e.g., rehashing wordlists)
14202: ... \ your stuff
14203: ; IS 'cold
14204: @end example
14205:
14206: @cindex turnkey image files
14207: @cindex image file, turnkey applications
14208: You can make a turnkey image by letting @code{'cold} execute a word
14209: (your turnkey application) that never returns; instead, it exits Gforth
14210: via @code{bye} or @code{throw}.
14211:
14212: @cindex command-line arguments, access
14213: @cindex arguments on the command line, access
14214: You can access the (image-specific) command-line arguments through the
14215: variables @code{argc} and @code{argv}. @code{arg} provides convenient
14216: access to @code{argv}.
14217:
14218: If @code{'cold} exits normally, Gforth processes the command-line
14219: arguments as files to be loaded and strings to be evaluated. Therefore,
14220: @code{'cold} should remove the arguments it has used in this case.
14221:
14222:
14223:
14224: doc-'cold
14225: doc-argc
14226: doc-argv
14227: doc-arg
14228:
14229:
14230:
14231: @c ******************************************************************
14232: @node Engine, Binding to System Library, Image Files, Top
14233: @chapter Engine
14234: @cindex engine
14235: @cindex virtual machine
14236:
14237: Reading this chapter is not necessary for programming with Gforth. It
14238: may be helpful for finding your way in the Gforth sources.
14239:
14240: The ideas in this section have also been published in Bernd Paysan,
14241: @cite{ANS fig/GNU/??? Forth} (in German), Forth-Tagung '93 and M. Anton
14242: Ertl, @cite{@uref{http://www.complang.tuwien.ac.at/papers/ertl93.ps.Z, A
14243: Portable Forth Engine}}, EuroForth '93.
14244:
14245: @menu
14246: * Portability::
14247: * Threading::
14248: * Primitives::
14249: * Performance::
14250: @end menu
14251:
14252: @node Portability, Threading, Engine, Engine
14253: @section Portability
14254: @cindex engine portability
14255:
14256: An important goal of the Gforth Project is availability across a wide
14257: range of personal machines. fig-Forth, and, to a lesser extent, F83,
14258: achieved this goal by manually coding the engine in assembly language
14259: for several then-popular processors. This approach is very
14260: labor-intensive and the results are short-lived due to progress in
14261: computer architecture.
14262:
14263: @cindex C, using C for the engine
14264: Others have avoided this problem by coding in C, e.g., Mitch Bradley
14265: (cforth), Mikael Patel (TILE) and Dirk Zoller (pfe). This approach is
14266: particularly popular for UNIX-based Forths due to the large variety of
14267: architectures of UNIX machines. Unfortunately an implementation in C
14268: does not mix well with the goals of efficiency and with using
14269: traditional techniques: Indirect or direct threading cannot be expressed
14270: in C, and switch threading, the fastest technique available in C, is
14271: significantly slower. Another problem with C is that it is very
14272: cumbersome to express double integer arithmetic.
14273:
14274: @cindex GNU C for the engine
14275: @cindex long long
14276: Fortunately, there is a portable language that does not have these
14277: limitations: GNU C, the version of C processed by the GNU C compiler
14278: (@pxref{C Extensions, , Extensions to the C Language Family, gcc.info,
14279: GNU C Manual}). Its labels as values feature (@pxref{Labels as Values, ,
14280: Labels as Values, gcc.info, GNU C Manual}) makes direct and indirect
14281: threading possible, its @code{long long} type (@pxref{Long Long, ,
14282: Double-Word Integers, gcc.info, GNU C Manual}) corresponds to Forth's
14283: double numbers@footnote{Unfortunately, long longs are not implemented
14284: properly on all machines (e.g., on alpha-osf1, long longs are only 64
14285: bits, the same size as longs (and pointers), but they should be twice as
14286: long according to @pxref{Long Long, , Double-Word Integers, gcc.info, GNU
14287: C Manual}). So, we had to implement doubles in C after all. Still, on
14288: most machines we can use long longs and achieve better performance than
14289: with the emulation package.}. GNU C is available for free on all
14290: important (and many unimportant) UNIX machines, VMS, 80386s running
14291: MS-DOS, the Amiga, and the Atari ST, so a Forth written in GNU C can run
14292: on all these machines.
14293:
14294: Writing in a portable language has the reputation of producing code that
14295: is slower than assembly. For our Forth engine we repeatedly looked at
14296: the code produced by the compiler and eliminated most compiler-induced
14297: inefficiencies by appropriate changes in the source code.
14298:
14299: @cindex explicit register declarations
14300: @cindex --enable-force-reg, configuration flag
14301: @cindex -DFORCE_REG
14302: However, register allocation cannot be portably influenced by the
14303: programmer, leading to some inefficiencies on register-starved
14304: machines. We use explicit register declarations (@pxref{Explicit Reg
14305: Vars, , Variables in Specified Registers, gcc.info, GNU C Manual}) to
14306: improve the speed on some machines. They are turned on by using the
14307: configuration flag @code{--enable-force-reg} (@code{gcc} switch
14308: @code{-DFORCE_REG}). Unfortunately, this feature not only depends on the
14309: machine, but also on the compiler version: On some machines some
14310: compiler versions produce incorrect code when certain explicit register
14311: declarations are used. So by default @code{-DFORCE_REG} is not used.
14312:
14313: @node Threading, Primitives, Portability, Engine
14314: @section Threading
14315: @cindex inner interpreter implementation
14316: @cindex threaded code implementation
14317:
14318: @cindex labels as values
14319: GNU C's labels as values extension (available since @code{gcc-2.0},
14320: @pxref{Labels as Values, , Labels as Values, gcc.info, GNU C Manual})
14321: makes it possible to take the address of @i{label} by writing
14322: @code{&&@i{label}}. This address can then be used in a statement like
14323: @code{goto *@i{address}}. I.e., @code{goto *&&x} is the same as
14324: @code{goto x}.
14325:
14326: @cindex @code{NEXT}, indirect threaded
14327: @cindex indirect threaded inner interpreter
14328: @cindex inner interpreter, indirect threaded
14329: With this feature an indirect threaded @code{NEXT} looks like:
14330: @example
14331: cfa = *ip++;
14332: ca = *cfa;
14333: goto *ca;
14334: @end example
14335: @cindex instruction pointer
14336: For those unfamiliar with the names: @code{ip} is the Forth instruction
14337: pointer; the @code{cfa} (code-field address) corresponds to ANS Forths
14338: execution token and points to the code field of the next word to be
14339: executed; The @code{ca} (code address) fetched from there points to some
14340: executable code, e.g., a primitive or the colon definition handler
14341: @code{docol}.
14342:
14343: @cindex @code{NEXT}, direct threaded
14344: @cindex direct threaded inner interpreter
14345: @cindex inner interpreter, direct threaded
14346: Direct threading is even simpler:
14347: @example
14348: ca = *ip++;
14349: goto *ca;
14350: @end example
14351:
14352: Of course we have packaged the whole thing neatly in macros called
14353: @code{NEXT} and @code{NEXT1} (the part of @code{NEXT} after fetching the cfa).
14354:
14355: @menu
14356: * Scheduling::
14357: * Direct or Indirect Threaded?::
14358: * DOES>::
14359: @end menu
14360:
14361: @node Scheduling, Direct or Indirect Threaded?, Threading, Threading
14362: @subsection Scheduling
14363: @cindex inner interpreter optimization
14364:
14365: There is a little complication: Pipelined and superscalar processors,
14366: i.e., RISC and some modern CISC machines can process independent
14367: instructions while waiting for the results of an instruction. The
14368: compiler usually reorders (schedules) the instructions in a way that
14369: achieves good usage of these delay slots. However, on our first tries
14370: the compiler did not do well on scheduling primitives. E.g., for
14371: @code{+} implemented as
14372: @example
14373: n=sp[0]+sp[1];
14374: sp++;
14375: sp[0]=n;
14376: NEXT;
14377: @end example
14378: the @code{NEXT} comes strictly after the other code, i.e., there is
14379: nearly no scheduling. After a little thought the problem becomes clear:
14380: The compiler cannot know that @code{sp} and @code{ip} point to different
14381: addresses (and the version of @code{gcc} we used would not know it even
14382: if it was possible), so it could not move the load of the cfa above the
14383: store to the TOS. Indeed the pointers could be the same, if code on or
14384: very near the top of stack were executed. In the interest of speed we
14385: chose to forbid this probably unused ``feature'' and helped the compiler
14386: in scheduling: @code{NEXT} is divided into several parts:
14387: @code{NEXT_P0}, @code{NEXT_P1} and @code{NEXT_P2}). @code{+} now looks
14388: like:
14389: @example
14390: NEXT_P0;
14391: n=sp[0]+sp[1];
14392: sp++;
14393: NEXT_P1;
14394: sp[0]=n;
14395: NEXT_P2;
14396: @end example
14397:
14398: There are various schemes that distribute the different operations of
14399: NEXT between these parts in several ways; in general, different schemes
14400: perform best on different processors. We use a scheme for most
14401: architectures that performs well for most processors of this
14402: architecture; in the furture we may switch to benchmarking and chosing
14403: the scheme on installation time.
14404:
14405:
14406: @node Direct or Indirect Threaded?, DOES>, Scheduling, Threading
14407: @subsection Direct or Indirect Threaded?
14408: @cindex threading, direct or indirect?
14409:
14410: @cindex -DDIRECT_THREADED
14411: Both! After packaging the nasty details in macro definitions we
14412: realized that we could switch between direct and indirect threading by
14413: simply setting a compilation flag (@code{-DDIRECT_THREADED}) and
14414: defining a few machine-specific macros for the direct-threading case.
14415: On the Forth level we also offer access words that hide the
14416: differences between the threading methods (@pxref{Threading Words}).
14417:
14418: Indirect threading is implemented completely machine-independently.
14419: Direct threading needs routines for creating jumps to the executable
14420: code (e.g. to @code{docol} or @code{dodoes}). These routines are inherently
14421: machine-dependent, but they do not amount to many source lines. Therefore,
14422: even porting direct threading to a new machine requires little effort.
14423:
14424: @cindex --enable-indirect-threaded, configuration flag
14425: @cindex --enable-direct-threaded, configuration flag
14426: The default threading method is machine-dependent. You can enforce a
14427: specific threading method when building Gforth with the configuration
14428: flag @code{--enable-direct-threaded} or
14429: @code{--enable-indirect-threaded}. Note that direct threading is not
14430: supported on all machines.
14431:
14432: @node DOES>, , D![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to gforth.ds CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [gforth] / gforth / doc